The Impact of Participation in Extracurricular Activities on School Achievement of French Middle School Students: Human Capital and Cultural Capital Revisited

The Impact of Participation in Extracurricular Activities on School Achievement of French Middle... Abstract The impact of participation in extracurricular activities on academic success has long been studied in the social sciences. This article aims at improving the measurement and understanding of this impact. Based on panel data regression models applied to a panel of French middle school students, it first provides a robust estimation of the impact of extracurricular activities on school outcomes (marks in French and Mathematics) and on a set of cognitive and non-cognitive skills. It finds a positive and significant impact on marks in French and Mathematics and scores on non-cognitive skills tests. No impact is found on cognitive skills. The article then investigates the underlying mechanisms of this impact. Its findings do not reinforce the transfer paradigm, according to which extracurricular activities provide students who participate in them with skills that they can reinvest in school life. Neither does it support the notion that such an impact may primarily be the result of students’ greater connivance with the cultural standards of teachers. Instead, it seems likely that what is mainly at stake in participation in extracurricular activities is families’ unequal capacity for extending the time of school supervision in their children’s free time. Therefore, insofar as the varying participation in these activities is strongly correlated to differences in students’ social and cultural background, participation in extracurricular activities would in itself contribute to reinforcing social inequalities in school achievement. Introduction Extracurricular activities have long been considered as a source of inequality among children in relation to school achievement (Covay and Carbonaro 2010; Dumais 2006; Lareau 2003; Lareau and Weininger 2008). According to a restrictive definition, extracurricular activities encompass non-compulsory activities that take place at school but are not included in the curriculum (Bartkus et al. 2012). Understood in a broader sense, they encompass all voluntary organized activities, whether these take place in the school setting or not: sports or artistic practices, such as musical activities, but also participation in youth organizations (e.g., scouting) or leisure clubs. The main rationale for studying the impact of extracurricular activities on educational outcomes is the idea that learning processes are not restricted to explicit transmission in the classroom context. Moreover, since extracurricular activities are not compulsory, their impact is expected to be unequal depending on students’ social background. This article addresses these questions in the current context in France. It is based on a panel of several thousand French middle school (collège) students. Collège is the French equivalent of middle school or junior high school. It corresponds to the first stage of secondary education, with all young people aged 11 to 15 receiving the same general education. So far, the impact of extracurricular activities has been extensively studied at the high school and university level, but less at the middle school level (Jansen 2016). Extracurricular time deserves special attention in the French context due to the specific way of organizing school hours, as compared to other European countries and to the United States. The French school system is characterized by a lower annual number of days of schooling than elsewhere, but the school days are longer (European Commission 2016). In addition, school time tends to be highly concentrated on the core school subjects and, unlike English-speaking countries, schools play a relatively minor role in the management of extracurricular activities in France, where they are mainly supported by the non-profit and volunteer sectors. As a result, extracurricular time remains largely beyond the control of the educational system and more dependent on family resources. In this respect, the French case, where school life and extracurricular activities are much more clearly separated than in many other countries, may contribute to a better identification and understanding of the intrinsic impact of extracurricular activities on school performance in general. The article first reviews the existing literature on extracurricular activities’ observed educational payoffs. It then carries out fixed-effects (FE) regression models in order to demonstrate the robustness of the impact of participation in extracurricular activities on the academic achievement of French middle school students. The rest of the article investigates the processes through which this impact may be seen to be produced. Three competing mechanisms are considered, according to which the impact of extracurricular activities on school achievement is seen to be primarily due to a process of skills accumulation and transfer (human capital model); to a process of social selection (cultural capital model); or to its time-structuring function (“concerted cultivation” model). Finally, the article questions the extent to which this impact varies according to the nature of the activities (cultural vs. non-cultural, outside school vs. inside school). The aim of the article is thus threefold. First, by exploiting the possibilities contained in the panel data on which it relies, it offers a more robust measure of the educational payoff of extracurricular activities than much of the existing research provides. Second, it suggests that the impact of extracurricular activities is more a matter of cultural capital than human capital, which is why these activities would contribute to reinforcing social inequalities in school achievement. Third, it strongly suggests that participation in extracurricular activities affects academic performance through the control it exerts on adolescents’ time use. Literature review: How and to what extent does participation in extracurricular activities affect school achievement? The robustness of extracurricular activities’ educational payoffs The relationship between cognitive development in children and adolescents, their academic performance, and how they spend their time has long been investigated (Larson and Verma 1999). Particular attention has been given to the time devoted to scheduled and supervised leisure activities, such as extracurricular activities (Flammer and Alsaker 1999), whose impact on academic performance has been extensively debated (Seow and Pan 2014; Shulruf 2010). It has long been argued that extracurricular activities might be detrimental to school success (Camp 1990; Coleman 1961), or at least play an ambivalent role in this respect (Broh 2002; Fejgin 1994; Marsh 1992; Marsh and Kleitman 2002; Steinberg 1996), insofar as the time they require is in direct competition with time devoted to academic pursuits. In addition, primarily recreational extracurricular activities have sometimes been deemed to encourage attitudes antagonistic to school values (Coleman 1961; Fejgin 1994). Since the early 2000s, however, an insistence on the positive effects of extracurricular activities has become dominant (Covay and Carbonaro 2010; Feldman and Matjasko 2005; Fletcher, Nickerson, and Wright 2003; Guest and Schneider 2003; Kaufman and Gabler 2004; Lareau 2003; Lipscomb 2007). More recently, several studies have put forward the notion of a possible threshold effect (Seow and Pan 2014): up to a certain point, academic achievement would benefit from participation in extracurricular activities, but excessive commitment of students’ time would result in declining performance (Marsh 1992; Marsh and Kleitman 2002). However, this hypothesis remains rather contentious (Mahoney, Harris, and Eccles 2006). Moreover, the impact of extracurricular activities on school outcomes seems to vary from one activity to another, with mixed results. While some studies insist on the stronger impact of sport in comparison with all other activities (Broh 2002; Marsh 1992; Marsh and Kleitman 2002; McNeal 1995), others emphasize the specific impact of cultural activities, especially those related to music (Hille and Schupp 2015), but with an even stronger impact when combined with sports activities (Cabane, Hille, and Lechner 2015), suggesting a cumulative effect: the more students are involved in diverse activities, whatever they are, the better their academic results (Lareau 2003). Human capital vs. cultural capital Most controversies in this field relate to the very nature of the impact of extracurricular activities on academic achievement, if any. The main divide is between interpretations in terms of human capital and interpretations in terms of cultural capital (Farkas 1996; Kaufman and Gabler 2004). Cognitive and non-cognitive skills Human capital approaches contend that involvement in these activities enhances students’ skills. Skill enhancement may be of a various nature, cognitive as well as non-cognitive. Cognitive skills enhancement assumes that participation in extracurricular activities has an indirect impact on school performances, by improving students’ intellectual abilities or by providing them with competencies that are transferable and applicable in the academic domain (cognitive transfer paradigm). This transfer mechanism has, however, received little support (Dettermann and Sternberg 1993) beyond some studies related to musical education and musical activities (Schellenberg 2004, 2006, 2011). Instead, much of the existing research suggests that participation in extracurricular activities affects academic performance by improving students’ non-cognitive abilities (Broh 2002; Covay and Carbonaro 2010; Holland and Andre 1987; Seow and Pan 2014). Non-cognitive skills include life skills such as organization, planning, time management (Holland and Andre 1987); effort, perseverance, discipline, emotional stability (Farkas 2003); self-esteem, perseverance (Broh 2002); locus of control, self-confidence (Fejgin 1994); self-reliance (Fletcher, Nickerson, and Wright 2003); valuing achievement, respect of adult authority, and the regulation of interactions with others (Covay and Carbonaro 2010). Cultural capital and child-rearing styles According to cultural capital theory (Bourdieu 1986), the impact of participation in extracurricular activities on school achievement is not at all a matter of the transfer of abilities, whether cognitive or non-cognitive; instead, it mainly acts as a status signal (Kaufman and Gabler 2004). In accordance with Bourdieu’s understanding of the notion of cultural capital, students’ participation in extracurricular activities is a demonstration of their cultural resources and their belonging to the upper classes. In this way, they also demonstrate cultural endowments, dispositions, and attitudes well suited to school requirements, shared with teachers, and rewarded as such. This interpretation assumes that participation in extracurricular activities has no intrinsic efficacy but is rather emblematic of the cultural arbitrariness of the school system. Schools tend to reward the possession and control of a certain number of attributes and cultural competences that tend to be monopolized by the ruling class (Bourdieu and Passeron [1970] 1977). This argument converges with Bowles’s and Gintis’s assertion that the persistence of class advantage across generations is due to the fit between family-inherited behavior traits and the implicit norms of teachers and schools, rather than the transmission of cognitive capacities from parents to children (Bowles and Gintis 1976, 2002). In this sense, participation in extracurricular activities may serve as an instrument of the upper classes’ social closure. Lareau’s research on parents’ educational styles, with its seminal contrast between the “concerted cultivation” and the “accomplishment of natural growth” models, might be considered as a variant of the cultural capital hypothesis (Lareau 2003). “Concerted cultivation” refers to a child-rearing style quite common in the middle class, which emphasizes reasoning and dialogue between children and their parents. “Concerted cultivation” is also based on the encouragement of children’s scheduled activities, deemed to improve their talents and abilities, of which engaging in extracurricular activities would be an emblematic case. “Accomplishment of natural growth,” which is more common among working-class families, focuses instead on authority and discipline rather than on exchange and consultation between children and their parents. In addition, children’s leisure time is mostly dedicated to unstructured and improvised activities. Watching TV at home or playing basketball on the street instead of going to a movie theater or playing football at a sports club may illustrate the contrast between these two opposite educational styles. Consistent with broader cultural dispositions, this contrast may lead to a comparative advantage for children raised in a “concerted cultivation” style. This kind of child-rearing environment fosters an aptitude for a certain kind of reasoning and interacting with adults and authorities that provides children with a greater sense of entitlement and self-confidence in social interactions that might be particularly rewarded in the school context. Similarly, the kind of commitment required by structured leisure activities may be well fitted to the time-structuring abilities required by school practices. Participation in extracurricular activities and social reproduction Many sociologists are driven to interpret the impact of extracurricular activities in terms of cultural capital rather than human capital because these interpretative models throw light on the role that participation in these activities plays in the reproduction of social inequalities. Participation in structured extracurricular activities is particularly known to vary across families, depending primarily on parents’ resources, whether economic (Chin and Phillips 2004) or cultural (Dumais 2006; Lareau 2003; Lareau and Weininger 2008; Weininger et al. 2015). Parents from higher socioeconomic status are much more likely to involve their children in extracurricular activities than parents from lower socioeconomic status. To the extent that participation in extracurricular activities is supposed to stimulate academic success, social inequalities in participation in these activities could therefore be seen as a mean through which social privilege is reproduced over generations. Various studies have found empirical evidence of such a mediating role when it comes to participation in extracurricular activities (Bodovski and Farkas 2008; Covay and Carbonaro 2010). Nonetheless, this mediating impact generally appears to be rather small (it explains only a relatively small proportion of the socioeconomic status effect on school performances). The pitfalls of correlational analysis Because of their correlational nature, many of the existing studies fail to adequately grasp the relationship between extracurricular activities and educational achievement because of self-selection, confounding factors, and unobserved heterogeneity (Holland and Andre 1987, 1988; Marsh 1992). Students who participate in these activities may indeed differ from non-participants in regard to some non-observable or non-measurable characteristics or dispositions likely to also have an impact on academic achievement. Various methodological options may be used to overcome, at least partially, the shortcomings of many of the usual analyses in this field. One of them relies on the implementation of fixed-effects (FE) regressions on panel data (Lipscomb 2007), where respondents are taken as their own controls. But this option is seldom explored in analyses of the impact of extracurricular activities on school outcomes, most often because of a lack of adequate data, that is, panel data. The remainder of this article will take advantage precisely of the panel structure of the data on which it is based to apply this kind of regression model. Data and variables The data comes from a French panel of secondary school students commissioned by the statistical studies department of the French Ministry of Education (Direction de l'évaluation, de la prospective et de la performance [DEPP]). It consists of a random sample of 35,000 French students who entered the first grade of middle school (sixth grade) in September 2007. The panel database includes a large amount of information on students’ trajectories and school outcomes recorded annually. It also includes information on students’ family environment provided by two subsequent mail-out surveys submitted to students’ parents in 2008 and 2011. First, the dataset registers the scores on the national assessment tests in French and Mathematics uniformly administered to all French middle school students at the beginning of their sixth grade (students on average aged between 11 and 12). They consist of two 45-minute sequences of exercises designed to measure essential skills in both subjects. Second, the dataset registers the scores obtained by the same students three years later (June 2011) in various subjects, including French and Mathematics, in the “Brevet national des Collèges” (BNC), the final middle school exam (students on average aged between 14 and 15). Third, the dataset includes scores on a set of tests specifically designed for the panel and aimed at measuring wide-ranging abilities, including both cognitive and conative dimensions. The cognitive dimension is measured by a global cognitive score (COG) calculated from the scores obtained on six tests of elementary skills in vocabulary, encyclopedic or long-term memory, mathematical reasoning, completion of incomplete sentences, silent reading comprehension, and assessment of logical reasoning (Ben Ali and Vourc’h 2015). The conative dimension is addressed through three subscales based on Albert Bandura’s (1990) multidimensional scales of perceived efficacy. The first indexes students’ perceived school efficiency (PESCH), that is, beliefs in their ability to succeed in different academic disciplines. The second corresponds to their perceived social efficiency (PESOC), that is, beliefs in their ability to initiate and maintain social relationships and to regulate interpersonal conflicts. The third indexes their perceived efficiency in self-regulation (PESELF), that is, their ability to resist pressure from their peers to engage in deviant behavior (Blanchard et al. 2013). These tests were also administered to students twice, for the first time at the beginning of 2008 and for the second time in 2011. Finally, in 2008 and 2011, students’ parents were asked, among other things, to indicate on a list of eight extracurricular activities those in which their children participated.1 This resulted in a list of eight dummy variables, including participation in a sports club (outside school), participation in the school sports association, registration in a public library, enrollment in a music academy or in a music school, enrollment in a drama class (out of school), participation in a youth organization (e.g., Scouts), enrollment in a youth center, or participation in an activity club (other than sports) at school.2 The addition of these eight dummy variables defines a scale of participation in extracurricular activities ranging from 0 to 8. The initial sample has been subjected to various sources of attrition. Not all the students originally included in the panel were subjected to the cognitive and conative tests. Similarly, not all parents responded to the postal surveys mentioned above. For the sake of coherence, the analysis is thus restricted to the pupils who were subjected to the tests and whose parents responded to the 2008 and 2011 surveys.3 In addition, the analysis is also restricted to the pupils who sat the BNC “on time,” that is, those who had not repeated a class in the interim, which was nearly 90 percent of the sample. As a result, the subsample on which subsequent analyses are based comprises 19,809 individuals. Finally, the various data collection steps generated some missing values that in all subsequent analysis have been treated by multiple imputation, in order to maintain the size of the sample constant.4 Research questions and hypotheses The first aim of the analysis is to produce robust estimates of the impact of participation in extracurricular activities on school achievement which, given previous research on the subject, is expected to be significant and positive. In addition, the hypothesis of a non-linear effect of participation in extracurricular activities, that is, the hypothesis that participation in these activities is beneficial up to a certain point only, will also be tested. H1a: Participation in extracurricular activities has a significantly positive impact on school outcomes. H1b: Up to a certain point, participation in extracurricular activities has a positive impact on academic performance and a negative impact thereafter. It can also be expected that the impact of extracurricular activities does not only depend on the cumulative number of activities in which students participate, but also on their content. H2: The impact of participation in extracurricular activities on school outcomes varies according to the nature of the activities in which students participate. The second goal of the analysis is to clarify the nature of the impact of extracurricular activities on academic achievement. The theory of human capital suggests that this impact is due to the accumulation of specific skills that students can reinvest in school learning. It has often been advanced that this impact would be primarily a matter of accumulation and transfer of non-cognitive skills, such as those related to self-reliance or perceived efficacy, rather than cognitive skills. H3: Participation in extracurricular activities has a greater impact on non-cognitive than on cognitive skills. Insofar as the nature and content of extracurricular activities can at times be very far from those associated with academic goals, its impact on school outcomes is expected to be mainly indirect. Controlling for its impact on cognitive and non-cognitive skills, the remaining impact of extracurricular activities on school outcomes is thus expected to be insignificant. H4: The indirect impact of participation in extracurricular activities on school outcomes is mediated by its direct impact on cognitive and non-cognitive skills. Finally, the cultural capital theory provides an alternative explanation of the mediating function of extracurricular activities. Participation in these activities is deemed to act as a signal of students’ social and cultural backgrounds and reinforce their conformity to the social and cultural norms that the school system tends to reward. Participation in extracurricular activities as such would thus contribute to the social closure and reproduction of the upper classes. Here, the impact of the social status and cultural resources of a student’s family on his or her academic performance is expected to be mediated by participation in extracurricular activities. H5: Participation in extracurricular activities mediates the impact of students’ inherited social status and family-transmitted cultural capital. Method and analytical strategy After standardization of the sixth-grade tests and BNC exam marks, the three sets of indicators for educational outcomes and skills will be treated as dependent variables in regression models measured twice, in 2008 and in 2011. The same holds for the indicators related to students’ involvement in extracurricular activities, also measured in 2008 and 2011. Some of the other characteristics of students and of their environment that deserve to be controlled for in the models because they could affect both the likelihood of students’ engagement in extracurricular activities and their academic results can also be regarded as potentially time-varying, for example students’ parents’ income and the number of their siblings. Other characteristics can be considered as intrinsically time-invariant. Some, such as gender, parents’ level of education, or social class, are observable and measurable. Others are not (unobserved heterogeneity). However, the panel structure of the data allows for the implementation of fixed-effects (FE) regression models that make the control of all these time-invariant characteristics possible, even when unobservable (Allison 2009). FE regression is based on the estimation of the impact of within variations, that is, the impact of a variation in time in an independent variable on the variation in time in a specific outcome for the same respondent, rather than between variations, that is, the variations observed in both dependent and independent variables between different respondents at the same time. In FE regression, all the time-invariant characteristics of the respondents, even when unmeasured, can thus be controlled for by using each individual as his or her own control. Accordingly, in order to estimate whether or not participation in extracurricular activities affects educational outcomes or skills, this article compares the value of these outcomes in 2008 and 2011 while controlling for variation in the extracurricular activities indicators and other relevant time-varying variables. Assuming that students’ other characteristics all remain constant in the interval, the effect estimate of participation in extracurricular activities can be considered as unbiased, that is, it cannot be attributed to confounding factors or omitted time-invariant variables.5 To put it more formally, let us consider the dependent variable Yit (Mathematics or French scores, cognitive or conative test scores), the time-varying predictors Xit (including indicators of participation in extracurricular activities), and the invariant predictors Zi. The basic model for estimating the impact of Xit on Yit can be written as follows:   yit=μt+βXit+γZi+αi+εitwhere μt is a period-specific intercept, β and γ are vectors of coefficients, αi is an error term attached to each individual, and εit is an error term attached to each individual at each point in time. As all the variables under consideration are observed at only two periods (T = 2), the model can be reformulated as   yi1=μ1+βXi1+γZi+αi+εi1and   yi2=μ2+βXi2+γZi+αi+εi2 Then, averaging the two preceding equations, we get   y̅i=μ̅+βX̅i+γZi+αi+ε̅iand subtracting the first equation from the second   yit−y̅i=(μt−μ̅)+β(Xit−X̅i)+(εit−ε̅i),which can be rewritten as   ÿit=μ̈t+βẌit+ε̈it,where the unobserved fixed effects αi as well as the γ coefficients for time-invariant characteristics (between effects, i.e., time-invariant effects but which vary from one individual to another) have disappeared, allowing for unbiased estimation of β coefficients (within effects, i.e., time-varying effects “within” each individual). To test hypotheses 1 to 4, the impact of participation in extracurricular activities on school outcomes is estimated first by means of OLS regressions performed separately on cross-sectional data in 2008 and 2011. These estimates are then compared to the estimates produced by FE regressions of the same outcomes on the same variables performed on panel data. Hybrid models are used to test hypothesis 5. Hybrid models are a class of models that overcome the inability of fixed-effects models to estimate the effect of time-invariant variables. One immediate alternative to FE models that does not suffer from the same disadvantage is the random-effects (RE) model, in which time-invariant independent variables can be included. The main difference between RE and FE models is that RE models assume that the time-invariant unobserved characteristics of the individuals, captured by the abovementioned αi error term, are not correlated to the predictors included in the model. This hypothesis is rarely verified, though, and the time-varying variables’ coefficients are thus biased. Most of the time, FE models must be preferred for this reason.6 To circumvent this problem, it has been proposed to make use of a hybrid form of the RE model that shares the ability of FE models to estimate unbiased coefficients for time-varying variables (Allison 2009; Wooldridge 2010; Schunck 2013). This is made possible by the decomposition of the time-varying effects ( β coefficients) into a between (X̅i) and a within component (Xit−X̅i). This hybrid model (Allison 2009) can be written as follows:   yit=μt+β(Xit−X̅i)+γZi+θ(X̅i)+αi+εit Interestingly, the β coefficients estimated by this hybrid model are the same as those estimated by the FE model, and they are seemingly unbiased. In addition, as this model is a random-effects model, time-invariant effects (Zi) can also be estimated.7 Results The social factors of participation in extracurricular activities Participation in extracurricular activities is unevenly frequent among students. As shown in tables 1a and 1b, in 2008 as well as in 2011, the average number of activities in which students participate is positively and significantly correlated with their parents’ status, social class, level of education, and income. Similarly, differences in rates of participation in almost all activities taken separately are significantly associated with these characteristics. The only exception is participation in a sports association, for which differences associated with social status and income are not significant in 2011. Moreover, the social disparities are maximal when it comes to truly extracurricular activities (participation in a sports club, in a music academy, in a drama class) and minimal when it comes to extracurricular activities that take place at school (participation in a sports association or in an activity club other than sports in the school context). In addition, the activities for which the differences in participation according to parents’ economic, educational, and social resources are the highest are also the most emblematic of the realm of highbrow culture (enrollment in a library, in a music academy, or in a drama class). In all probability, the unequal participation of children from unequal social backgrounds does not necessarily reflect an unequal desire to participate. Rather, prior studies suggest that these social differences come mainly from an unequal access to the material and cultural resources that condition participation (Chin and Phillips 2004). Table 1a. Descriptive statistics (2008) 2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Table 1a. Descriptive statistics (2008) 2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Table 1b. Descriptive statistics (2011) 2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Table 1b. Descriptive statistics (2011) 2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Besides, comparison between tables 1a and 1b displays a substantial reduction in the average number of extracurricular activities participated in by students between 2008 and 2011. The relative frequency of each of the activities taken separately follows the same tendency, which is coherent with a global trend usually observed at these ages, at least in the French context. Prior studies have shown that early adolescence (between 10 and 15 years of age) often corresponds to a decline in reading (Baudelot, Cartier, and Détrez 1999) as well as a relative withdrawal from many other cultural activities (Octobre et al. 2010). This trend generally corresponds to the growing importance of informal peer relationships when students get older, detrimental to other activities, and to the progressive empowerment of children, whose parents generally prescribe most of the extracurricular activities performed at younger ages (ibid.). Participation in extracurricular activities and school achievement Table 2 displays the results of two OLS and FE regression models of the marks obtained in French and Mathematics on the scale of participation in extracurricular activities, ranging from 0 to 8. A first OLS regression model is estimated on two separate cross-sectional datasets, one in 2008 and the other in 2011, where the impact of participation in activities is controlled for gender, parents’ social status, parents’ income, parents’ education, and number of siblings.8 Marks, social status, education, and income are standardized (H1a). A second OLS model adds a quadratic term for the scale of participation in extracurricular activities, in order to test for the hypothesis of a non-linear effect (H1b). Two FE regression models that include the same variables as the two previous OLS regressions, except for the time-invariant variables (parents’ education and status), are then tested. As before, the second FE model tests for the presence of a non-linear effect (H1b). Both FE models add several time-varying controls that could potentially act as confounding factors. These include change of school between the two dates, change of area of residence, serious illness, death of father, mother, brother, or sister, and loss of employment by father or mother. Table 2. OLS and fixed-effects regressions of Mathematics and French marks   French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005    French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. Table 2. OLS and fixed-effects regressions of Mathematics and French marks   French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005    French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. In the first OLS regression model, the estimated impact of the number of activities is globally significant for Mathematics and French marks, in 2008 as well as in 2011. In the first FE model, the impact of the number of activities remains significant. In comparison with the OLS regression, however, this impact is two to three times lower. The second OLS model depicts a non-linear effect of participation in extracurricular activities in 2008 and 2011, for marks in both French and Mathematics. However, this non-linear effect is only partially confirmed by FE models. The coefficient for the quadratic term of the scale of participation in extracurricular activities is of the expected sign for both marks in French and Mathematics, but is only significant for the latter, with a turning point just below 4 (3.9). As 95 percent of the sample display a maximum score of 4 on the scale of participation in extracurricular activities, this inversion of the effect concerns only a marginal fraction of the pupils who are particularly involved in these activities. Overall, these results are partially supportive of hypothesis 1. Involvement in extracurricular activities is positively and significantly related to school achievement indicators such as marks in French and Mathematics (H1a), but the strength of this impact seems highly overestimated by OLS regression. Part of the effect seemingly associated with participation in extracurricular activities is in fact due to unobserved heterogeneity between respondents that may cause both participation in these activities and school success. Besides this, the non-linear effect of participation in extracurricular activities (H1b) is only supported for Mathematics marks. Finally, relative to the value of the standard deviations of marks in French and Mathematics, the addition of one activity is associated with a relatively modest gain of 0.8 points (on a scale of 0 to 100) in French and of 1.3 points in Mathematics. In other words, the difference between a student who does not participate in any activity and a student who participates in 4 would be of 3.3 points in French and, taking into account a possible non-linear effect, of 5.3 points in Mathematics (on a scale from 0 to 100). The varying impact of participation according to the nature of extracurricular activities The FE regression model presented in table 2 was then rerun with the participation scale replaced by the eight dummy variables from which it derives. The parameter estimates of the impact of participation in these eight activities on marks in French and Mathematics are summarized in figure 1. In the graph, the estimated values of the regression coefficients of the eight activities are reported together with their 95 percent confidence interval. Figure 1. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on marks in French and Mathematics (with 95 percent CI) Control variables: Parents’ income, level of education, number of siblings, change of school between 2008 and 2011l, change of place of residence, father’s job loss, mother’s job loss, serious illness, father’s death, mother’s death, sibling’s death. Source: MENESR-DEPP, panel 2007. Figure 1. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on marks in French and Mathematics (with 95 percent CI) Control variables: Parents’ income, level of education, number of siblings, change of school between 2008 and 2011l, change of place of residence, father’s job loss, mother’s job loss, serious illness, father’s death, mother’s death, sibling’s death. Source: MENESR-DEPP, panel 2007. The reported estimates range from −0.03 (impact of enrollment in a youth center on mark in Mathematics) to 0.07 (impact of participation in an activity club at school on mark in French and impact of enrollment in a library on mark in Mathematics). Not all these effects are significant. The only activities whose impact is significant for both French and Mathematics marks are enrollment in a library, enrollment in a music academy or school of music, and participation in an activity club at school (excluding sports activities). The impact of participation in a sports club is significant in regard to Mathematics marks only, whereas participation in a sports association at school is significant for French marks only. When compared to the value of standard deviations in French and Mathematics marks, these effects appear relatively modest. Concerning marks in Mathematics, these range from an extra 0.6 points (on a scale of 0 to 100) for the impact of participation in a sports club to an extra 1.5 points for the impact of enrollment in a library. Concerning French marks, these range from an extra 0.5 points for enrollment in a music academy or in the sports association at school to an extra 1.2 points for participation in an activity club at school. However, the computation of confidence intervals reproduced in figure 1 suggests that most of the differences observed in the magnitude of the impact of the various activities are not statistically significant. Overall these results give ambivalent support to the hypothesis that the various extracurricular activities unevenly affect school outcomes (H2). Several of the extracurricular activities under consideration do not significantly impact school outcomes, but those with significant impact do not significantly differ from each other in regard to the strength of their impact. The impact of extracurricular activities on cognitive and conative skills Table 3 displays the estimates of the impact of extracurricular activities on the scores obtained by the student panelists on the cognitive and non-cognitive tests included in the surveys of 2008 and 2011. The first indicator corresponds to the global score on cognitive tests (COG), followed by three indicators that correspond to the scores on conative tests of the academic dimension (PESCH), the social dimension (PESOC), and the self-regulation dimension (PESELF) of the perceived efficiency of students. For each of the four variables, the first two columns correspond to OLS regressions on cross-sectional data (2008 and 2011) and the third column to an FE model performed on panel data. Independent variables are the same as in the previous models for French and Mathematics marks. The test for a non-linear effect, which proved to be insignificant in each case, is not reproduced in the table. Table 3. OLS and fixed-effects regressions of cognitive and conative test scores   Cog  Pesoc  Pesch  Peself  OLS  FE  OLS  FE  OLS  FE  OLS  FE  2008  2011    2008  2011    2008  2011    2008  2011    Extracurricular activities  0.095***  0.107***  −0.002  0.074***  0.095***  0.031***  0.069***  0.087***  0.018**  0.045***  0.042***  0.033***    (0.006)  (0.006)  (0.003)  (0.006)  (0.006)  (0.006)  (0.007)  (0.006)  (0.006)  (0.006)  (0.006)  (0.007)  Gender  −0.022  −0.011    −0.212***  −0.402***    0.269***  0.228***    0.043**  0.039**      (0.013)  (0.013)    (0.015)  (0.014)    (0.015)  (0.015)    (0.015)  (0.014)    Parents' income  0.057***  0.043*  −0.000  −0.008  0.006  −0.020**  0.005  0.007    0.019**  −0.000  0.021*    (0.009)  (0.018)  (0.003)  (0.008)  (0.007)  (0.007)  (0.008)  (0.008)    (0.006)  (0.006)  (0.009)  Siblings  −0.057***  −0.068***  0.007  0.005  0.015*  −0.017  −0.017*  −0.007    −0.052***  0.003  0.006    (0.006)  (0.006)  (0.006)  (0.007)  (0.007)  (0.013)  (0.007)  (0.007)    (0.008)  (0.007)  (0.015)  Parents' education  0.254***  0.278***    −0.005  −0.030**    0.079***  0.094***    0.108***  0.028**      (0.009)  (0.009)    (0.011)  (0.010)    (0.010)  (0.010)    (0.010)  (0.010)    Parents' status  0.105***  0.121***    0.028**  0.001    0.034***  0.049***    0.054***  0.016      (0.008)  (0.009)    (0.009)  (0.009)    (0.010)  (0.009)    (0.009)  (0.009)    Change of school      −0.034**      0.030      −0.011      −0.012        (0.012)      (0.027)      (0.027)      (0.031)  Change of place of residence      0.080**      −0.086      −0.012      0.046        (0.029)      (0.061)      (0.061)      (0.068)  Father’s job loss      −0.005      −0.002      −0.017      −0.007        (0.011)      (0.024)      (0.023)      (0.027)  Mother’s job loss      −0.011      0.010      −0.003      0.029        (0.011)      (0.022)      (0.022)      (0.025)  Serious illness      −0.009      −0.075      0.010      0.034        (0.028)      (0.054)      (0.060)      (0.073)  Father’s death      −0.037      0.032      −0.072      0.106        (0.040)      (0.084)      (0.098)      (0.123)  Mother’s death      −0.004      −0.110      0.016      0.201        (0.066)      (0.117)      (0.128)      (0.156)  Sibling’s death      0.049      0.251*      −0.016      0.074        (0.065)      (0.105)      (0.111)      (0.142)  Intercept  0.001  0.031  0.105***  0.185***  0.455***  −0.019  −0.489***  −0.441***  0.031  −0.044  −0.096***  −0.049    (0.025)  (0.024)  (0.015)  (0.030)  (0.027)  (0.032)  (0.029)  (0.027)  (0.031)  (0.028)  (0.027)  (0.037)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.190  0.218  0.013  0.023  0.055  0.006  0.047  0.049  0.007  0.040  0.006  0.006    Cog  Pesoc  Pesch  Peself  OLS  FE  OLS  FE  OLS  FE  OLS  FE  2008  2011    2008  2011    2008  2011    2008  2011    Extracurricular activities  0.095***  0.107***  −0.002  0.074***  0.095***  0.031***  0.069***  0.087***  0.018**  0.045***  0.042***  0.033***    (0.006)  (0.006)  (0.003)  (0.006)  (0.006)  (0.006)  (0.007)  (0.006)  (0.006)  (0.006)  (0.006)  (0.007)  Gender  −0.022  −0.011    −0.212***  −0.402***    0.269***  0.228***    0.043**  0.039**      (0.013)  (0.013)    (0.015)  (0.014)    (0.015)  (0.015)    (0.015)  (0.014)    Parents' income  0.057***  0.043*  −0.000  −0.008  0.006  −0.020**  0.005  0.007    0.019**  −0.000  0.021*    (0.009)  (0.018)  (0.003)  (0.008)  (0.007)  (0.007)  (0.008)  (0.008)    (0.006)  (0.006)  (0.009)  Siblings  −0.057***  −0.068***  0.007  0.005  0.015*  −0.017  −0.017*  −0.007    −0.052***  0.003  0.006    (0.006)  (0.006)  (0.006)  (0.007)  (0.007)  (0.013)  (0.007)  (0.007)    (0.008)  (0.007)  (0.015)  Parents' education  0.254***  0.278***    −0.005  −0.030**    0.079***  0.094***    0.108***  0.028**      (0.009)  (0.009)    (0.011)  (0.010)    (0.010)  (0.010)    (0.010)  (0.010)    Parents' status  0.105***  0.121***    0.028**  0.001    0.034***  0.049***    0.054***  0.016      (0.008)  (0.009)    (0.009)  (0.009)    (0.010)  (0.009)    (0.009)  (0.009)    Change of school      −0.034**      0.030      −0.011      −0.012        (0.012)      (0.027)      (0.027)      (0.031)  Change of place of residence      0.080**      −0.086      −0.012      0.046        (0.029)      (0.061)      (0.061)      (0.068)  Father’s job loss      −0.005      −0.002      −0.017      −0.007        (0.011)      (0.024)      (0.023)      (0.027)  Mother’s job loss      −0.011      0.010      −0.003      0.029        (0.011)      (0.022)      (0.022)      (0.025)  Serious illness      −0.009      −0.075      0.010      0.034        (0.028)      (0.054)      (0.060)      (0.073)  Father’s death      −0.037      0.032      −0.072      0.106        (0.040)      (0.084)      (0.098)      (0.123)  Mother’s death      −0.004      −0.110      0.016      0.201        (0.066)      (0.117)      (0.128)      (0.156)  Sibling’s death      0.049      0.251*      −0.016      0.074        (0.065)      (0.105)      (0.111)      (0.142)  Intercept  0.001  0.031  0.105***  0.185***  0.455***  −0.019  −0.489***  −0.441***  0.031  −0.044  −0.096***  −0.049    (0.025)  (0.024)  (0.015)  (0.030)  (0.027)  (0.032)  (0.029)  (0.027)  (0.031)  (0.028)  (0.027)  (0.037)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.190  0.218  0.013  0.023  0.055  0.006  0.047  0.049  0.007  0.040  0.006  0.006  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. Table 3. OLS and fixed-effects regressions of cognitive and conative test scores   Cog  Pesoc  Pesch  Peself  OLS  FE  OLS  FE  OLS  FE  OLS  FE  2008  2011    2008  2011    2008  2011    2008  2011    Extracurricular activities  0.095***  0.107***  −0.002  0.074***  0.095***  0.031***  0.069***  0.087***  0.018**  0.045***  0.042***  0.033***    (0.006)  (0.006)  (0.003)  (0.006)  (0.006)  (0.006)  (0.007)  (0.006)  (0.006)  (0.006)  (0.006)  (0.007)  Gender  −0.022  −0.011    −0.212***  −0.402***    0.269***  0.228***    0.043**  0.039**      (0.013)  (0.013)    (0.015)  (0.014)    (0.015)  (0.015)    (0.015)  (0.014)    Parents' income  0.057***  0.043*  −0.000  −0.008  0.006  −0.020**  0.005  0.007    0.019**  −0.000  0.021*    (0.009)  (0.018)  (0.003)  (0.008)  (0.007)  (0.007)  (0.008)  (0.008)    (0.006)  (0.006)  (0.009)  Siblings  −0.057***  −0.068***  0.007  0.005  0.015*  −0.017  −0.017*  −0.007    −0.052***  0.003  0.006    (0.006)  (0.006)  (0.006)  (0.007)  (0.007)  (0.013)  (0.007)  (0.007)    (0.008)  (0.007)  (0.015)  Parents' education  0.254***  0.278***    −0.005  −0.030**    0.079***  0.094***    0.108***  0.028**      (0.009)  (0.009)    (0.011)  (0.010)    (0.010)  (0.010)    (0.010)  (0.010)    Parents' status  0.105***  0.121***    0.028**  0.001    0.034***  0.049***    0.054***  0.016      (0.008)  (0.009)    (0.009)  (0.009)    (0.010)  (0.009)    (0.009)  (0.009)    Change of school      −0.034**      0.030      −0.011      −0.012        (0.012)      (0.027)      (0.027)      (0.031)  Change of place of residence      0.080**      −0.086      −0.012      0.046        (0.029)      (0.061)      (0.061)      (0.068)  Father’s job loss      −0.005      −0.002      −0.017      −0.007        (0.011)      (0.024)      (0.023)      (0.027)  Mother’s job loss      −0.011      0.010      −0.003      0.029        (0.011)      (0.022)      (0.022)      (0.025)  Serious illness      −0.009      −0.075      0.010      0.034        (0.028)      (0.054)      (0.060)      (0.073)  Father’s death      −0.037      0.032      −0.072      0.106        (0.040)      (0.084)      (0.098)      (0.123)  Mother’s death      −0.004      −0.110      0.016      0.201        (0.066)      (0.117)      (0.128)      (0.156)  Sibling’s death      0.049      0.251*      −0.016      0.074        (0.065)      (0.105)      (0.111)      (0.142)  Intercept  0.001  0.031  0.105***  0.185***  0.455***  −0.019  −0.489***  −0.441***  0.031  −0.044  −0.096***  −0.049    (0.025)  (0.024)  (0.015)  (0.030)  (0.027)  (0.032)  (0.029)  (0.027)  (0.031)  (0.028)  (0.027)  (0.037)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.190  0.218  0.013  0.023  0.055  0.006  0.047  0.049  0.007  0.040  0.006  0.006    Cog  Pesoc  Pesch  Peself  OLS  FE  OLS  FE  OLS  FE  OLS  FE  2008  2011    2008  2011    2008  2011    2008  2011    Extracurricular activities  0.095***  0.107***  −0.002  0.074***  0.095***  0.031***  0.069***  0.087***  0.018**  0.045***  0.042***  0.033***    (0.006)  (0.006)  (0.003)  (0.006)  (0.006)  (0.006)  (0.007)  (0.006)  (0.006)  (0.006)  (0.006)  (0.007)  Gender  −0.022  −0.011    −0.212***  −0.402***    0.269***  0.228***    0.043**  0.039**      (0.013)  (0.013)    (0.015)  (0.014)    (0.015)  (0.015)    (0.015)  (0.014)    Parents' income  0.057***  0.043*  −0.000  −0.008  0.006  −0.020**  0.005  0.007    0.019**  −0.000  0.021*    (0.009)  (0.018)  (0.003)  (0.008)  (0.007)  (0.007)  (0.008)  (0.008)    (0.006)  (0.006)  (0.009)  Siblings  −0.057***  −0.068***  0.007  0.005  0.015*  −0.017  −0.017*  −0.007    −0.052***  0.003  0.006    (0.006)  (0.006)  (0.006)  (0.007)  (0.007)  (0.013)  (0.007)  (0.007)    (0.008)  (0.007)  (0.015)  Parents' education  0.254***  0.278***    −0.005  −0.030**    0.079***  0.094***    0.108***  0.028**      (0.009)  (0.009)    (0.011)  (0.010)    (0.010)  (0.010)    (0.010)  (0.010)    Parents' status  0.105***  0.121***    0.028**  0.001    0.034***  0.049***    0.054***  0.016      (0.008)  (0.009)    (0.009)  (0.009)    (0.010)  (0.009)    (0.009)  (0.009)    Change of school      −0.034**      0.030      −0.011      −0.012        (0.012)      (0.027)      (0.027)      (0.031)  Change of place of residence      0.080**      −0.086      −0.012      0.046        (0.029)      (0.061)      (0.061)      (0.068)  Father’s job loss      −0.005      −0.002      −0.017      −0.007        (0.011)      (0.024)      (0.023)      (0.027)  Mother’s job loss      −0.011      0.010      −0.003      0.029        (0.011)      (0.022)      (0.022)      (0.025)  Serious illness      −0.009      −0.075      0.010      0.034        (0.028)      (0.054)      (0.060)      (0.073)  Father’s death      −0.037      0.032      −0.072      0.106        (0.040)      (0.084)      (0.098)      (0.123)  Mother’s death      −0.004      −0.110      0.016      0.201        (0.066)      (0.117)      (0.128)      (0.156)  Sibling’s death      0.049      0.251*      −0.016      0.074        (0.065)      (0.105)      (0.111)      (0.142)  Intercept  0.001  0.031  0.105***  0.185***  0.455***  −0.019  −0.489***  −0.441***  0.031  −0.044  −0.096***  −0.049    (0.025)  (0.024)  (0.015)  (0.030)  (0.027)  (0.032)  (0.029)  (0.027)  (0.031)  (0.028)  (0.027)  (0.037)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.190  0.218  0.013  0.023  0.055  0.006  0.047  0.049  0.007  0.040  0.006  0.006  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. For each of the indicators under consideration, the OLS regressions display significant and positive impact of participation in extracurricular activities. But when it comes to the FE models, this impact is no longer significant for the score obtained on the cognitive test (COG). The significant impact observed in OLS regressions might thus be entirely due to unobservable characteristics of the respondents that affect both cognitive test scores and participation in extracurricular activities. By contrast, the impact of participation in extracurricular activities on non-cognitive test scores remains highly significant in FE models, although reduced. This impact, which ranges from 0.018 to 0.033, remains quite substantial when compared to the value of the FE models’ intercepts. Overall, the contrast between the results observed for the scores obtained on cognitive and non-cognitive tests is highly supportive of the hypothesis that participation in extracurricular activities has a greater impact on non-cognitive than on cognitive skills (H3). Detailed analysis of the impact of each activity taken separately conveys interesting nuances, however. As shown in figure 2, none has any positive impact on cognitive test scores, and participation in an activity club at school even appears to have a slightly negative impact. More interesting are the disparities displayed as to the varying impact of the eight activities on the conative test scores, which are particularly pronounced in respect to the perceived social efficiency score (PESOC). Here, a clear contrast appears between the positive impact of enrollment in drama classes, participation in a sports club, a sports association, or an activity club, and the negative impact of enrollment in a music academy. In addition, the most culturally legitimate activities, with maximal impact on school outcomes, such as enrollment in a library or in a music academy, are not the most beneficial in this context. Figure 2. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on cognitive and conative tests scores (with 95 percent CI) Control variables: Parents’ income, level of education, number of siblings, change of school between 2008 and 2011l, change of place of residence, father’s job loss, mother’s job loss, serious illness, father’s death, mother’s death, sibling’s death. Source: MENESR-DEPP, panel 2007. Figure 2. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on cognitive and conative tests scores (with 95 percent CI) Control variables: Parents’ income, level of education, number of siblings, change of school between 2008 and 2011l, change of place of residence, father’s job loss, mother’s job loss, serious illness, father’s death, mother’s death, sibling’s death. Source: MENESR-DEPP, panel 2007. In the next step of the analysis, the hypothesis that the impact of extracurricular activities on school achievement would be mediated by the development of cognitive and non-cognitive skills implicitly required for academic success has been tested. To test this hypothesis, marks in French and Mathematics are first regressed on the number of extracurricular activities. A second model adds the global cognitive test score, and a third the three non-cognitive test scores to the variables on the right side of the equation. A substantial reduction of the direct effect of participation in extracurricular activities when controlled for scores on cognitive and non-cognitive skills would advocate for a mediation mechanism. Estimates reproduced in table 4 display the variation in the effect of participation in extracurricular activities on school outcomes when the three nested models are estimated by means of FE regressions on panel data. As shown in table 4, the impact of participation in extracurricular activities slightly decreases from model 1 to model 2 and from model 2 to model 3 when estimated by means of FE regressions. But the values of the coefficients estimated by models 2 and 3 remain quite close to those estimated by model 1. Table 4. Fixed-effects regressions of Mathematics and French marks, controlling by cognitive and conative test scores   French  Mathematics    Model 1  Model 2  Model 3  Model 1  Model 2  Model 3  Extracurricular activities  0.050***  0.047***  0.046***  0.072***  0.068***  0.067***    (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  Extracurricular activities2  −0.003  −0.002  −0.002  −0.009***  −0.009***  −0.009***    (0.003)  (0.003)  (0.003)  (0.003)  (0.003)  (0.002)  Parents’ income  −0.007  −0.007  −0.007  −0.005  −0.005  −0.006    (0.005)  (0.005)  (0.005)  (0.007)  (0.006)  (0.006)  Siblings  0.024*  0.023*  0.023*  0.041***  0.040***  0.039***    (0.011)  (0.010)  (0.010)  (0.010)  (0.010)  (0.010)  Cognitive score (cog)    0.168***  0.146***    0.188***  0.155***      (0.013)  (0.013)    (0.014)  (0.014)  Perceived social efficiency (pesoc)      −0.045***      −0.078***        (0.007)      (0.007)  Perceived school efficiency (pesch)      0.090***      0.143***        (0.008)      (0.008)  Perceived efficiency in self-regulation (peself)      0.027***      0.039***        (0.006)      (0.006)  Intercept  −0.087***  −0.100***  −0.098***  −0.114***  −0.129***  −0.126***    (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  N  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.012  0.550  0.572  0.007  0.591  0.498    French  Mathematics    Model 1  Model 2  Model 3  Model 1  Model 2  Model 3  Extracurricular activities  0.050***  0.047***  0.046***  0.072***  0.068***  0.067***    (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  Extracurricular activities2  −0.003  −0.002  −0.002  −0.009***  −0.009***  −0.009***    (0.003)  (0.003)  (0.003)  (0.003)  (0.003)  (0.002)  Parents’ income  −0.007  −0.007  −0.007  −0.005  −0.005  −0.006    (0.005)  (0.005)  (0.005)  (0.007)  (0.006)  (0.006)  Siblings  0.024*  0.023*  0.023*  0.041***  0.040***  0.039***    (0.011)  (0.010)  (0.010)  (0.010)  (0.010)  (0.010)  Cognitive score (cog)    0.168***  0.146***    0.188***  0.155***      (0.013)  (0.013)    (0.014)  (0.014)  Perceived social efficiency (pesoc)      −0.045***      −0.078***        (0.007)      (0.007)  Perceived school efficiency (pesch)      0.090***      0.143***        (0.008)      (0.008)  Perceived efficiency in self-regulation (peself)      0.027***      0.039***        (0.006)      (0.006)  Intercept  −0.087***  −0.100***  −0.098***  −0.114***  −0.129***  −0.126***    (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  N  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.012  0.550  0.572  0.007  0.591  0.498  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. Table 4. Fixed-effects regressions of Mathematics and French marks, controlling by cognitive and conative test scores   French  Mathematics    Model 1  Model 2  Model 3  Model 1  Model 2  Model 3  Extracurricular activities  0.050***  0.047***  0.046***  0.072***  0.068***  0.067***    (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  Extracurricular activities2  −0.003  −0.002  −0.002  −0.009***  −0.009***  −0.009***    (0.003)  (0.003)  (0.003)  (0.003)  (0.003)  (0.002)  Parents’ income  −0.007  −0.007  −0.007  −0.005  −0.005  −0.006    (0.005)  (0.005)  (0.005)  (0.007)  (0.006)  (0.006)  Siblings  0.024*  0.023*  0.023*  0.041***  0.040***  0.039***    (0.011)  (0.010)  (0.010)  (0.010)  (0.010)  (0.010)  Cognitive score (cog)    0.168***  0.146***    0.188***  0.155***      (0.013)  (0.013)    (0.014)  (0.014)  Perceived social efficiency (pesoc)      −0.045***      −0.078***        (0.007)      (0.007)  Perceived school efficiency (pesch)      0.090***      0.143***        (0.008)      (0.008)  Perceived efficiency in self-regulation (peself)      0.027***      0.039***        (0.006)      (0.006)  Intercept  −0.087***  −0.100***  −0.098***  −0.114***  −0.129***  −0.126***    (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  N  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.012  0.550  0.572  0.007  0.591  0.498    French  Mathematics    Model 1  Model 2  Model 3  Model 1  Model 2  Model 3  Extracurricular activities  0.050***  0.047***  0.046***  0.072***  0.068***  0.067***    (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  (0.012)  Extracurricular activities2  −0.003  −0.002  −0.002  −0.009***  −0.009***  −0.009***    (0.003)  (0.003)  (0.003)  (0.003)  (0.003)  (0.002)  Parents’ income  −0.007  −0.007  −0.007  −0.005  −0.005  −0.006    (0.005)  (0.005)  (0.005)  (0.007)  (0.006)  (0.006)  Siblings  0.024*  0.023*  0.023*  0.041***  0.040***  0.039***    (0.011)  (0.010)  (0.010)  (0.010)  (0.010)  (0.010)  Cognitive score (cog)    0.168***  0.146***    0.188***  0.155***      (0.013)  (0.013)    (0.014)  (0.014)  Perceived social efficiency (pesoc)      −0.045***      −0.078***        (0.007)      (0.007)  Perceived school efficiency (pesch)      0.090***      0.143***        (0.008)      (0.008)  Perceived efficiency in self-regulation (peself)      0.027***      0.039***        (0.006)      (0.006)  Intercept  −0.087***  −0.100***  −0.098***  −0.114***  −0.129***  −0.126***    (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  (0.018)  N  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.012  0.550  0.572  0.007  0.591  0.498  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. As shown in figure 3, the results are the same when the scale variable of participation in extracurricular activities is replaced by the eight activities taken separately. Their impact on marks in both French and Mathematics remains relatively unchanged when controlled for the students’ scores on the cognitive and non-cognitive tests. Figure 3. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on marks in French and Mathematics controlled by cognitive and conative test scores (with 95 percent CI) Cog: Cognitive test score Pesch: Perceived school efficiency test score Pesoc: Perceived social efficiency test score Peself: Perceived efficiency in self-regulation test score Source: MENESR-DEPP, panel 2007. Figure 3. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on marks in French and Mathematics controlled by cognitive and conative test scores (with 95 percent CI) Cog: Cognitive test score Pesch: Perceived school efficiency test score Pesoc: Perceived social efficiency test score Peself: Perceived efficiency in self-regulation test score Source: MENESR-DEPP, panel 2007. Together, these results lead to the rejection of the hypothesis that the impact of participation in extracurricular activities on school achievement would be predominantly mediated by its impact on cognitive and non-cognitive skills (H4), at least as these skills are measured in the 2007 panel data. Participation in extracurricular activities and the reproduction of privilege The focus henceforth will be on the mediating role that participation in extracurricular activities might play in the reproduction of social and cultural privilege. Two nested models are estimated. To overcome the inability of FE models to cope with the estimation of time-invariant characteristics such as parents’ level of education and social status, the models under consideration are hybrid models. A first model includes parents’ income, status, and level of education, as well as the respondents’ gender and the number of their siblings. A second model adds the number of extracurricular activities in which the student participates. Comparing the estimated values of the effect of parents’ status, income, and level of education in both models provides an overview of the extent to which participation in extracurricular activities mediates the effect of social and cultural assets inherited from parents. In the estimation of the hybrid models, time-varying effects are broken down into between and within components, but the interpretation is restricted to the latter. Between effects are consequently reported in italics in table 5. Table 5. Hybrid random-effects nested regressions of French and Mathematics marks on participation in extracurricular activities and social and cultural inherited assets   French  Mathematics    Model 1    Model 2    Model 1    Model 2      French  95% CI  French  95% CI  Maths  95% CI  Maths  95% CI  Gender  0.399***  [0.377,0.421]  0.401***  [0.379,0.422]  −0.155***  [−0.177,−0.133]  −0.153***  [−0.175,−0.131]    (0.011)    (0.011)    (0.011)    (0.011)    Parents' status  0.127***  [0.113,0.141]  0.116***  [0.102,0.129]  0.115***  [0.101,0.128]  0.103***  [0.089,0.116]    (0.007)    (0.007)    (0.007)    (0.007)    Parent's education  0.263***  [0.249,0.277]  0.235***  [0.221,0.250]  0.270***  [0.256,0.285]  0.242***  [0.227,0.256]    (0.007)    (0.007)    (0.007)    (0.007)    d_income  −0.001  [−0.011,0.009]  −0.002  [−0.012,0.008]  0.002  [−0.008,0.012]  0.001  [−0.009,0.012]    (0.005)    (0.005)    (0.005)    (0.005)    d_siblings  0.020*  [0.003,0.038]  0.017  [−0.001,0.034]  0.037***  [0.018,0.055]  0.034***  [0.015,0.052]    (0.009)    (0.009)    (0.009)    (0.009)    d_extracurricular activities      0.045***  [0.036,0.054]      0.040***  [0.031,0.049]        (0.005)        (0.005)    m_income  0.061***  [0.047,0.074]  0.058***  [0.045,0.072]  0.075***  [0.062,0.089]  0.073***  [0.059,0.086]    (0.007)    (0.007)    (0.007)    (0.007)    m_siblings  −0.063***  [−0.073,−0.052]  −0.068***  [−0.079,−0.058]  −0.045***  [−0.056,−0.034]  −0.051***  [−0.061,−0.040]    (0.005)    (0.005)    (0.005)    (0.005)    m_extracurricular activities      0.110**  0.120]      0.115**  [0.104,0.126]        (0.005)        (0.006)    Intercept  −0.512***  [−0.550,−0.474]  −0.699*  [−0.741,−0.657]  0.303***  [0.264,0.341]  0.107*  [0.064,0.149]    (0.020)    (0.021)    (0.020)    (0.022)    N  19,809    19,809    19,809    19,809    R2  0.232    0.247    0.203    0.221      French  Mathematics    Model 1    Model 2    Model 1    Model 2      French  95% CI  French  95% CI  Maths  95% CI  Maths  95% CI  Gender  0.399***  [0.377,0.421]  0.401***  [0.379,0.422]  −0.155***  [−0.177,−0.133]  −0.153***  [−0.175,−0.131]    (0.011)    (0.011)    (0.011)    (0.011)    Parents' status  0.127***  [0.113,0.141]  0.116***  [0.102,0.129]  0.115***  [0.101,0.128]  0.103***  [0.089,0.116]    (0.007)    (0.007)    (0.007)    (0.007)    Parent's education  0.263***  [0.249,0.277]  0.235***  [0.221,0.250]  0.270***  [0.256,0.285]  0.242***  [0.227,0.256]    (0.007)    (0.007)    (0.007)    (0.007)    d_income  −0.001  [−0.011,0.009]  −0.002  [−0.012,0.008]  0.002  [−0.008,0.012]  0.001  [−0.009,0.012]    (0.005)    (0.005)    (0.005)    (0.005)    d_siblings  0.020*  [0.003,0.038]  0.017  [−0.001,0.034]  0.037***  [0.018,0.055]  0.034***  [0.015,0.052]    (0.009)    (0.009)    (0.009)    (0.009)    d_extracurricular activities      0.045***  [0.036,0.054]      0.040***  [0.031,0.049]        (0.005)        (0.005)    m_income  0.061***  [0.047,0.074]  0.058***  [0.045,0.072]  0.075***  [0.062,0.089]  0.073***  [0.059,0.086]    (0.007)    (0.007)    (0.007)    (0.007)    m_siblings  −0.063***  [−0.073,−0.052]  −0.068***  [−0.079,−0.058]  −0.045***  [−0.056,−0.034]  −0.051***  [−0.061,−0.040]    (0.005)    (0.005)    (0.005)    (0.005)    m_extracurricular activities      0.110**  0.120]      0.115**  [0.104,0.126]        (0.005)        (0.006)    Intercept  −0.512***  [−0.550,−0.474]  −0.699*  [−0.741,−0.657]  0.303***  [0.264,0.341]  0.107*  [0.064,0.149]    (0.020)    (0.021)    (0.020)    (0.022)    N  19,809    19,809    19,809    19,809    R2  0.232    0.247    0.203    0.221    Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Note: “d_” subtracts refer to within effects and “m_” subtracts refer to between effects, reported in italics. Source: MENESR-DEPP, panel 2007. Table 5. Hybrid random-effects nested regressions of French and Mathematics marks on participation in extracurricular activities and social and cultural inherited assets   French  Mathematics    Model 1    Model 2    Model 1    Model 2      French  95% CI  French  95% CI  Maths  95% CI  Maths  95% CI  Gender  0.399***  [0.377,0.421]  0.401***  [0.379,0.422]  −0.155***  [−0.177,−0.133]  −0.153***  [−0.175,−0.131]    (0.011)    (0.011)    (0.011)    (0.011)    Parents' status  0.127***  [0.113,0.141]  0.116***  [0.102,0.129]  0.115***  [0.101,0.128]  0.103***  [0.089,0.116]    (0.007)    (0.007)    (0.007)    (0.007)    Parent's education  0.263***  [0.249,0.277]  0.235***  [0.221,0.250]  0.270***  [0.256,0.285]  0.242***  [0.227,0.256]    (0.007)    (0.007)    (0.007)    (0.007)    d_income  −0.001  [−0.011,0.009]  −0.002  [−0.012,0.008]  0.002  [−0.008,0.012]  0.001  [−0.009,0.012]    (0.005)    (0.005)    (0.005)    (0.005)    d_siblings  0.020*  [0.003,0.038]  0.017  [−0.001,0.034]  0.037***  [0.018,0.055]  0.034***  [0.015,0.052]    (0.009)    (0.009)    (0.009)    (0.009)    d_extracurricular activities      0.045***  [0.036,0.054]      0.040***  [0.031,0.049]        (0.005)        (0.005)    m_income  0.061***  [0.047,0.074]  0.058***  [0.045,0.072]  0.075***  [0.062,0.089]  0.073***  [0.059,0.086]    (0.007)    (0.007)    (0.007)    (0.007)    m_siblings  −0.063***  [−0.073,−0.052]  −0.068***  [−0.079,−0.058]  −0.045***  [−0.056,−0.034]  −0.051***  [−0.061,−0.040]    (0.005)    (0.005)    (0.005)    (0.005)    m_extracurricular activities      0.110**  0.120]      0.115**  [0.104,0.126]        (0.005)        (0.006)    Intercept  −0.512***  [−0.550,−0.474]  −0.699*  [−0.741,−0.657]  0.303***  [0.264,0.341]  0.107*  [0.064,0.149]    (0.020)    (0.021)    (0.020)    (0.022)    N  19,809    19,809    19,809    19,809    R2  0.232    0.247    0.203    0.221      French  Mathematics    Model 1    Model 2    Model 1    Model 2      French  95% CI  French  95% CI  Maths  95% CI  Maths  95% CI  Gender  0.399***  [0.377,0.421]  0.401***  [0.379,0.422]  −0.155***  [−0.177,−0.133]  −0.153***  [−0.175,−0.131]    (0.011)    (0.011)    (0.011)    (0.011)    Parents' status  0.127***  [0.113,0.141]  0.116***  [0.102,0.129]  0.115***  [0.101,0.128]  0.103***  [0.089,0.116]    (0.007)    (0.007)    (0.007)    (0.007)    Parent's education  0.263***  [0.249,0.277]  0.235***  [0.221,0.250]  0.270***  [0.256,0.285]  0.242***  [0.227,0.256]    (0.007)    (0.007)    (0.007)    (0.007)    d_income  −0.001  [−0.011,0.009]  −0.002  [−0.012,0.008]  0.002  [−0.008,0.012]  0.001  [−0.009,0.012]    (0.005)    (0.005)    (0.005)    (0.005)    d_siblings  0.020*  [0.003,0.038]  0.017  [−0.001,0.034]  0.037***  [0.018,0.055]  0.034***  [0.015,0.052]    (0.009)    (0.009)    (0.009)    (0.009)    d_extracurricular activities      0.045***  [0.036,0.054]      0.040***  [0.031,0.049]        (0.005)        (0.005)    m_income  0.061***  [0.047,0.074]  0.058***  [0.045,0.072]  0.075***  [0.062,0.089]  0.073***  [0.059,0.086]    (0.007)    (0.007)    (0.007)    (0.007)    m_siblings  −0.063***  [−0.073,−0.052]  −0.068***  [−0.079,−0.058]  −0.045***  [−0.056,−0.034]  −0.051***  [−0.061,−0.040]    (0.005)    (0.005)    (0.005)    (0.005)    m_extracurricular activities      0.110**  0.120]      0.115**  [0.104,0.126]        (0.005)        (0.006)    Intercept  −0.512***  [−0.550,−0.474]  −0.699*  [−0.741,−0.657]  0.303***  [0.264,0.341]  0.107*  [0.064,0.149]    (0.020)    (0.021)    (0.020)    (0.022)    N  19,809    19,809    19,809    19,809    R2  0.232    0.247    0.203    0.221    Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Note: “d_” subtracts refer to within effects and “m_” subtracts refer to between effects, reported in italics. Source: MENESR-DEPP, panel 2007. The parameter estimates displayed in table 5 show a slight decrease of the impact of parents’ social status on French and Mathematics marks from model 1 to model 2. Relatively small, the differences between the corresponding coefficients are also statistically insignificant, however, as suggested by the comparison of the 95 percent confidence intervals of the estimated coefficients reported in the table. The effect of income is remarkably stable and not significant from one model to another. Contrastingly, for both French and Mathematics marks, the comparison of the values and confidence intervals of the impact of parents’ education in the two models displays a more substantial and significant reduction from model 1 to model 2. This result suggests that participation in extracurricular activities mediates the impact of parents’ education on the academic success of their offspring. Overall, these results partially support hypothesis 5. The mediating effect of extracurricular activities on the transmission of parents’ educational resources suggests that the participation of children in these activities may reflect an educational investment on the part of their parents conditioned by their own educational resources. But the results observed with respect to the transmission of the effect of parents’ social status does not conform to the idea that participation in these activities would primarily act as a status signal rewarded as such by the school system. Discussion French middle school students’ extracurricular activities have a significant impact on their academic performance. All significant impacts are positive, although an inversion of the impact is depicted at extremely high levels of participation (more than 4 activities) in respect to marks in Mathematics. Overall, the hypothesis that participation in extracurricular activities might be negatively oriented toward school is not supported (Coleman 1961), at least for the age range considered in the analysis developed in this article. The addition of further observation points beyond the 2008 and 2011 surveys would allow for a better assessment of the long-term impact of participation in extracurricular activities. However, when estimated by means of FE models on panel data, these effects are two to three times weaker than when estimated by means of OLS regression on cross-sectional data. But the persistence of the impact of participation in extracurricular activities, even if this is reduced, when controlling for unobserved heterogeneity regarding students and their environment, is the advantage brought by FE modeling and is a guarantee of its robustness. Moreover, not all academic and skills-related outcomes are equally impacted by participation in extracurricular activities. Mathematics and French scores as well as the scores obtained at several conative tests are significantly though moderately improved, while scores at cognitive tests are almost unaffected. In addition, the impact may vary depending on the content of the activities, with even a few slightly significant negative effects (effect of participating in an activity club on cognitive score and effect of enrollment in a music academy on perceived social efficiency). This contrasting impact raises questions about the scope and limitations of three distinct mechanisms through which participation in these activities may be seen to positively affect school-related outcomes. The first mechanism would be the so-called “transfer paradigm” (Dettermann and Sternberg 1993), according to which skills and competencies (i.e., human capital) acquired in extracurricular activities might be transferred and implemented in other domains. Consistent with previous research on this issue, the results of the analysis conducted here do not demonstrate the cognitive component of this transfer mechanism. By contrast, participation in extracurricular activities is found to have a significant impact on some of the perceived efficiency scores obtained by the panel students. This finding gives some credit to the idea that the impact of participation in extracurricular activities might be due to the transfer of non-cognitive rather than cognitive skills. However, the FE regression models do not clearly demonstrate that the impact of participation in extracurricular activities on non-cognitive skills mediates its impact on French and Mathematics marks. Nonetheless, this lack of support for the transfer paradigm does not preclude the possibility that other cognitive or non-cognitive skills not measured in the data examined in this article may mediate the impact of participation in extracurricular activities on academic achievement. Further analyses would be necessary on that point. The second mechanism takes participation in extracurricular activities to be an accumulation of cultural capital, in Bourdieu’s sense of the term. According to cultural capital theory, participation in these activities acts as a signal of social and cultural complicity that is addressed by the students to their teachers and may bias teachers’ judgment. It relies on the idea that, consciously or not, teachers tend to reward students’ compliance with the norms of the dominant culture. Understood in this sense, however, this cultural capital-oriented interpretation suffers from serious flaws. First, whereas several of the activities that have a significant impact on school outcomes, such as enrollment in a library or in a music academy, belong to the realm of highbrow culture that might be valued by teachers, others do not, as is the case for participation in a sports club or in an activity club at school. Second, in order to bias teachers’ judgment, students’ participation in extracurricular activities must be to some extent reflected in their daily behavior at school, which is not evidently the case for most of the activities discussed here, which do not all take place in the school context. Moreover, the fact that the impact of participation in extracurricular activities is roughly equivalent for Mathematics and French marks does not seem fully compatible with this interpretation, as the cultural capital bias, if any, might be expected to be stronger for the latter than for the former. Third, it should be borne in mind that the marks in French and Mathematics on which participation in these activities has a net impact are not continuous assessment scores given to students by their professors throughout the school year, but marks obtained in standardized and anonymous tests and exams. One can easily imagine the extent to which cultural distance or proximity between students and teachers might affect their relationship, including teachers’ judgment of students’ performance in everyday school life. It is much more difficult to demonstrate that the cultural bias of teachers’ judgment and the cultural markers associated with student participation in extracurricular activities are strong enough to affect the assessment of such anonymous tests and exams. Finally, the mediating impact of extracurricular activities displayed in the latest series of models (table 5) provides ambivalent support for the theory of cultural capital. The mediating impact observed in relation to parents’ education fits relatively well with Bourdieu’s insight that cultural capital, as any form of capital, results from an accumulated labor (Bourdieu 1986), which might be illustrated by the time parents devote to supporting the extracurricular activities of their children. But the absence of such a mediating impact on the social status of parents does not support Bourdieu’s notion of cultural capital as an instrument of social reproduction and social closure through the “symbolic violence” of “cultural arbitrariness” (Bourdieu and Passeron [1970] 1977). Extracurricular activities do not seem as such to act as a signal of cultural conformity to the ruling class culture that the students send to their teachers. Considering both the number of activities in which students enroll and the specific impact attached to the various activities, a third process emerges, which is somewhat independent of their contents. Structured and supervised activities may exert a structuring impact on children’s ability to manage time scarcity, whereas those whose leisure is less structured and supervised may lack this kind of ability. The educational benefit of participation in extracurricular activities could therefore be based on a cumulative process. The more students participate in structured extracurricular activities, whatever these may be, the better their academic performance. Time spent with these activities may thus have a positive impact on school achievement to the extent that it mechanically reduces unstructured leisure time that is detrimental to school achievement. Such generic impact of involvement in time-consuming leisure activities recalls the “concerted cultivation”–oriented educational style theorized by Annette Lareau (2003). The prevalence of a mechanism of this kind would explain the relatively weak disparities between the various activities with regard to the magnitude of their impact on school outcomes. The highlighting of a non-linear effect, at least for marks in Mathematics, also suggests that parents’ ability to supervise and thus potentially moderate children’s extracurricular activities is at stake. This third mechanism deserves to be further investigated with more inclusive information on students’ leisure time, including both structured and non-structured leisure, activities strongly encouraged by the school and less legitimated practices (media, games, social networks, etc.). This would help better identify and measure the contrasting impact of adolescents’ structured and unstructured leisure time. The lack of information on time spent on activities is another limitation. Based on available data, the analysis tends to conflate individuals who may have unequal levels of participation. Moreover, this third mechanism is not exclusive of the other two. To the extent that participation in extracurricular activities involves the acquisition of time management skills, its structuring power can be understood from the point of view of human capital theory. Similarly, the uneven amount of time devoted to extracurricular activities can be interpreted in terms of cultural capital. That teachers positively value the traits exhibited by pupils encouraged by their parents to participate in extracurricular activities might mirror parents’ ability to impose in the school context standards of assessment that work to their advantage (Lareau and Weininger 2003). In addition, the effect of participation in extracurricular activities could also be interpreted from the perspective of its implications in terms of social capital. In fact, many existing studies on related topics interpret the impact of extracurricular activities as an effect of the increase in volume of interactions with other students involved in participation in these activities (Broh 2002; Hansen, Larson, and Dworkin 2003). Participation in extracurricular activities would thus indirectly promote students’ attachment to their group of peers and to the adults involved in the management of these activities, as well as to the school and its norms. Unfortunately, the data under consideration in the present study does not include adequate measurement of students’ social ties. However, the differential impact of extracurricular activities on students’ perceived efficiency in social relations, with a clear contrast between sports and cultural practices, argues for a need of further studies taking greater account of this perspective.9 Conclusion This article makes two new contributions to the study of the impact of participation in extracurricular activities on school achievement. First, relying on regression techniques designed for panel data, it demonstrates the robustness of the impact of extracurricular activities on school achievement. At the same time, it attests to this impact being weaker than is usually estimated by means of usual regression techniques on cross-sectional data. Second, it puts forth a deeper investigation of the process by which these activities impact school outcomes and leads to a reappraisal of the concept of cultural capital that includes Lareau’s approach to class differences in parents’ child-rearing styles. In that regard, it seems to be the case that what is mainly at stake in participation in extracurricular activities is the unequal capacity of families to extend the time of school supervision in their children’s free time. This is not to say that cultural capital biases, in the sense originally coined by Bourdieu and Passeron, might not exist during school life. According to the findings presented in this article, however, extracurricular activities do not appear to be the main place where these biases occur. Finally, although substantially modest, the impact of participation in extracurricular activities may have significant implications in terms of school inequalities, especially in the French context, where schools exert very limited control over activities mainly left to the discretion of families. In itself, unequal participation in voluntary extracurricular activities reinforces social inequalities in school achievement insofar as participation is strongly correlated with students’ social and cultural background. The extent to which this unequal investment in extracurricular activities can be mainly put down to material constraints (Chin and Phillips 2004) or class-based cultural orientations (Lareau et al. 2015) is beyond the scope of this article. To the extent that what is at stake is the management of adolescents’ time, the article suggests at least that reducing the unequal impact of participation in extracurricular activities would involve integrating them as much as possible into the school curriculum itself, to compensate for the disadvantage experienced by pupils from disadvantaged backgrounds. More generally, the results displayed in this article lead to reconsidering the social process by which child-rearing styles may affect school achievement and school inequalities. The school payoff of the concerted cultivation style theorized by Lareau is generally associated with the children’s sense of entitlement and self-confidence due to the inclination of their parents to encourage their language practices and their interactions with adults. Organized leisure activities are usually seen as a key medium through which this encouragement operates. Yet, while empirically established, this impact may have been a little overemphasized so far. Greater attention should therefore be paid to the diffuse impact of a set of everyday practices, verbal and non-verbal interactions between children and their parents. This calls at least for further research linking school performance with time budgets of children and families and for a broader assessment of the ways by which cultural capital is accumulated and transmitted during out-of-school time. More difficult to grasp, this complex set of practices and habits leads to the formation of social gaps between families that are also more difficult to compensate for through targeted policies aimed at encouraging a particular type of practice or activity. Notes 1 Parents were only asked to indicate whether their children were currently participating or not in each of the eight activities at the time of the survey. 2 This item is much more heterogeneous than the other seven. It can encompass various activities, such as parlor games, manual work, cookery, and so on. 3 For this subpopulation, the data provider (DEPP) calculated a weighting variable that has been used in all subsequent analyses. 4 All subsequent regression analyses have been performed twice: the first time with multiple imputation of missing values (using the “mi impute chained” command in Stata 14), the second time with restriction to complete cases. In all circumstances, the results proved to be extremely close (results on request). 5 Of course, one cannot rule out the confounding effect of other time-varying factors, which should also be controlled for as far as possible in the models. 6 The choice between RE and FE models can be decided by means of the Hausman test, which basically tests whether the unique errors αi are correlated with the regressors, the null hypothesis being that they are not. The test was performed for all subsequent analyses. In all cases, the null hypothesis was rejected, which was definitely in favor of FE models. 7 The inclusion of the between component of the time-varying effects (X̅i) simply ensures that effect estimates γ of time-invariant characteristics Zi included in the model are corrected for between differences in Xit (Schunck 2013). 8 The social status of parents is measured by a status scale based on an MDS analysis performed on the structure of friendship ties between social groups taken from the intermediate level of the French classification of socioprofessional categories (42 categories) (Cousteaux and Lemel, 2004). Education corresponds to the number of years of education of the most educated parent. 9 This interpretation in terms of social capital may be less relevant to activities mainly carried out outside school, as is especially the case in France, than it is to activities mainly carried out within school, though. About the Author Philippe Coulangeon is Senior Research Fellow at the French National Center for Scientific Research (CNRS). His research deals with social stratification, culture, and education in contemporary France. His recently published work appears in Poetics, Cultural Sociology, Revue Française de Sociologie, and International Encyclopedia of the Social and Behavioral Sciences. References Allison, Paul D. 2009. Fixed Effects Regression Models , vol. 160. Thousands Oaks, CA: SAGE Publications. Google Scholar CrossRef Search ADS   Bandura, Albert. 1990. Multidimensional Scales of Perceived Academic Efficacy . Stanford, CA: Stanford University Press. Baudelot, C., M. Cartier, and C. Détrez. 1999. Et pourtant, ils lisent... . Paris: Le Seuil. Bartkus, K. R., B. Nemelka, M. Nemelka, and P. Gardner. 2012. “ Clarifying the Meaning of Extracurricular Activity: A Literature Review of Definitions.” American Journal of Business Education  (Online) 5( 6): 693. Ben Ali, L., and R. Vourc’h. 2015. “Évolution des acquis cognitifs au collège au regard de l’environnement de l’élève. Constat et mise en perspective longitudinale.” Éducation and formations, no. 86–87, pp. 211–34, MENESR-DEPP. http://cache.media.education.gouv.fr/file/revue_86-87/57/4/depp-2015-EF-86-87-evolution-acquis-cognitifs-au-college-au-regard-environnement-eleve_424574.pdf. Blanchard, S., A. Lieury, M. Le Cam, and T. Rocher. 2013. Motivation et sentiment d'efficacité personnelle chez 30 000 élèves de 6e du collège français. Bulletin de psychologie  1: 23– 35. Google Scholar CrossRef Search ADS   Bodovski, K., and G. Farkas. 2008. “ ‘Concerted Cultivation’ and Unequal Achievement in Elementary School.” Social Science Research  37( 3): 903– 19. Google Scholar CrossRef Search ADS   Bourdieu, Pierre. 1986. “The Forms of Capital.” In Handbook of Theory and Research for the Sociology of Education , edited by John G. Richardson, 241– 58. New York: Greenwood. Bourdieu, P., and J. C. Passeron. [ 1970] 1977. Reproduction in Education, Culture and Society . London: Sage. Bowles, S., and H. Gintis. 1976. Schooling in Capitalist America: Educational Reform and the Contradictions of American Life . New York: Basic Books. ———. 2002. “ Schooling in Capitalist America Revisited.” Sociology of Education  75( 1): 1– 18. Google Scholar CrossRef Search ADS   Broh, Beckett A. 2002. “ Linking Extracurricular Programming to Academic Achievement: Who Benefits and Why?” Sociology of Education  75( 1): 69– 95. Google Scholar CrossRef Search ADS   Cabane, C., A. Hille, and M. Lechner. 2015. “Mozart or Pelé? The Effects of Teenagers Participation in Music and Sports.” SOEP Papers on Multidisciplinary Panel Data Research 749. Camp, William G. 1990. “ Participation in Student Activities and Achievement: A Covariance Structural Analysis.” Journal of Educational Research  83( 5): 272– 78. Google Scholar CrossRef Search ADS   Chin, T., and M. Phillips. 2004. “ Social Reproduction and Child-Rearing Practices: Social Class, Children’s Agency, and the Summer Activity Gap.” Sociology of Education  77( 3): 185– 210. Google Scholar CrossRef Search ADS   Coleman, James S. 1961. The Adolescent Society . New York: Free Press of Glencoe. Cousteaux, A.-S., and Y. Lemel. 2004. “Étude de l’homophilie socioprofessionnelle à travers l’enquête contacts.” Document de travail, Centre de Recherche en Économie et Statistique, 10. http://crest.science/RePEc/wpstorage/2004-10.pdf. Covay, E., and W. Carbonaro. 2010. “ After the Bell: Participation in Extracurricular Activities, Classroom Behavior, and Academic Achievement.” Sociology of Education  83( 1): 20– 45. Google Scholar CrossRef Search ADS   Dettermann, D. K., and R. J. Sternberg, eds. 1993. Transfer on Trial: Intelligence, Cognition, and Instruction . Norwood, NJ: Ablex Publications. Dumais, Susan A. 2006. “ Elementary School Students’ Extracurricular Activities: The Effects of Participation on Achievement and Teachers’ Evaluations.” Sociological Spectrum  26( 2): 117– 47. Google Scholar CrossRef Search ADS   European Commission/EACEA/Eurydice. 2016. The Organisation of School Time in Europe. Primary and General Secondary Education–2016/17. Eurydice Facts and Figures . Luxembourg: Publications Office of the European Union. Farkas, George. 1996. Human Capital or Cultural Capital? Ethnicity and Poverty Groups in an Urban School District . New York: Walter de Gruter. ———. 2003. “ Cognitive Skills and Noncognitive Traits and Behaviors in Stratification Processes.” Annual Review of Sociology  29( 1): 541– 62. Google Scholar CrossRef Search ADS   Fejgin, Naomi. 1994. “ Participation in High School Competitive Sports: A Subversion of School Mission or Contribution to Academic Goals?” Sociology of Sport Journal  11( 3): 211– 30. Google Scholar CrossRef Search ADS   Feldman, A. F., and J. L. Matjasko. 2005. “ The Role of School-Based Extracurricular Activities in Adolescent Development: A Comprehensive Review and Future Directions.” Review of Educational Research  75( 2): 159– 210. Google Scholar CrossRef Search ADS   Flammer, A., and F. D. Alsaker. 1999. The Adolescent Experience: European and American Adolescents in the 1990s . Mahwah, NJ: Lawrence Erlbaum Associates. Fletcher, A. C., P. Nickerson, and K. L Wright. 2003. “ Structured Leisure Activities in Middle Childhood: Links to Well‐Being.” Journal of Community Psychology  31( 6): 641– 59. Google Scholar CrossRef Search ADS   Guest, A., and B. Schneider. 2003. “ Adolescents’ Extracurricular Participation in Context: The Mediating Effects of Schools, Communities, and Identity.” Sociology of Education  76( 2): 89– 109. Google Scholar CrossRef Search ADS   Hansen, D. M., R. W. Larson, and J. B. Dworkin. 2003. “ What Adolescents Learn in Organized Youth Activities: A Survey of Self‐Reported Developmental Experiences.” Journal of Research on Adolescence  13( 1): 25– 55. Google Scholar CrossRef Search ADS   Hille, A., and J. Schupp. 2015. “ How Learning a Musical Instrument Affects the Development of Skills.” Economics of Education Review  44: 56– 82. Google Scholar CrossRef Search ADS   Holland, A., and T. Andre. 1987. “ Participation in Extracurricular Activities in Secondary School: What Is Known, What Needs to Be Known?” Review of Educational Research  57( 4): 437– 66. Google Scholar CrossRef Search ADS   ———. 1988. “ Beauty Is in the Eye of the Reviewer.” Review of Educational Research  58( 1): 113– 18. Google Scholar CrossRef Search ADS   Jansen, L. 2016. “The Academic Impact of Extracurricular Activities on Middle School Students.” Doctoral dissertation, Lindenwood University. Kaufman, J., and J. Gabler. 2004. “ Cultural Capital and the Extracurricular Activities of Girls and Boys in the College Attainment Process.” Poetics  32( 2): 145– 68. Google Scholar CrossRef Search ADS   Lareau, Annette. 2003. Unequal Childhoods: Class, Race, and Family Life . Berkeley: University of California Press. Lareau, A., and E. B. Weininger. 2003. “ Cultural Capital in Educational Research: A Critical Assessment.” Theory and Society  32( 5): 567– 606. Google Scholar CrossRef Search ADS   ———. 2008. “ Time, Work, and Family Life: Reconceptualizing Gendered Time Patterns Through the Case of Children’s Organized Activities.” Sociological Forum  23( 3): 419– 54. Google Scholar CrossRef Search ADS   Larson, R. W., and S. Verma. 1999. “ How Children and Adolescents Spend Time across the World: Work, Play, and Developmental Opportunities.” Psychological Bulletin  125( 6): 701– 36. Google Scholar CrossRef Search ADS PubMed  Lipscomb, Stephen. 2007. “ Secondary School Extracurricular Involvement and Academic Achievement: A Fixed Effects Approach.” Economics of Education Review  26( 4): 463– 72. Google Scholar CrossRef Search ADS   Mahoney, J. L., A. L. Harris, and J. S. Eccles. 2006. “ Organized Activity Participation, Positive Youth Development, and the Over-Scheduling Hypothesis.” Social Policy Report  20( 4):3–31. Society for Research in Child Development. Marsh, Herbert W. 1992. “ Extracurricular Activities: Beneficial Extension of the Traditional Curriculum or Subversion of Academic Goals?” Journal of Educational Psychology  84( 4): 553– 62. Google Scholar CrossRef Search ADS   Marsh, H. W., and S. Kleitman. 2002. “ Extracurricular School Activities: The Good, the Bad, and the Nonlinear.” Harvard Educational Review  72( 4): 464– 511. Google Scholar CrossRef Search ADS   McNeal, Ralph B. Jr. 1995. “ Extracurricular Activities and High School Dropouts.” Sociology of Education  68( 1): 62– 80. Google Scholar CrossRef Search ADS   Octobre, S., C. Détrez, P. Mercklé, and N. Berthomier. 2010. L’enfance des loisirs. Trajectoires communes et parcours individuels de la fin de l’enfance à la grande adolescence . Paris: La Documentation Française. Schellenberg, E. Glenn. 2004. “ Music Lessons Enhance IQ.” Psychological Science  15( 8): 511– 14. Google Scholar CrossRef Search ADS PubMed  ———. 2006. “ Long-Term Positive Associations between Music Lessons and IQ.” Journal of Educational Psychology  98( 2): 457– 68. Google Scholar CrossRef Search ADS   ———. 2011. “ Examining the Association between Music Lessons and Intelligence.” British Journal of Psychology  102( 3): 283– 302. Google Scholar CrossRef Search ADS PubMed  Schunck, R. 2013. “ Within and between estimates in random-effects models: Advantages and drawbacks of correlated random effects and hybrid models.” Stata Journal  13( 1): 65– 76. Seow, P. S., and G. Pan. 2014. “ A Literature Review of the Impact of Extracurricular Activities Participation on Students’ Academic Performance.” Journal of Education for Business  89( 7): 361– 66. Google Scholar CrossRef Search ADS   Shulruf, Boaz. 2010. “ Do Extra-Curricular Activities in Schools Improve Educational Outcomes? A Critical Review and Meta-Analysis of the Literature.” International Review of Education  56( 5–6): 591– 612. Google Scholar CrossRef Search ADS   Steinberg, L. 1996. Beyond the Classroom: Why School Reform Has Failed and What Parents Need to Do . New York: Simon and Schuster. Weininger, Elliot B., Annette Lareau, and Dalton Conley. 2015. " What money doesn't buy: Class resources and children’s participation in organized extracurricular activities." Social Forces  94( 2): 479– 503. Google Scholar CrossRef Search ADS   Wooldridge, J. M. 2010. Econometric analysis of cross section and panel data . 2nd ed. Cambridge, MA: MIT press. © The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Forces Oxford University Press

The Impact of Participation in Extracurricular Activities on School Achievement of French Middle School Students: Human Capital and Cultural Capital Revisited

Social Forces , Volume Advance Article – Mar 22, 2018

Loading next page...
 
/lp/ou_press/the-impact-of-participation-in-extracurricular-activities-on-school-bV292W9ZpX
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the University of North Carolina at Chapel Hill. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
0037-7732
eISSN
1534-7605
D.O.I.
10.1093/sf/soy016
Publisher site
See Article on Publisher Site

Abstract

Abstract The impact of participation in extracurricular activities on academic success has long been studied in the social sciences. This article aims at improving the measurement and understanding of this impact. Based on panel data regression models applied to a panel of French middle school students, it first provides a robust estimation of the impact of extracurricular activities on school outcomes (marks in French and Mathematics) and on a set of cognitive and non-cognitive skills. It finds a positive and significant impact on marks in French and Mathematics and scores on non-cognitive skills tests. No impact is found on cognitive skills. The article then investigates the underlying mechanisms of this impact. Its findings do not reinforce the transfer paradigm, according to which extracurricular activities provide students who participate in them with skills that they can reinvest in school life. Neither does it support the notion that such an impact may primarily be the result of students’ greater connivance with the cultural standards of teachers. Instead, it seems likely that what is mainly at stake in participation in extracurricular activities is families’ unequal capacity for extending the time of school supervision in their children’s free time. Therefore, insofar as the varying participation in these activities is strongly correlated to differences in students’ social and cultural background, participation in extracurricular activities would in itself contribute to reinforcing social inequalities in school achievement. Introduction Extracurricular activities have long been considered as a source of inequality among children in relation to school achievement (Covay and Carbonaro 2010; Dumais 2006; Lareau 2003; Lareau and Weininger 2008). According to a restrictive definition, extracurricular activities encompass non-compulsory activities that take place at school but are not included in the curriculum (Bartkus et al. 2012). Understood in a broader sense, they encompass all voluntary organized activities, whether these take place in the school setting or not: sports or artistic practices, such as musical activities, but also participation in youth organizations (e.g., scouting) or leisure clubs. The main rationale for studying the impact of extracurricular activities on educational outcomes is the idea that learning processes are not restricted to explicit transmission in the classroom context. Moreover, since extracurricular activities are not compulsory, their impact is expected to be unequal depending on students’ social background. This article addresses these questions in the current context in France. It is based on a panel of several thousand French middle school (collège) students. Collège is the French equivalent of middle school or junior high school. It corresponds to the first stage of secondary education, with all young people aged 11 to 15 receiving the same general education. So far, the impact of extracurricular activities has been extensively studied at the high school and university level, but less at the middle school level (Jansen 2016). Extracurricular time deserves special attention in the French context due to the specific way of organizing school hours, as compared to other European countries and to the United States. The French school system is characterized by a lower annual number of days of schooling than elsewhere, but the school days are longer (European Commission 2016). In addition, school time tends to be highly concentrated on the core school subjects and, unlike English-speaking countries, schools play a relatively minor role in the management of extracurricular activities in France, where they are mainly supported by the non-profit and volunteer sectors. As a result, extracurricular time remains largely beyond the control of the educational system and more dependent on family resources. In this respect, the French case, where school life and extracurricular activities are much more clearly separated than in many other countries, may contribute to a better identification and understanding of the intrinsic impact of extracurricular activities on school performance in general. The article first reviews the existing literature on extracurricular activities’ observed educational payoffs. It then carries out fixed-effects (FE) regression models in order to demonstrate the robustness of the impact of participation in extracurricular activities on the academic achievement of French middle school students. The rest of the article investigates the processes through which this impact may be seen to be produced. Three competing mechanisms are considered, according to which the impact of extracurricular activities on school achievement is seen to be primarily due to a process of skills accumulation and transfer (human capital model); to a process of social selection (cultural capital model); or to its time-structuring function (“concerted cultivation” model). Finally, the article questions the extent to which this impact varies according to the nature of the activities (cultural vs. non-cultural, outside school vs. inside school). The aim of the article is thus threefold. First, by exploiting the possibilities contained in the panel data on which it relies, it offers a more robust measure of the educational payoff of extracurricular activities than much of the existing research provides. Second, it suggests that the impact of extracurricular activities is more a matter of cultural capital than human capital, which is why these activities would contribute to reinforcing social inequalities in school achievement. Third, it strongly suggests that participation in extracurricular activities affects academic performance through the control it exerts on adolescents’ time use. Literature review: How and to what extent does participation in extracurricular activities affect school achievement? The robustness of extracurricular activities’ educational payoffs The relationship between cognitive development in children and adolescents, their academic performance, and how they spend their time has long been investigated (Larson and Verma 1999). Particular attention has been given to the time devoted to scheduled and supervised leisure activities, such as extracurricular activities (Flammer and Alsaker 1999), whose impact on academic performance has been extensively debated (Seow and Pan 2014; Shulruf 2010). It has long been argued that extracurricular activities might be detrimental to school success (Camp 1990; Coleman 1961), or at least play an ambivalent role in this respect (Broh 2002; Fejgin 1994; Marsh 1992; Marsh and Kleitman 2002; Steinberg 1996), insofar as the time they require is in direct competition with time devoted to academic pursuits. In addition, primarily recreational extracurricular activities have sometimes been deemed to encourage attitudes antagonistic to school values (Coleman 1961; Fejgin 1994). Since the early 2000s, however, an insistence on the positive effects of extracurricular activities has become dominant (Covay and Carbonaro 2010; Feldman and Matjasko 2005; Fletcher, Nickerson, and Wright 2003; Guest and Schneider 2003; Kaufman and Gabler 2004; Lareau 2003; Lipscomb 2007). More recently, several studies have put forward the notion of a possible threshold effect (Seow and Pan 2014): up to a certain point, academic achievement would benefit from participation in extracurricular activities, but excessive commitment of students’ time would result in declining performance (Marsh 1992; Marsh and Kleitman 2002). However, this hypothesis remains rather contentious (Mahoney, Harris, and Eccles 2006). Moreover, the impact of extracurricular activities on school outcomes seems to vary from one activity to another, with mixed results. While some studies insist on the stronger impact of sport in comparison with all other activities (Broh 2002; Marsh 1992; Marsh and Kleitman 2002; McNeal 1995), others emphasize the specific impact of cultural activities, especially those related to music (Hille and Schupp 2015), but with an even stronger impact when combined with sports activities (Cabane, Hille, and Lechner 2015), suggesting a cumulative effect: the more students are involved in diverse activities, whatever they are, the better their academic results (Lareau 2003). Human capital vs. cultural capital Most controversies in this field relate to the very nature of the impact of extracurricular activities on academic achievement, if any. The main divide is between interpretations in terms of human capital and interpretations in terms of cultural capital (Farkas 1996; Kaufman and Gabler 2004). Cognitive and non-cognitive skills Human capital approaches contend that involvement in these activities enhances students’ skills. Skill enhancement may be of a various nature, cognitive as well as non-cognitive. Cognitive skills enhancement assumes that participation in extracurricular activities has an indirect impact on school performances, by improving students’ intellectual abilities or by providing them with competencies that are transferable and applicable in the academic domain (cognitive transfer paradigm). This transfer mechanism has, however, received little support (Dettermann and Sternberg 1993) beyond some studies related to musical education and musical activities (Schellenberg 2004, 2006, 2011). Instead, much of the existing research suggests that participation in extracurricular activities affects academic performance by improving students’ non-cognitive abilities (Broh 2002; Covay and Carbonaro 2010; Holland and Andre 1987; Seow and Pan 2014). Non-cognitive skills include life skills such as organization, planning, time management (Holland and Andre 1987); effort, perseverance, discipline, emotional stability (Farkas 2003); self-esteem, perseverance (Broh 2002); locus of control, self-confidence (Fejgin 1994); self-reliance (Fletcher, Nickerson, and Wright 2003); valuing achievement, respect of adult authority, and the regulation of interactions with others (Covay and Carbonaro 2010). Cultural capital and child-rearing styles According to cultural capital theory (Bourdieu 1986), the impact of participation in extracurricular activities on school achievement is not at all a matter of the transfer of abilities, whether cognitive or non-cognitive; instead, it mainly acts as a status signal (Kaufman and Gabler 2004). In accordance with Bourdieu’s understanding of the notion of cultural capital, students’ participation in extracurricular activities is a demonstration of their cultural resources and their belonging to the upper classes. In this way, they also demonstrate cultural endowments, dispositions, and attitudes well suited to school requirements, shared with teachers, and rewarded as such. This interpretation assumes that participation in extracurricular activities has no intrinsic efficacy but is rather emblematic of the cultural arbitrariness of the school system. Schools tend to reward the possession and control of a certain number of attributes and cultural competences that tend to be monopolized by the ruling class (Bourdieu and Passeron [1970] 1977). This argument converges with Bowles’s and Gintis’s assertion that the persistence of class advantage across generations is due to the fit between family-inherited behavior traits and the implicit norms of teachers and schools, rather than the transmission of cognitive capacities from parents to children (Bowles and Gintis 1976, 2002). In this sense, participation in extracurricular activities may serve as an instrument of the upper classes’ social closure. Lareau’s research on parents’ educational styles, with its seminal contrast between the “concerted cultivation” and the “accomplishment of natural growth” models, might be considered as a variant of the cultural capital hypothesis (Lareau 2003). “Concerted cultivation” refers to a child-rearing style quite common in the middle class, which emphasizes reasoning and dialogue between children and their parents. “Concerted cultivation” is also based on the encouragement of children’s scheduled activities, deemed to improve their talents and abilities, of which engaging in extracurricular activities would be an emblematic case. “Accomplishment of natural growth,” which is more common among working-class families, focuses instead on authority and discipline rather than on exchange and consultation between children and their parents. In addition, children’s leisure time is mostly dedicated to unstructured and improvised activities. Watching TV at home or playing basketball on the street instead of going to a movie theater or playing football at a sports club may illustrate the contrast between these two opposite educational styles. Consistent with broader cultural dispositions, this contrast may lead to a comparative advantage for children raised in a “concerted cultivation” style. This kind of child-rearing environment fosters an aptitude for a certain kind of reasoning and interacting with adults and authorities that provides children with a greater sense of entitlement and self-confidence in social interactions that might be particularly rewarded in the school context. Similarly, the kind of commitment required by structured leisure activities may be well fitted to the time-structuring abilities required by school practices. Participation in extracurricular activities and social reproduction Many sociologists are driven to interpret the impact of extracurricular activities in terms of cultural capital rather than human capital because these interpretative models throw light on the role that participation in these activities plays in the reproduction of social inequalities. Participation in structured extracurricular activities is particularly known to vary across families, depending primarily on parents’ resources, whether economic (Chin and Phillips 2004) or cultural (Dumais 2006; Lareau 2003; Lareau and Weininger 2008; Weininger et al. 2015). Parents from higher socioeconomic status are much more likely to involve their children in extracurricular activities than parents from lower socioeconomic status. To the extent that participation in extracurricular activities is supposed to stimulate academic success, social inequalities in participation in these activities could therefore be seen as a mean through which social privilege is reproduced over generations. Various studies have found empirical evidence of such a mediating role when it comes to participation in extracurricular activities (Bodovski and Farkas 2008; Covay and Carbonaro 2010). Nonetheless, this mediating impact generally appears to be rather small (it explains only a relatively small proportion of the socioeconomic status effect on school performances). The pitfalls of correlational analysis Because of their correlational nature, many of the existing studies fail to adequately grasp the relationship between extracurricular activities and educational achievement because of self-selection, confounding factors, and unobserved heterogeneity (Holland and Andre 1987, 1988; Marsh 1992). Students who participate in these activities may indeed differ from non-participants in regard to some non-observable or non-measurable characteristics or dispositions likely to also have an impact on academic achievement. Various methodological options may be used to overcome, at least partially, the shortcomings of many of the usual analyses in this field. One of them relies on the implementation of fixed-effects (FE) regressions on panel data (Lipscomb 2007), where respondents are taken as their own controls. But this option is seldom explored in analyses of the impact of extracurricular activities on school outcomes, most often because of a lack of adequate data, that is, panel data. The remainder of this article will take advantage precisely of the panel structure of the data on which it is based to apply this kind of regression model. Data and variables The data comes from a French panel of secondary school students commissioned by the statistical studies department of the French Ministry of Education (Direction de l'évaluation, de la prospective et de la performance [DEPP]). It consists of a random sample of 35,000 French students who entered the first grade of middle school (sixth grade) in September 2007. The panel database includes a large amount of information on students’ trajectories and school outcomes recorded annually. It also includes information on students’ family environment provided by two subsequent mail-out surveys submitted to students’ parents in 2008 and 2011. First, the dataset registers the scores on the national assessment tests in French and Mathematics uniformly administered to all French middle school students at the beginning of their sixth grade (students on average aged between 11 and 12). They consist of two 45-minute sequences of exercises designed to measure essential skills in both subjects. Second, the dataset registers the scores obtained by the same students three years later (June 2011) in various subjects, including French and Mathematics, in the “Brevet national des Collèges” (BNC), the final middle school exam (students on average aged between 14 and 15). Third, the dataset includes scores on a set of tests specifically designed for the panel and aimed at measuring wide-ranging abilities, including both cognitive and conative dimensions. The cognitive dimension is measured by a global cognitive score (COG) calculated from the scores obtained on six tests of elementary skills in vocabulary, encyclopedic or long-term memory, mathematical reasoning, completion of incomplete sentences, silent reading comprehension, and assessment of logical reasoning (Ben Ali and Vourc’h 2015). The conative dimension is addressed through three subscales based on Albert Bandura’s (1990) multidimensional scales of perceived efficacy. The first indexes students’ perceived school efficiency (PESCH), that is, beliefs in their ability to succeed in different academic disciplines. The second corresponds to their perceived social efficiency (PESOC), that is, beliefs in their ability to initiate and maintain social relationships and to regulate interpersonal conflicts. The third indexes their perceived efficiency in self-regulation (PESELF), that is, their ability to resist pressure from their peers to engage in deviant behavior (Blanchard et al. 2013). These tests were also administered to students twice, for the first time at the beginning of 2008 and for the second time in 2011. Finally, in 2008 and 2011, students’ parents were asked, among other things, to indicate on a list of eight extracurricular activities those in which their children participated.1 This resulted in a list of eight dummy variables, including participation in a sports club (outside school), participation in the school sports association, registration in a public library, enrollment in a music academy or in a music school, enrollment in a drama class (out of school), participation in a youth organization (e.g., Scouts), enrollment in a youth center, or participation in an activity club (other than sports) at school.2 The addition of these eight dummy variables defines a scale of participation in extracurricular activities ranging from 0 to 8. The initial sample has been subjected to various sources of attrition. Not all the students originally included in the panel were subjected to the cognitive and conative tests. Similarly, not all parents responded to the postal surveys mentioned above. For the sake of coherence, the analysis is thus restricted to the pupils who were subjected to the tests and whose parents responded to the 2008 and 2011 surveys.3 In addition, the analysis is also restricted to the pupils who sat the BNC “on time,” that is, those who had not repeated a class in the interim, which was nearly 90 percent of the sample. As a result, the subsample on which subsequent analyses are based comprises 19,809 individuals. Finally, the various data collection steps generated some missing values that in all subsequent analysis have been treated by multiple imputation, in order to maintain the size of the sample constant.4 Research questions and hypotheses The first aim of the analysis is to produce robust estimates of the impact of participation in extracurricular activities on school achievement which, given previous research on the subject, is expected to be significant and positive. In addition, the hypothesis of a non-linear effect of participation in extracurricular activities, that is, the hypothesis that participation in these activities is beneficial up to a certain point only, will also be tested. H1a: Participation in extracurricular activities has a significantly positive impact on school outcomes. H1b: Up to a certain point, participation in extracurricular activities has a positive impact on academic performance and a negative impact thereafter. It can also be expected that the impact of extracurricular activities does not only depend on the cumulative number of activities in which students participate, but also on their content. H2: The impact of participation in extracurricular activities on school outcomes varies according to the nature of the activities in which students participate. The second goal of the analysis is to clarify the nature of the impact of extracurricular activities on academic achievement. The theory of human capital suggests that this impact is due to the accumulation of specific skills that students can reinvest in school learning. It has often been advanced that this impact would be primarily a matter of accumulation and transfer of non-cognitive skills, such as those related to self-reliance or perceived efficacy, rather than cognitive skills. H3: Participation in extracurricular activities has a greater impact on non-cognitive than on cognitive skills. Insofar as the nature and content of extracurricular activities can at times be very far from those associated with academic goals, its impact on school outcomes is expected to be mainly indirect. Controlling for its impact on cognitive and non-cognitive skills, the remaining impact of extracurricular activities on school outcomes is thus expected to be insignificant. H4: The indirect impact of participation in extracurricular activities on school outcomes is mediated by its direct impact on cognitive and non-cognitive skills. Finally, the cultural capital theory provides an alternative explanation of the mediating function of extracurricular activities. Participation in these activities is deemed to act as a signal of students’ social and cultural backgrounds and reinforce their conformity to the social and cultural norms that the school system tends to reward. Participation in extracurricular activities as such would thus contribute to the social closure and reproduction of the upper classes. Here, the impact of the social status and cultural resources of a student’s family on his or her academic performance is expected to be mediated by participation in extracurricular activities. H5: Participation in extracurricular activities mediates the impact of students’ inherited social status and family-transmitted cultural capital. Method and analytical strategy After standardization of the sixth-grade tests and BNC exam marks, the three sets of indicators for educational outcomes and skills will be treated as dependent variables in regression models measured twice, in 2008 and in 2011. The same holds for the indicators related to students’ involvement in extracurricular activities, also measured in 2008 and 2011. Some of the other characteristics of students and of their environment that deserve to be controlled for in the models because they could affect both the likelihood of students’ engagement in extracurricular activities and their academic results can also be regarded as potentially time-varying, for example students’ parents’ income and the number of their siblings. Other characteristics can be considered as intrinsically time-invariant. Some, such as gender, parents’ level of education, or social class, are observable and measurable. Others are not (unobserved heterogeneity). However, the panel structure of the data allows for the implementation of fixed-effects (FE) regression models that make the control of all these time-invariant characteristics possible, even when unobservable (Allison 2009). FE regression is based on the estimation of the impact of within variations, that is, the impact of a variation in time in an independent variable on the variation in time in a specific outcome for the same respondent, rather than between variations, that is, the variations observed in both dependent and independent variables between different respondents at the same time. In FE regression, all the time-invariant characteristics of the respondents, even when unmeasured, can thus be controlled for by using each individual as his or her own control. Accordingly, in order to estimate whether or not participation in extracurricular activities affects educational outcomes or skills, this article compares the value of these outcomes in 2008 and 2011 while controlling for variation in the extracurricular activities indicators and other relevant time-varying variables. Assuming that students’ other characteristics all remain constant in the interval, the effect estimate of participation in extracurricular activities can be considered as unbiased, that is, it cannot be attributed to confounding factors or omitted time-invariant variables.5 To put it more formally, let us consider the dependent variable Yit (Mathematics or French scores, cognitive or conative test scores), the time-varying predictors Xit (including indicators of participation in extracurricular activities), and the invariant predictors Zi. The basic model for estimating the impact of Xit on Yit can be written as follows:   yit=μt+βXit+γZi+αi+εitwhere μt is a period-specific intercept, β and γ are vectors of coefficients, αi is an error term attached to each individual, and εit is an error term attached to each individual at each point in time. As all the variables under consideration are observed at only two periods (T = 2), the model can be reformulated as   yi1=μ1+βXi1+γZi+αi+εi1and   yi2=μ2+βXi2+γZi+αi+εi2 Then, averaging the two preceding equations, we get   y̅i=μ̅+βX̅i+γZi+αi+ε̅iand subtracting the first equation from the second   yit−y̅i=(μt−μ̅)+β(Xit−X̅i)+(εit−ε̅i),which can be rewritten as   ÿit=μ̈t+βẌit+ε̈it,where the unobserved fixed effects αi as well as the γ coefficients for time-invariant characteristics (between effects, i.e., time-invariant effects but which vary from one individual to another) have disappeared, allowing for unbiased estimation of β coefficients (within effects, i.e., time-varying effects “within” each individual). To test hypotheses 1 to 4, the impact of participation in extracurricular activities on school outcomes is estimated first by means of OLS regressions performed separately on cross-sectional data in 2008 and 2011. These estimates are then compared to the estimates produced by FE regressions of the same outcomes on the same variables performed on panel data. Hybrid models are used to test hypothesis 5. Hybrid models are a class of models that overcome the inability of fixed-effects models to estimate the effect of time-invariant variables. One immediate alternative to FE models that does not suffer from the same disadvantage is the random-effects (RE) model, in which time-invariant independent variables can be included. The main difference between RE and FE models is that RE models assume that the time-invariant unobserved characteristics of the individuals, captured by the abovementioned αi error term, are not correlated to the predictors included in the model. This hypothesis is rarely verified, though, and the time-varying variables’ coefficients are thus biased. Most of the time, FE models must be preferred for this reason.6 To circumvent this problem, it has been proposed to make use of a hybrid form of the RE model that shares the ability of FE models to estimate unbiased coefficients for time-varying variables (Allison 2009; Wooldridge 2010; Schunck 2013). This is made possible by the decomposition of the time-varying effects ( β coefficients) into a between (X̅i) and a within component (Xit−X̅i). This hybrid model (Allison 2009) can be written as follows:   yit=μt+β(Xit−X̅i)+γZi+θ(X̅i)+αi+εit Interestingly, the β coefficients estimated by this hybrid model are the same as those estimated by the FE model, and they are seemingly unbiased. In addition, as this model is a random-effects model, time-invariant effects (Zi) can also be estimated.7 Results The social factors of participation in extracurricular activities Participation in extracurricular activities is unevenly frequent among students. As shown in tables 1a and 1b, in 2008 as well as in 2011, the average number of activities in which students participate is positively and significantly correlated with their parents’ status, social class, level of education, and income. Similarly, differences in rates of participation in almost all activities taken separately are significantly associated with these characteristics. The only exception is participation in a sports association, for which differences associated with social status and income are not significant in 2011. Moreover, the social disparities are maximal when it comes to truly extracurricular activities (participation in a sports club, in a music academy, in a drama class) and minimal when it comes to extracurricular activities that take place at school (participation in a sports association or in an activity club other than sports in the school context). In addition, the activities for which the differences in participation according to parents’ economic, educational, and social resources are the highest are also the most emblematic of the realm of highbrow culture (enrollment in a library, in a music academy, or in a drama class). In all probability, the unequal participation of children from unequal social backgrounds does not necessarily reflect an unequal desire to participate. Rather, prior studies suggest that these social differences come mainly from an unequal access to the material and cultural resources that condition participation (Chin and Phillips 2004). Table 1a. Descriptive statistics (2008) 2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Table 1a. Descriptive statistics (2008) 2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  2008  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 357.3  Chi2(3): 14.1  Chi2(3): 733.4  Chi2(3): 68.6  Chi2(3): 97.2  Chi2(3): 202.4  Chi2(3): 859.6  Chi2(3): 57.2  Chi2(3): 135.1  p < 0.0001  p = 0.003  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.7  27.7  47.2  23.5  3.6  42.4  10.6  9.3  2.2  QS2  1.8  29.9  52.6  25.3  6.0  44.7  13.6  8.2  2.7  QS3  2.0  29.1  61.1  26.7  6.2  49.2  20.2  7.3  3.3  QS4  2.4  30.6  71.7  30.3  8.3  55.8  31.3  5.7  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Head of household socio-economic class (ESeC)  F(8): 141.3  Chi2(8): 24.6  Chi2(8): 783.6  Chi2(8): 94.1  Chi2(8): 103.0  Chi2(8): 271.2  Chi2(8): 848.7  Chi2(8): 69.4  Chi2(8): 143.8  p < 0.0001  p = 0.002  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  2.4  30.2  71.5  29.4  8.4  54.4  30.8  5.9  6.6  2—Lower grade professionals  2.3  31.3  65.3  30.5  7.1  57.2  25.9  7.2  4.2  3—Higher-grade white-collar workers  1.9  30.9  53.3  26  6.0  47.3  14.4  8.1  2.6  4—Petty bourgeoisie or self-employed  1.9  26.9  60.8  25.1  6.4  44.0  20.3  6.9  3.6  5—Farmers  1.8  29.6  55.4  30.6  5.8  37.6  15.9  5.2  3.1  6—Higher-grade blue-collar workers  2.0  30.5  61.4  25.9  5.7  49.0  16.7  6.9  2.5  7—Lower-grade white-collar workers  1.6  27.1  43.3  21  5.4  41.4  9.5  10.7  2.2  8—Skilled workers  1.7  27.7  48.8  24.2  3.5  43.2  11.0  8.8  2.4  9—Semi- and non-skilled workers  1.6  27.5  43.3  21.9  4.0  40.5  9.7  10.5  1.8  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  Parents’ education (1)  F(1): 1,387.2  Chi2(1): 25.5  Chi2(1): 728.1  Chi2(1): 155.1  Chi2(1): 80.9  Chi2(1): 330.2  Chi2(1): 774.0  Chi2(1): 30.5  Chi2(1): 152.1  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.7  27.8  48.9  22.7  4.4  41.5  11.3  8.7  2.0  >: median  2.3  30.8  67.9  30.6  7.5  55.0  26.9  6.7  5.5  All  2.0  29.2  57.8  26.4  5.9  47.9  18.6  7.7  3.6  Parents’ income quartiles  F(3): 393.5  Chi2(3): 9.1  Chi2(3): 1,200.0  Chi2(3): 117.7  Chi2(3): 61.4  Chi2(3): 117.7  Chi2(3): 974.7  Chi2(3): 64.2  Chi2(3): 125.0  p < 0.0001  p = 0.029  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.6  28.1  40.6  21.6  4.7  43.6  10.0  10.2  2.2  Q2  1.8  28.8  53.8  24.5  4.9  44.9  14.3  7.3  2.9  Q3  2.1  31  63.3  29  5.7  48.5  17.6  7.5  3.0  Q4  2.4  28.8  73.3  30.2  8.2  54.3  32.8  5.9  6.3  All  2.0  29.2  57.6  26.3  5.9  47.8  18.5  7.7  3.6  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Table 1b. Descriptive statistics (2011) 2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Table 1b. Descriptive statistics (2011) 2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  2011  Average number of extracurricular activities  Participation in: (in percentage)      Sports association  Sports club  Club  Drama class  Library  Music academy  Youth center  Youth organization  Parents’ social status quartiles  F(3): 255.4  Chi2(3): 6.7  Chi2(3): 684.4  Chi2(3): 19.1  Chi2(3): 66.3  Chi2(3): 95.6  Chi2(3): 791.9  Chi2(3): 94.6  Chi2(3): 144.3  p < 0.0001  p = 0.081  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  QS1  1.3  20.5  42.7  11.6  2.9  30.3  8.9  10.4  2.0  QS2  1.4  22.1  48.1  12.4  4.6  32.3  12.6  9.1  2.8  QS3  1.6  21.7  55.3  13.7  5.1  34.3  17.5  7.8  3.2  QS4  1.9  22.6  66.0  14.3  6.2  39.4  28.2  5.3  6.4  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Head of household socio-economic class (ESeC)  F(8):97.1  Chi2(8): 16.6  Chi2(8): 668.8  Chi2(8): 32.3  Chi2(8): 66.9  Chi2(8): 155.4  Chi2(8): 776.5  Chi2(8): 100.9  Chi2(8): 160.0  p < 0.0001  p = 0.034  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  1—Higher grade professionals  1.8  21.9  65.4  13.5  6.0  37.8  27.7  5.4  6.7  2—Lower grade professionals  1.8  22.6  60.6  16.0  6.0  41.6  22.9  7.1  4.0  3—Higher-grade white-collar workers  1.5  21.4  48.7  12.4  4.4  34.1  13.1  9.3  2.8  4—Petty bourgeoisie or self-employed  1.5  21.5  53.8  12.3  4.9  29.0  16.1  7.0  2.8  5—Farmers  1.4  24.6  49.5  14.0  5.8  25.5  13.9  5.9  4.6  6—Higher-grade blue-collar workers  1.6  23.7  56.3  13.2  4.9  33.3  16.2  8.7  3.0  7—Lower-grade white-collar workers  1.3  20.3  40.2  12.5  4.1  35.0  9.8  10.5  1.8  8—Skilled workers  1.3  20.1  43.7  12.0  2.9  30.8  9.4  10.2  2.0  9—Semi- and non-skilled workers  1.2  21.5  40.2  10.6  2.9  29.1  7.4  11.1  1.7  All  1.5  21.6  52.5  12.9  4.6  33.9  16.4  8.3  3.5  Parents’ education (1)  F(1): 1,084.1  Chi2(1): 15.9  Chi2(1): 778.8  Chi2(1): 33.5  Chi2(1): 57.7  Chi2(1): 218.8  Chi2(1): 751.6  Chi2(1): 48.4  Chi2(1): 168.9  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  <median  1.3  20.5  43.2  11.6  3.5  29.2  9.7  9.5  1.8  >: median  1.8  22.9  63.3  14.5  5.8  39.3  24.2  6.8  5.5  All  1.5  21.6  52.6  12.9  4.6  33.9  16.5  8.2  3.5  Parents’ income quartiles  F(3): 238.9  Chi2(3): 2.4  Chi2(3): 955.0  Chi2(3): 10.6  Chi2(3): 31.8  Chi2(3): 31.8  Chi2(3): 803.2  Chi2(3): 69.4  Chi2(3): 120.5  p < 0.0001  p = 0.497  p < 0.0001  p = 0.014  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  p < 0.0001  Q1  1.3  21.4  37.9  11.7  4.0  32.0  8.8  10.8  1.7  Q2  1.4  21.1  48.8  12.7  3.7  33.0  11.8  8.7  3.0  Q3  1.6  22.3  57.5  13.8  4.9  34.2  17.6  7.9  3.3  Q4  1.9  22.1  68.2  13.8  6.0  37.4  29.0  5.9  6.1  All  1.5  21.7  52.6  13.0  4.6  34.1  16.5  8.4  3.5  (1) number of years of education of the parent with the highest degree. F test on means difference in italics in column one, with df in parentheses. Chi square test of independence in italic in columns 2–9, with df in parentheses. Source: MENESR-DEPP, panel 2007. Besides, comparison between tables 1a and 1b displays a substantial reduction in the average number of extracurricular activities participated in by students between 2008 and 2011. The relative frequency of each of the activities taken separately follows the same tendency, which is coherent with a global trend usually observed at these ages, at least in the French context. Prior studies have shown that early adolescence (between 10 and 15 years of age) often corresponds to a decline in reading (Baudelot, Cartier, and Détrez 1999) as well as a relative withdrawal from many other cultural activities (Octobre et al. 2010). This trend generally corresponds to the growing importance of informal peer relationships when students get older, detrimental to other activities, and to the progressive empowerment of children, whose parents generally prescribe most of the extracurricular activities performed at younger ages (ibid.). Participation in extracurricular activities and school achievement Table 2 displays the results of two OLS and FE regression models of the marks obtained in French and Mathematics on the scale of participation in extracurricular activities, ranging from 0 to 8. A first OLS regression model is estimated on two separate cross-sectional datasets, one in 2008 and the other in 2011, where the impact of participation in activities is controlled for gender, parents’ social status, parents’ income, parents’ education, and number of siblings.8 Marks, social status, education, and income are standardized (H1a). A second OLS model adds a quadratic term for the scale of participation in extracurricular activities, in order to test for the hypothesis of a non-linear effect (H1b). Two FE regression models that include the same variables as the two previous OLS regressions, except for the time-invariant variables (parents’ education and status), are then tested. As before, the second FE model tests for the presence of a non-linear effect (H1b). Both FE models add several time-varying controls that could potentially act as confounding factors. These include change of school between the two dates, change of area of residence, serious illness, death of father, mother, brother, or sister, and loss of employment by father or mother. Table 2. OLS and fixed-effects regressions of Mathematics and French marks   French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005    French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. Table 2. OLS and fixed-effects regressions of Mathematics and French marks   French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005    French marks  Mathematics marks  OLS  FE  OLS  FE  2008    2011        2008    2011        Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Model 1  Model 2  Extracurricular activities  0.086***  0.147***  0.094***  0.138***  0.036***  0.048***  0.089***  0.169***  0.097***  0.157***  0.030***  0.070***    (0.006)  (0.017)  (0.006)  (0.016)  (0.005)  (0.012)  (0.006)  (0.018)  (0.006)  (0.016)  (0.005)  (0.012)  Extracurricular activities2    −0.014***    −0.011**    −0.003    −0.018***    −0.016***    −0.009***      (0.003)    (0.004)    (0.003)    (0.004)    (0.004)    (0.003)  Gender  0.352***  0.355***  0.450***  0.453***      −0.252***  −0.248***  −0.056***  −0.052***        (0.014)  (0.014)  (0.014)  (0.014)      (0.014)  (0.013)  (0.014)  (0.014)      Parents’ education  0.230***  0.229***  0.261***  0.260***      0.219***  0.217***  0.287***  0.286***        (0.009)  (0.009)  (0.009)  (0.009)      (0.009)  (0.009)  (0.009)  (0.009)      Parents’ status  0.107***  0.106***  0.135***  0.135***      0.097***  0.097***  0.121***  0.120***        (0.008)  (0.008)  (0.009)  (0.009)      (0.009)  (0.009)  (0.010)  (0.010)      Parents’ income  0.048***  0.048***  0.040*  0.039*  −0.007  −0.007  0.055***  0.055***  0.053**  0.052**  −0.005  −0.005    (0.008)  (0.008)  (0.017)  (0.017)  (0.005)  (0.005)  (0.009)  (0.009)  (0.019)  (0.019)  (0.007)  (0.007)  Siblings  −0.071***  −0.070***  −0.053***  −0.052***  0.023*  0.023*  −0.051***  −0.050***  −0.039***  −0.038***  0.041***  0.040***    (0.007)  (0.007)  (0.007)  (0.007)  (0.011)  (0.011)  (0.007)  (0.007)  (0.006)  (0.006)  (0.010)  (0.010)  Change of school          0.003  0.004          0.002  0.005            (0.022)  (0.022)          (0.022)  (0.022)  Change of place of residence          −0.080  −0.080          −0.156***  −0.156***            (0.049)  (0.049)          (0.047)  (0.047)  Father’s job loss          0.018  0.017          −0.014  −0.014            (0.018)  (0.018)          (0.018)  (0.018)  Mother’s job loss          −0.024  −0.024          −0.000  −0.000            (0.018)  (0.018)          (0.017)  (0.017)  Serious illness          −0.090*  −0.089*          −0.033  −0.031            (0.045)  (0.045)          (0.043)  (0.043)  Father’s death          −0.073  −0.073          −0.122  −0.124            (0.073)  (0.073)          (0.072)  (0.071)  Mother’s death          0.020  0.020          −0.012  −0.011            (0.107)  (0.107)          (0.100)  (0.099)  Sibling’s death          −0.008  −0.008          −0.035  −0.034            (0.082)  (0.082)          (0.093)  (0.094)  Intercept  −0.538***  −0.591***  −0.804***  −0.833***  −0.065*  −0.074**  0.345***  0.276***  −0.060*  −0.101***  −0.067**  −0.093***    (0.026)  (0.029)  (0.026)  (0.028)  (0.026)  (0.027)  (0.026)  (0.029)  (0.026)  (0.028)  (0.025)  (0.027)  N  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  19,809  R2  0.193  0.194  0.217  0.218  0.009  0.010  0.172  0.174  0.191  0.192  0.003  0.005  Standard errors in parentheses. * p < 0.05 ** p < 0.01 *** p < 0.001 Source: MENESR-DEPP, panel 2007. In the first OLS regression model, the estimated impact of the number of activities is globally significant for Mathematics and French marks, in 2008 as well as in 2011. In the first FE model, the impact of the number of activities remains significant. In comparison with the OLS regression, however, this impact is two to three times lower. The second OLS model depicts a non-linear effect of participation in extracurricular activities in 2008 and 2011, for marks in both French and Mathematics. However, this non-linear effect is only partially confirmed by FE models. The coefficient for the quadratic term of the scale of participation in extracurricular activities is of the expected sign for both marks in French and Mathematics, but is only significant for the latter, with a turning point just below 4 (3.9). As 95 percent of the sample display a maximum score of 4 on the scale of participation in extracurricular activities, this inversion of the effect concerns only a marginal fraction of the pupils who are particularly involved in these activities. Overall, these results are partially supportive of hypothesis 1. Involvement in extracurricular activities is positively and significantly related to school achievement indicators such as marks in French and Mathematics (H1a), but the strength of this impact seems highly overestimated by OLS regression. Part of the effect seemingly associated with participation in extracurricular activities is in fact due to unobserved heterogeneity between respondents that may cause both participation in these activities and school success. Besides this, the non-linear effect of participation in extracurricular activities (H1b) is only supported for Mathematics marks. Finally, relative to the value of the standard deviations of marks in French and Mathematics, the addition of one activity is associated with a relatively modest gain of 0.8 points (on a scale of 0 to 100) in French and of 1.3 points in Mathematics. In other words, the difference between a student who does not participate in any activity and a student who participates in 4 would be of 3.3 points in French and, taking into account a possible non-linear effect, of 5.3 points in Mathematics (on a scale from 0 to 100). The varying impact of participation according to the nature of extracurricular activities The FE regression model presented in table 2 was then rerun with the participation scale replaced by the eight dummy variables from which it derives. The parameter estimates of the impact of participation in these eight activities on marks in French and Mathematics are summarized in figure 1. In the graph, the estimated values of the regression coefficients of the eight activities are reported together with their 95 percent confidence interval. Figure 1. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on marks in French and Mathematics (with 95 percent CI) Control variables: Parents’ income, level of education, number of siblings, change of school between 2008 and 2011l, change of place of residence, father’s job loss, mother’s job loss, serious illness, father’s death, mother’s death, sibling’s death. Source: MENESR-DEPP, panel 2007. Figure 1. View largeDownload slide FE regression coefficients estimates of the impact of participation in eight extracurricular activities on marks in French and Mathematics (with 95 percent CI) Control variables: Parents’ income, level of education, number of siblings, change of school between 2008 and 2011l, change of place of residence, father’s job loss, mother’s job loss, serious illness, father’s death, mother’s death, sibling’s death. Source: MENESR-DEPP, panel 2007. The reported estimates range from −0.03 (impact of enrollment in a youth center on mark in Mathematics) to 0.07 (impact of participation in an activity club at school on mark in French and impact of enrollment in a library on mark in Mathematics). Not all these effects are significant. The only activities whose impact is significant for both French and Mathematics marks are enrollment in a library, enrollment in a music academy or school of music, and participation in an activity club at school (excluding sports activities). The impact of participation in a sports club is significant in regard to Mathematics marks only, whereas participation in a sports association at school is significant for French marks only. When compared to the value of standard deviations in French and Mathematics marks, these effects appear relatively modest. Concerning marks in Mathematics, these range from an extra 0.6 points (on a scale of 0 to 100) for the impact of participation in a sports club to an extra 1.5 points for the impact of enrollment in a library. Concerning French marks, these range from an extra 0.5 points for enrollment in a music academy or in the sports association at school to an extra 1.2 points for participation in an activity club at school. However, the computation of confidence intervals reproduced in figure 1 suggests that most of the differences observed in the magnitude of the impact of the various activities are not statistically significant. Overall these results give ambivalent support to the hypothesis that the various extracurricular activities unevenly affect school outcomes (H2). Several of the extracurricular activities under consideration do not significantly impact school outcomes, but those with significant impact do not significantly differ from each other in regard to the strength of their impact. The impact of extracurricular activities on cognitive and conative skills Table 3 displays the estimates of the impact of extracurricular activities on the scores obtained by the student panelists on the cognitive and non-cognitive tests included in the surveys of 2008 and 2011. The first indicator corresponds to the global score on cognitive tests (COG), followed by three indicators that correspond to the scores on conative tests of the academic dimension (PESCH), the social dimension (PESOC), and the self-regulation dimension (PESELF) of the perceived efficiency of students. For each of the four variables, the first two columns correspond to OLS regressions on cross-sectional data (2008 and 2011) and the third column to an FE model performed on panel data. Independent variables are the same as in the previous models for French and Mathematics marks. The test for a non-linear effect, which proved to be insignificant in each case, is not reproduced in the table. Table 3. OLS and fixed-effects regressions of cognitive and conative test scores   Cog  Pesoc  Pesch  Peself  OLS  FE  OLS  FE  OLS  FE  OLS  FE  2008  2011    2008  2011    2008  2011    2008  2011    Extracurricular activities  0.095***  0.107***  −0.002  0.074***  0.095***  0.031***  0.069***  0.087***  0.018**  0.045***  0.042***  0.033***    (0.006)  (0.006)  (0.003)  (0.006)  (0.006)  (0.006)  (0.007)  (0.006)  (0.006)  (0.006)  (0.006)  (0.007)  Gender  −0.022  −0.011    −0.212***  −0.402***    0.269