Abstract This study seeks to deepen our understanding of the factors that explain individuals’ willingness to self-censor (WtSC)—the proclivity to withhold an opinion from an audience perceived to disagree with that opinion. It does so by testing the “impressionable years” hypothesis, which states that the historical context experienced between the age of 18 and 25 years has a lasting effect on individual dispositions such as WtSC. The study was conducted in Chile, an ideal case to explore possible cohort effects because of the profound political changes experienced there in the past 50 years. Analysis of an original cross-sectional survey shows that—as expected—people who came of age in periods of political repression exhibit significantly higher levels of WtSC later in life compared with those who grew up during less repressive times. Creating opportunities for formal and informal deliberation among citizens is a crucial aspect of democratic systems that seek to promote political participation—systems in which citizens can aspire to be part of the decision-making process and are not simply content with electing representatives (Dahl, 1989; Delli-Carpini et al., 2004; Chambers, 2003; Fishkin, 1995). Citizens, however, often refrain from expressing themselves, even when freedom from censorship and coercion is legally protected. Among the explanations for this behavior are the lack of opportunities, motivation, and/or the ability to express political opinions (Glynn, Hayes, & Shanahan, 1997; Oshagan, 1996). Yet, even when there is opportunity, motivation, and ability, work on the spiral of silence theory shows that individuals may still refrain from expressing their opinions if they perceive their opinion belongs to the minority (Donsbach, Salmon, & Tsfati, 2014). This is particularly true for people who generally fear confronting what is perceived as a hostile audience, that is, people with high levels of willingness to self-censor (WtSC; see Hayes, Glynn, & Shanahan, 2005a, 2005b). When citizens refrain from expressing a political opinion out of fear of disagreement and its consequences, the democratic ideal of giving equal voice to all citizens is hindered. Thus, an important task of public opinion scholars is to examine what triggers people’s disposition to self-censor their opinions. The available evidence suggests that WtSC is related to individuals’ socialization processes. For example, it has been found that communication patterns experienced in childhood, such as an authoritarian style within family and school (Rojas & Hopke, 2010; Rojas, 2012), can delve into a personality trait that leads to a preference to avoid discussion with those who have different points of view. In the current study, we extend this line of work by arguing that not only microlevel variables such as family or school matter, but that the macro-context, such as the level of political activity and/or repression in a country, can also socialize individuals to have a higher or lower propensity to self-censor the expression of minority opinions (Lup, 2015). More specifically, this article applies to the study of WtSC the “impressionable years” or “critical period” hypothesis, which posits that events experienced in adolescence and early adulthood have a disproportionate effect on attitudes in later life (Schuman & Corning, 2006, 2017; Schuman, Corning, & Schwartz, 2012; Tessler, Konold, & Reif, 2004). Thus, we check whether growing up in a politically repressive context has a long-lasting positive impact on WtSC, while doing it in a politically liberal context has a durable negative effect on it. To test this expectation, we conducted an original survey in Chile, a country that provides an ideal scenario to find macro-level influences on individual predispositions, as it has transitioned from a brutal dictatorship to a successful, mature democracy in less than 50 years. This article is organized as follows: first, we review the concept of WtSC and the main theoretical approximations on the relation between political processes and generations. Second, we provide a brief background on Chile and show its utility for the purposes of this study. After detailing the data and the method used, we present the results of the statistical analysis. In the last section, we discuss limitations and propose directions for future research. The Concept of WtSC Democracy is conceived as the political expression of a social organization in which mutual understanding is the most important form of social integration (Rojas, 2006). In a deliberative model of democracy, citizens not only voice sporadic judgments, such as voting, but also participate in the dialogue that occurs within the decision-making process (Dahl, 1989). This conceptual construction of a deliberative democracy, in which citizens debate public affairs, dictates the need for having informed citizens, who are free from both state- and self-censorship (Delli-Carpini et al., 2004; Fishkin, 1995). Public opinion research, however, has found that people vary in their disposition to express and share their political opinions, especially among audiences perceived as hostile to those opinions. The most elaborate argument on why people may refrain from public expression of political opinion was elaborated by Noelle-Neumann’s (1974) spiral of silence theory, which established that individuals’ perceptions of the distribution of public opinion determines their likelihood of expressing or not a political opinion. Using what she called a “quasi-statistical sense,” people are more willing to express their opinion to an audience if they perceive that it agrees with that opinion. When they perceive that the audience may be hostile to that opinion, however, individuals will feel pressured to silence it. To test the robustness of this key aspect of the theory, Glynn, Hayes, and Shanahan (1997; also Glynn & Huge, 2014) conducted a meta-analysis on studies that explore the relationship between perceptions of support for one’s opinions and willingness to express these opinions. Consistent with spiral of silence theory, they concluded that most empirical research shows a significant relationship between the perceived climate of opinion and political expression, although this correlation is small in magnitude. It is in the context of spiral of silence theory that the concept of WtSC was developed, which Hayes et al. (2005a, 2005b) defined as the inhibition of expressing one’s opinion in front of people who might not agree with it. Defined as a trait or disposition, WtSC has been found to explain why the effects of the opinion climate on individual expression are more prevalent for some individuals (e.g., those high in WtSC) and rather weak for others (e.g., those low in WtSC). Importantly, this individual difference approach seems to apply worldwide. Using data from an international sample on four continents, Matthes and his colleagues (2012) found that fear of social isolation is positively correlated with WtSC in all studied countries, except China. Although the main purpose of developing the WtSC concept was to identify a moderator variable of the opinion climate–individual expression relationship, studying the causal antecedents of WtSC has attracted scholarly attention as well. Thus, existing research has shown that it is closely related to shyness, communication apprehension, anxiety about social interaction, self-awareness about what others think, fear of negative evaluation, lower argumentation skills, low self-esteem, less interest in politics, and having experienced an authoritarian communication setting at home or in school during childhood (Hayes et al., 2005a, 2005b; Rojas, 2012; Rojas & Hopke, 2010; Valenzuela, 2011). Other individual-level antecedents of public expression of political opinions could also be applicable to WtSC, such as importance of the topic discussed for the individual (Kim, Han, Shanahan, & Berdayes, 2004), the role of majority opinion within reference groups (Hayes, 2007; Oshagan, 1996), and attitude certainty—that is, the subjective sense of conviction with which one holds an attitude (Matthes, Rios Morrison, & Schemer, 2010; also Lasorsa, 1992). What is lacking in the study of antecedents of WtSC is the role played by contextual variables, such as a country’s specific political history. For instance, in Lup’s (2015) study of informal political conversation networks, it was found that in countries with an antidemocratic past people are less willing to engage in political discussions than in countries with long-lasting democratic systems. In the current study, we extend this line of research to examine whether the political contexts, as manifested by the collective memories of a generation (Schuman & Corning, 2006, 2017), are significant causes of individuals’ WtSC. Generations and Political Dispositions Existing research suggests that a major determinant of WtSC is political socialization, understood as a series of developmental processes through which individuals acquire political orientation and patterns of political behavior (Dennis & Easton, 1969). In the context of WtSC, as explained above, studies by Rojas (2012; also Rojas & Hopke, 2010) have shown that the degree of authoritarianism with which people are educated during childhood and early adolescence could explain their proclivity to self-censor their opinion in adult life. In this case, authoritarian communication is conceptualized as interpersonal communication that promotes acceptance of authority and does not stimulate rational argument and expression of feelings. But school and family are not the only relevant agents of socialization. Social scientists have demonstrated that historical and political contexts experienced before adulthood can also produce long-lasting effects on individuals’ behaviors (see, e.g., Griffin, 2004; Jennings, 1996; Schuman & Corning, 2017). This theoretical perspective starts with Mannheim’s (1952) first analysis of generations, in which he argues that even when different age cohorts experience the same historical events (e.g. wars, economic crisis, and so forth), the influence of particular events is deeper in young people. He posits that individuals’ formative—or “impressionable”—years (i.e., the years between the age of 18 and 25 years) are crucial in political socialization, so that people living the same sociopolitical events in this period of life constitute a distinguishable generation. More recent studies support Mannheim’s (1952) argument on the long-lasting influence of the “impressionable years” on political attitudes and dispositions. In their 50-year longitudinal study on women, Alwin, Cohen, and Newcomb (1991) reported that the attitudes they developed during college were sustained over the course of their lives. Likewise, Schuman and colleagues (Schuman & Corning, 2006, 2017, Schuman et al., 2012; also Corning & Schuman, 2015) found that national and global events experienced during the “impressionable years,” such as the Kennedy assassination for Americans born in the 1950s or the fall of the Berlin Wall for young adults in Germany in 1989, have a disproportionate effect on memories, attitudes, and actions later in adult life. Analyzing a non-Western society, Tessler, Konold, and Reif (2004) found evidence that Algerians who lived their late adolescence and early adulthood during president Houari Boumédiène’s regime—a 13-year period marked by low political participation and state-driven economic policies—share attitudes that distinguish them from other cohorts. Other findings also support the hypothesis of the “impressionable years” hypothesis, such as studies showing that openness to new political attitudes decreases with age (Alwin & Krosnick, 1991; Osborne, Sears, & Valentino, 2011). Based on the robustness of the literature on the effects of generational memory on political dispositions and attitudes, in the current study, we posit for the existence of a relationship between the political context in which people came of age and levels of WtSC later in life. The expectation is that, in addition to the microlevel antecedents reported by prior studies on self-censorship, the “impressionable years” hypothesis offers one route for identifying macrolevel influences on WtSC. More specifically, our central hypothesis posits that: H1. People who experienced most of their “impressionable years” (i.e., between the ages of 18 and 25 years) in politically repressive contexts have higher levels of WtSC. Conversely, people who experienced most of their “impressionable years” in politically expressive or participatory times have lower levels of WtSC. Chile as a Case of Study Owing to the profound political transformations experienced by Chile in the past 50 years, the recent history of that country offers a particularly appropriate context for testing the “impressionable years” hypothesis as applied to WtSC. After a scenario of increasing political polarization during the late 1960s, and in the international context of the Cold War, socialist Salvador Allende was elected president in 1970. For the next 3 years, the ideological battle between leftist, centrist, and right-wing parties reached a high point because of Allende’s ambitious program of social reforms. On September 11, 1973, a military coup overthrew Allende and his Unidad Popular government, leaving all authority in the hands of a right-wing, U.S.-backed military junta led by general Augusto Pinochet. Congress and political parties were dissolved, all democratically elected authorities and unions were proscribed, and the media, censored. As could be expected, violation of human rights became systematic, in forms as diverse as exiles, executions, tortures, and disappearances. Thus, between 1974 and 1981, political participation and public expressions of dissent were practically nonexistent. In 1982, however, a severe economic crisis triggered massive street demonstrations against Pinochet’s dictatorship. The protest movement that ensued continued organizing demonstrations well over 1986, which forced the military to open the political process, legalize most political parties in the opposition, and permit expressions of opposition to the regime in the media. The pressure against the military spiked in 1988, when a plebiscite was called to decide the permanence of Pinochet in power (Valenzuela & Brandão, 2015). The dictatorship was defeated and in 1989 Patricio Aylwin, the candidate of a center-left coalition, won the first democratic presidential election in 20 years. Return to democracy, however, initiated a complex political era. Pinochet remained as the army’s commander-in-chief and, consequently, threatened the Aylwin administration with several so-called “military exercises” over the 1990s. At the same time, both Aylwin and his successor, Eduardo Frei, sought to prevent a return of military rule by advancing a consensual system of policymaking between the governing coalition and the center-right opposition. Thus, open confrontation and disagreement were discouraged by political elites across the ideological spectrum. The election of Ricardo Lagos in 2000—the first socialist president elected since—triggered a new era of political liberalization, with several constitutional reforms directed at consolidating civilian control of the military and strengthening democratic normalization. Thus, during the first decade of the 21st century, Chile witnessed a reemergence of political disagreement and citizen action. Vivid examples of this process are the massive student demonstrations of 2006 and 2011, and the environmental movement of 2010 and 2011 (see Fleet & Guzmán-Concha, 2017; Scherman, Arriagada, & Valenzuela, 2015). Method Data To test the generational hypothesis as applicable to WtSC, this study used an original survey conducted in 2011 among a representative sample of the general population living in Chile’s three largest metropolitan areas: Greater Santiago, Greater Valparaíso, and Greater Concepción. Together, these areas contain 64% of the country’s adult population. The questionnaire was developed by the authors, whereas fieldwork was carried on by the School of Journalism of Diego Portales University and Feedback, a professional polling firm, between August 19 and September 11, 2011. The survey used in this study is part of a research program initiated in 2009 to analyze political participation and media use of youth (people aged between 18 and 29 years), and is conducted every year to a sample of 1,000 people (for more details, see Scherman, Arriagada, & Valenzuela, 2016). In 2011, we included an additional sample of 737 participants aged 30 years or more with the purpose of producing a sample that is representative of the adult age population while, at the same time, is comparable with the samples used in prior and subsequent years. Thus, the total sample size was 1,737 individuals aged 18 years and older. Before the analysis, the sample was weighted according to population parameters. The margin of error of the survey was ±2.4%, with a confidence level of 95% and an 80% response rate (American Association for Public Opinion Research RR6 calculation). Although the total sample size of the survey was 1,737, only 881 respondents could be assigned to one of the four cohorts analyzed, as only these respondents fulfilled Tessler et al.’s (2004) criteria of living at least 6 years of the 8 between 18 and 25 years in the same historical period (i.e., the remaining 856 respondents were excluded because they lived <6 of their “impressionable years” within the same period). Finally, 80 individuals from the 881 were dropped as well because of missing values in independent and/or dependent variables. Thus, the final analysis was conducted over 801 individuals. Importantly, comparing the covariate distribution of the final sample with the initial sample or with a benchmark survey revealed few differences, as detailed in Table A1. Variables To measure WtSC, the key dependent variable, we used the scale developed initially by Hayes, Glynn, and Shanahan (2005a, 2005b) and applied in Chile with success by Matthes, Hayes, Rojas, Shen, Min, and Dylko (2012). Through a set of eight questions, the scale measures individuals’ willingness to withhold their opinion from an audience perceived to disagree with that opinion. In the current study, we used a shortened version of six items, which has been used in Latin America by Rojas and others (Rojas, 2012; Rojas & Hopke, 2010). Specifically, each respondent was asked their level of agreement or disagreement, using a five-point scale (1 = strongly disagree, 5 = strongly agree), to the following statements: “It is difficult for me to express my opinion if I think others won’t agree with what I say”; “There have been many times when I have thought others around me were wrong but I didn’t let them know”; “When I disagree with others, I’d rather go along with them than argue about it”; “I’d feel uncomfortable if someone asked my opinion and I knew that he or she wouldn’t agree with me”; “I tend speak my opinion only around friends or other people I trust”; and “It is safer to keep quiet than publicly speak an opinion that you know most others don’t share.” These six items were then averaged into a single scale of WtSC (M = 2.51; SD = 1.13; Cronbach’s α = 0.85). The explanatory variables are organized into three categories: generational memory, authoritarian communication in school and family, and sociodemographic controls. To measure political generations, we used the respondent’s age at the time of the survey to classify the sample into four dummy-coded political generations according to Chile’s recent history. Prior literature on the “impressionable years” hypothesis (Mannheim, 1952; Osborne, Sears, & Valentino, 2011; Schuman & Corning, 2006; Schuman et al., 2012; Tessler et al, 2004) establishes the period from 18 to 25 years old as the most important in political socialization: it includes the end of school, university (when it corresponds), the right to vote, and first experiences of employment. Following the work of Tessler et al. (2004), we defined the cohorts as individuals that lived the same historical period at least 6 years of the 8 between 18 and 25 years. All others (e.g., those that lived three in one period and three in another) were excluded from the analysis, so as to make a clearer comparison. Based on the political history of Chile between 1973 and 2011 described before, we measured four cohorts: Rise of the dictatorship (impressionable years: 1974–1981; 16.1% of the sample). A period characterized by strong political control, systematic violation of human rights, and enforced prohibition of public meetings or demonstrations. Decline of the dictatorship (impressionable years: 1982–1989; 19.0% of the sample). A period of massive demonstrations against military rule, characterized by the reemergence of political dissent that culminated with the 1988 plebiscite and the 1989 election. Transition to democracy (impressionable years: 1990–2000; 30.2% of the sample). Characterized as a period of democratic transition, when both political elites and citizens feared a return of military rule and, thus, open political conflict was explicitly avoided by promoting so-called “consensus politics.” The new democracy (impressionable years: 2001–2011; 34.7% of the sample). During this period, democratic rule was consolidated, there was a reemergence of citizen participation, and confrontation within political elites increased. In addition to measuring the impact of the macro-political context, we also measured a more immediate source of political socialization: the patterns of communication of survey participants with their parents and teachers during childhood, before their “impressionable years.” Specifically, we measured authoritarian communication (Rojas, 2012; Rojas & Hopke, 2010), a construct gauged with two questions asking about how easy it was for the respondents, as children, to express their disagreements with their parents and teachers, respectively. We used a scale of 1–5, where 1 meant “very easy” and 5 “very difficult” (M = 2.94; SD = 1.38; inter-item r = .40). To control for potential confounds, all analyses included as covariates the following (for the use of these controls, see Hayes et al., 2005a, 2005b; Gil de Zuñiga, Puig-I-Abril & Rojas, 2009; Moy & Scheufele, 2000): interest in politics (measured with two questions: interest in political news and interest in talking politics with family and friends; M = 2.71; SD = 1.14; inter-item r = .62), gender (52.8% women), and educational attainment (primary = 10.9%; high school = 49.5%; and more than high school = 39.7%; M = 2.18; SD = 0.70). Considering that self-censorship is more likely among those who during the Pinochet years identified with the left, we also controlled for ideological self-placement (left-wing = 24.1%). Results To test the hypothesis that political generation predicts WtSC, we conducted an analysis of covariance (ANCOVA). This analysis allows us to determine whether there are statistically significant differences between generations on WtSC, controlling for demographic variables, authoritarian communication, and political predispositions. The results revealed that generation is, indeed, a significant predictor of WtSC, F(3,764) = 15.69, p < .001. As shown in Table 1, the group that experienced the most repressive stage of the dictatorship in its formative years exhibits a higher level of WtSC (adjusted M = 2.88) than the generation that follows immediately (adjusted M = 2.36), which came of age during the decline of the dictatorship and the revival of political mobilization. On the other hand, the generation that lived its impressionable years during the tense democratic transition—when many feared a return of military rule—shows higher levels of WtSC (adjusted M = 2.71) than both previous and subsequent generations. In fact, its mean levels of WtSC are equivalent to the “rise of the dictatorship” generation. Finally, those who came of age in the historical context of a mature democracy with full civilian control of the military exhibit lower levels of WtSC (adjusted M = 2.22) than those belonging to the “transition to democracy” generation. In substantive terms, the “new democracy” cohort is nearly 17% less disposed to self-censor than the “rise of the dictatorship” cohort. Table 1 Marginal Means of WtSC as a Function of Political Generation (ANCOVA) Cohort membership WtSC Adjusted M (SE) N Rise of the dictatorship (1974–1981) 2.88a 0.10 127 Decline of the dictatorship (1982–1989) 2.36b 0.09 139 Transition to democracy (1990–2000) 2.71a 0.07 248 The new democracy (2001–2011) 2.22b 0.06 287 Cohort membership WtSC Adjusted M (SE) N Rise of the dictatorship (1974–1981) 2.88a 0.10 127 Decline of the dictatorship (1982–1989) 2.36b 0.09 139 Transition to democracy (1990–2000) 2.71a 0.07 248 The new democracy (2001–2011) 2.22b 0.06 287 Note: Means sharing a superscript letter do not differ significantly (at p < .05 or less) after a correction for multiple tests using Sidak’s method. Adjusted M = adjusted or estimated marginal mean; SE = standard error; n = observations. Data are weighted. ANCOVA = analysis of covariance; WtSC = willingness to self-censor. Table 1 Marginal Means of WtSC as a Function of Political Generation (ANCOVA) Cohort membership WtSC Adjusted M (SE) N Rise of the dictatorship (1974–1981) 2.88a 0.10 127 Decline of the dictatorship (1982–1989) 2.36b 0.09 139 Transition to democracy (1990–2000) 2.71a 0.07 248 The new democracy (2001–2011) 2.22b 0.06 287 Cohort membership WtSC Adjusted M (SE) N Rise of the dictatorship (1974–1981) 2.88a 0.10 127 Decline of the dictatorship (1982–1989) 2.36b 0.09 139 Transition to democracy (1990–2000) 2.71a 0.07 248 The new democracy (2001–2011) 2.22b 0.06 287 Note: Means sharing a superscript letter do not differ significantly (at p < .05 or less) after a correction for multiple tests using Sidak’s method. Adjusted M = adjusted or estimated marginal mean; SE = standard error; n = observations. Data are weighted. ANCOVA = analysis of covariance; WtSC = willingness to self-censor. Regarding the control variables, most of them relate to WtSC in ways consistent with prior research (Hayes et al., 2005a, 2005b). In Table 2, an analysis of estimated slope coefficients from the ANCOVA shows that education has a negative relationship with WtSC (b = −0.45, p < .001). Although the association between authoritarian communication and WtSC is positive—in line with Rojas (2012) and Rojas and Hopke (2010)—it is somewhat surprising its rather weak relationship (b = 0.05, p = .059). We will come back to this specific result in the discussion section. While political interest was found to be negatively related to WtSC, the relationship is rather weak as well (b = −0.06, p = .088). Finally, neither sex (b = 0.07, p = .313) nor identifying with a leftist ideology (b = −0.10, p = .258) is significantly related to WtSC once all cohorts are included in the estimation. Thus, while not of main interest in the current study, the analysis of the control variables is consistent with extant research on WtSC, which provides reassurance that our results are not idiosyncratic. Table 2 Estimated Slope Coefficients of Covariates on WtSC Ordinary Least Squares (OLS) b (SE) Rise of the dictatorship (1974–1981) (Reference category) Decline of the dictatorship (1982–1989) −0.52*** 0.13 Transition to democracy (1990–2000) −0.16 0.12 The new democracy (2001–2011) −0.66*** 0.12 Authoritarian communication 0.05+ 0.03 Gender (female) −0.07 0.07 Education −0.45*** 0.06 Political interest −0.06+ 0.03 Political identification (left) −0.10 0.09 Constant 3.85*** 0.19 b (SE) Rise of the dictatorship (1974–1981) (Reference category) Decline of the dictatorship (1982–1989) −0.52*** 0.13 Transition to democracy (1990–2000) −0.16 0.12 The new democracy (2001–2011) −0.66*** 0.12 Authoritarian communication 0.05+ 0.03 Gender (female) −0.07 0.07 Education −0.45*** 0.06 Political interest −0.06+ 0.03 Political identification (left) −0.10 0.09 Constant 3.85*** 0.19 Note: Cells report unstandardized (b) OLS regression coefficients with standard errors (SE) in parentheses. Data are weighted. R2 = .20. F(8,764) = 24.61***. WtSC = willingness to self-censor. +p < .10; * p < .05; ** p < .01; *** p < .001. Table 2 Estimated Slope Coefficients of Covariates on WtSC Ordinary Least Squares (OLS) b (SE) Rise of the dictatorship (1974–1981) (Reference category) Decline of the dictatorship (1982–1989) −0.52*** 0.13 Transition to democracy (1990–2000) −0.16 0.12 The new democracy (2001–2011) −0.66*** 0.12 Authoritarian communication 0.05+ 0.03 Gender (female) −0.07 0.07 Education −0.45*** 0.06 Political interest −0.06+ 0.03 Political identification (left) −0.10 0.09 Constant 3.85*** 0.19 b (SE) Rise of the dictatorship (1974–1981) (Reference category) Decline of the dictatorship (1982–1989) −0.52*** 0.13 Transition to democracy (1990–2000) −0.16 0.12 The new democracy (2001–2011) −0.66*** 0.12 Authoritarian communication 0.05+ 0.03 Gender (female) −0.07 0.07 Education −0.45*** 0.06 Political interest −0.06+ 0.03 Political identification (left) −0.10 0.09 Constant 3.85*** 0.19 Note: Cells report unstandardized (b) OLS regression coefficients with standard errors (SE) in parentheses. Data are weighted. R2 = .20. F(8,764) = 24.61***. WtSC = willingness to self-censor. +p < .10; * p < .05; ** p < .01; *** p < .001. Still, it could be argued that cohort effects are an artifact of defining political generations based on the historical context. Thus, we replicated the prior analysis using what Tessler and colleagues (2004) labeled “context free” cohort categories. The same respondents included in Table 1 (N= 801) were now divided into eight 5-year age groups, beginning with those who were 20–24 years old at the time of the survey, followed by those who were 25–29 years old, and so on through those who were 55–59 years old. Each age group was dummy coded and then entered simultaneously into a linear regression with the first group as the reference category, plus the controls. The fit statistics of this alternative model (F(12,760) = 16.09, p < .001, adjusted R2 = .190, Root Mean Square Error (RMSE) = 1.013, Akaike Information Criterion (AIC) = 2,226.35, Bayesian Information Criterion (BIC) = 2,286.81) are clearly worse than those of the proposed model (F(8,764) = 24.61, p < .001, adjusted R2 = 0.197, RMSE = 1.009, AIC = 2,216.09, BIC = 2,257.94). Furthermore, the results of this ahistorical model show that only when there was a complete overlap between age group and cohort membership did a statistically significant relationship arise. For instance, those in the 40–44 years of age group (all of them members of the transition to democracy cohort) exhibited higher levels of WtSC than those in the reference category (b = 0.63, p < .001). However, those who were in the immediately older group, between 45 and 49 years (i.e., all of them members of the decline of the dictatorship cohort) scored as low in the WtSC scale as those in the 20–24 years of age group, the reference category (b = 0.07, p = .635). Last, it could be argued that models that assume a continuous dependent variable, such as ANCOVA and OLS, are inappropriate given the ordinal nature of the items used to create the WtSC scale. Nevertheless, when we conduct an ordinal logit regression, the results are not contradictory with any of the key findings. Discussion This study analyzes a relevant individual-level disposition for the proper functioning of democracy: WtSC, which refers to citizens’ propensity to suppress the expression of opinions in front of a hostile audience to those opinions. Democracy not only demands freedom of expression but also citizens who are willing to exercise this freedom and participate in the debate of public affairs through the expression of their points of view, both when supported by the majority and when they represent minority viewpoints. The value of expression of different—if not disagreeing—perspectives is highlighted by numerous authors. Arendt (1968) proposed that if countries should be governed by those who are elected by their citizens, citizens should be able to form their opinions and base their decisions in the presence of a public debate that effectively includes all possible visions of reality and public affairs. Disagreement is particularly valued among deliberative democracy theorists, who expect that free discussions between contradicting opinions will foster intra and interpersonal reflection, raising the quality of a polity’s public sphere (Habermas, 1996). John Stuart Mill (1859) is profusely quoted to expose the benefits of discussions between opposite positions, which include learning normatively superior points of view, and allowing citizens to be persuaded to modify their opinions. The specific results of the current study are consistent with the notion that the sociopolitical context in which individuals lived during their “impressionable years” (18–25 years) is a significant antecedent of WtSC. Our statistical analysis suggests that Chileans who spent most of their early adulthood in a relatively repressive political context, like in the rise of the dictatorship and in the transition to democracy periods, continue to show—decades later—higher levels of WtSC than Chileans who came of age in less repressive contexts, such as the decline of the dictatorship and the new democracy periods. Thus, we find that the period from youth to first adulthood, which usually coincides in part with university years, is influential, in line with previous studies that test the generational hypothesis. For example, Alwin, Cohen, and Newcomb (1991) demonstrated that the political attitudes acquired during the university are maintained for the rest of life. Lup (2015) showed that people from countries with undemocratic backgrounds are less willing to engage in informal political conversations, and Schuman and Corning (2006; Schuman et al., 2012) found that national and international events in critical years have a disproportionate effect on people’s political attitudes. Still, finding that WtSC is higher among those who lived their impressionable years during the most brutal period of Pinochet’s dictatorship is to be expected. The same is true with the finding that the youngest cohort in our study shows the lowest levels of WtSC; this generation came of age in a mature democratic environment. More remarkable are the results showing that the transition to democracy generation is as predisposed to self-censorship as the rise of the dictatorship generation. To understand this result, it must be noted that the 1990s was a period of enormous political tension in Chile. Unlike what happened in Spain and other “third wave” democracies, the elected authorities of the Aylwin and Frei administrations had to coexist with a former dictator who maintained a significant share of political power, both within and outside the Chilean army. For instance, the Constitution—written and approved by the military junta in 1980—allowed Pinochet to remain commander-in-chief and, after retirement, senator for life. Thus, both government and significant portions of the electorate experienced institutional fragility and fear of a return to military rule. To prevent such an outcome, the civic-led governments of the transitional period sought to project stability by means of promoting political agreement among all democratic forces, discourage open political dissent, and demobilize the social movements that had fought the dictatorship in the 1980s (Paley, 2001, p. 91). In this context, a disposition to self-censor political opinions is more likely to arise. Another noteworthy finding relates to the relationship between WtSC and authoritarian communication. Our analysis shows that generational memory, rather than authoritarian communication, has a significant association with WtSC. Considering that both variables measure an individual’s socialization process, albeit at different stages of the life cycle, this finding could be interpreted as socialization at the macro level—represented by generations—having a larger influence on WtSC than socialization at the more micro level of home and school. Alternatively, there is the possibility that authoritarian communication represents a dimension of WtSC. For instance, it could well be that a student who would rather censor her nonconforming ideas than complain to a school teacher would already have a personality that tends toward WtSC and may not have an authoritarian instructor. Nevertheless, our data do not support this conclusion: the relationship between these two variables is rather weak (r = .17, p < .001), and the antecedents of each seem to be different. While gender and political ideology have stronger relationships with authoritarian communication than with WtSC, the opposite is true for education. It would be desirable for future research to probe deeper on the nature of both authoritarian communication and WtSC. However it may be, the study suggests that just as important as studying the immediate contexts of individuals to explain dispositions such as WtSC, the collective memories of a generation may also play a vital role in shaping political communication behaviors. There is also a strong relationship between the educational level and WtSC. Belonging to a more educated group is linked to a higher willingness to participate in public discussion. The importance of socio-economic variables is consistent with other studies conducted in Chile that explore issues related to political participation, both electoral and nonelectoral (Scherman & Arriagada 2012; Scherman, Arriagada, & Valenzuela, 2015). The relevance showed by this variable—not found in studies on WtSC carried out in other Latin American countries (Rojas & Hopke, 2010) may be related to economic inequality (Etchegaray & Matus, 2015). Although the inequality level in Chile is similar to that of Brazil, Guatemala, and Panama, it is still the highest among the 35 Organisation for Economic Co-operation and Development countries (OECD, 2016, p. 111). It is important to address a number of limitations of the study. The data were obtained from a representative, cross-sectional study. As such, it does not allow for firmly establishing causal–effect relationships between the variables. For instance, the same generational groups we studied would also have experienced similar childhoods and educational systems, so the consistencies within groups found in this study cannot be confidently attributed only to the political and historical events taking place during the “impressionable years” of respondents. A more robust design would be an age-period-cohort analysis. While this kind of analysis cannot be conducted with variables that have not been asked repeatedly in surveys, such as WtSC, proxy variables (e.g., political discussion frequency) could be subjected to this type of longitudinal research. Relatedly, examining mediating and moderating variables of the relationship between generational memory and WtSC could provide an additional route for future research. It would also be advisable for future investigations to consider psychological variables or personality traits that relate to opinion expression inhibition in the presence of a perceived hostile audience. Limitations notwithstanding, the study represents progress on three accounts. First, the analysis identifies generational memory as a hitherto unknown, yet important, antecedent of WtSC. This brings new data to understand why different social groups differ in important dispositions related to political expression. Second, the study extends the application of the “impressionable years” hypothesis to key provisions of political communication, such as WtSC. To date, most work in public opinion and collective memory relates to political attitudes and behaviors, not psychological dispositions. Third, by testing for cohort effects on WtSC, we show a way of connecting microlevel and macrolevel variables in the analysis of individual-level differences. Future research should replicate and extend our work to other dispositions that may be influenced by generational memory. Nicolle Etchegaray is an assistant professor in the school of Communications at Diego Portales University. She is also a doctoral student on Sociology, in the School of Sociology at Pontificia Universidad Católica de Chile (Catholic University of Chile). Although his doctoral thesis is associated with the area of political communication, his research areas include media effects, formation of public opinion and the role of the media in the formation and increase of social stereotypes. Andrés Scherman is an associate professor in the School of Communications at Diego Portales university, where he also directs the International Master in Communication. His research is concerned with audience analysis, political communication, digital social network use, and the influence of interactive technologies on political knowledge, attitudes, and behaviors. Sebastián Valenzuela is an associate professor in the School of Communications at Pontificia Universidad Católica de Chile (Catholic University of Chile), where he teaches and does consulting work on social media, journalism, and political communication. He is also an associate researcher at Chile's National Research Center for Integrated Natural Disasters Management (CIGIDEN/Fondap/15110017) and the Millennium Institute for Foundational Research on Data, as well as editor-in-chief of Cuadernos.info, a peer-reviewed, open-access scientific journal that publishes research on communication and media in Latin America. His research is concerned with the role of the news media, social media and ICTs on public opinion formation—the broad effects of journalism and interactive technologies on political knowledge, attitudes, and behaviors. Appendix 1 Table A1 Percentage Distribution of Selected Characteristics Initial sample (N = 1,737) Final sample (N = 801) Benchmark survey Gender: Male 47.9 47.2 49.0 Female 52.1 52.8 51.0 Educational attainment: High school or less 66.1 64.8 69.8 Some college 19.1 20.8 15.3 Bachelor’s degree or higher 14.8 14.4 14.8 Political interest: Low 38.7 51.3 43.9 Moderate 13.5 10.8 27.0 High 47.8 37.9 29.1 Political identification: Left-wing 28.6 24.1 18.6 Authoritarian communication: Low 52.1 42.1 N/A Moderate 10.3 15.1 N/A High 37.7 42.8 N/A Initial sample (N = 1,737) Final sample (N = 801) Benchmark survey Gender: Male 47.9 47.2 49.0 Female 52.1 52.8 51.0 Educational attainment: High school or less 66.1 64.8 69.8 Some college 19.1 20.8 15.3 Bachelor’s degree or higher 14.8 14.4 14.8 Political interest: Low 38.7 51.3 43.9 Moderate 13.5 10.8 27.0 High 47.8 37.9 29.1 Political identification: Left-wing 28.6 24.1 18.6 Authoritarian communication: Low 52.1 42.1 N/A Moderate 10.3 15.1 N/A High 37.7 42.8 N/A Notes: Benchmark survey is Centro de Estudios Públicos (CEP’s) 64th survey, a face-to-face, nationally representative survey (N = 1,554) conducted in June–July 2011 (available at: https://www.cepchile.cl/estudio-nacional-de-opinion-publica-junio-julio-2011-incluye-tema/cep/2016-03-04/095600.html). N/A = not available. Initial sample (N = 1,737) Final sample (N = 801) Benchmark survey Gender: Male 47.9 47.2 49.0 Female 52.1 52.8 51.0 Educational attainment: High school or less 66.1 64.8 69.8 Some college 19.1 20.8 15.3 Bachelor’s degree or higher 14.8 14.4 14.8 Political interest: Low 38.7 51.3 43.9 Moderate 13.5 10.8 27.0 High 47.8 37.9 29.1 Political identification: Left-wing 28.6 24.1 18.6 Authoritarian communication: Low 52.1 42.1 N/A Moderate 10.3 15.1 N/A High 37.7 42.8 N/A Initial sample (N = 1,737) Final sample (N = 801) Benchmark survey Gender: Male 47.9 47.2 49.0 Female 52.1 52.8 51.0 Educational attainment: High school or less 66.1 64.8 69.8 Some college 19.1 20.8 15.3 Bachelor’s degree or higher 14.8 14.4 14.8 Political interest: Low 38.7 51.3 43.9 Moderate 13.5 10.8 27.0 High 47.8 37.9 29.1 Political identification: Left-wing 28.6 24.1 18.6 Authoritarian communication: Low 52.1 42.1 N/A Moderate 10.3 15.1 N/A High 37.7 42.8 N/A Notes: Benchmark survey is Centro de Estudios Públicos (CEP’s) 64th survey, a face-to-face, nationally representative survey (N = 1,554) conducted in June–July 2011 (available at: https://www.cepchile.cl/estudio-nacional-de-opinion-publica-junio-julio-2011-incluye-tema/cep/2016-03-04/095600.html). N/A = not available. References Alwin D. F. , Cohen R. L. , Newcomb T. M. ( 1991 ). Political attitudes over the life span: The Bennington women after fifty years . Madison, WI : University of Wisconsin Press . Alwin D. F. , Krosnick J. A. ( 1991 ). Aging, cohorts, and the stability of sociopolitical orientations over the life span . American Journal of Sociology , 97 , 169 – 195 . doi: 10.1086/229744 Google Scholar CrossRef Search ADS Arendt H. ( 1968 ). Imperialism: Part two of the origins of totalitarianism . Florida : Mariner Books . Chambers S. ( 2003 ). Deliberative democratic theory . Annual Review of Political Science , 6 , 307 – 326 . doi: 10.1146/annurev.polisci.6.121901.085538 Google Scholar CrossRef Search ADS Corning A. , Schuman H. ( 2015 ). Generations and collective memory . Chicgo, IL : The University of Chicago Press . Google Scholar CrossRef Search ADS Dahl R. ( 1989 ). Democracy and its critics . New Haven, CT : Yale University Press . Delli-Carpini M. X. , Cook F. L. , Jacobs L. R. ( 2004 ). Public deliberation, discursive participation, and citizen engagement: A review of the empirical literature . Annual Review of Political Science , 7 , 315 – 344 . Retrieved from https://doi.org/10.1146/annurev.polisci.7.121003.091630 Google Scholar CrossRef Search ADS Donsbach W. , Salmon C. T. , & Tsfati Y. (Eds.). ( 2014 ). The spiral of silence: New perspectives on communication and public opinion . New York, NY : Routledge . Dennis J. , Easton D. ( 1969 ). Children in the political system: Origins of political legitimacy . New York, NY : McGraw-Hill Book Company . Etchegaray N. , Matus A. ( 2015 ). Evolution of poverty coverage between 2005 and 2014: What has changed and what hasn’t in primetime television newscasts in Chile . Cuadernos.info , 36 , 53 – 69 . doi:10.7764/cdi.36.727 Google Scholar CrossRef Search ADS Fishkin J. ( 1995 ). The voice of the people . New Haven, CT : Yale University Press . Fleet N. , Guzmán‐Concha C. ( 2017 ). Mass higher education and the 2011 student movement in Chile: Material and ideological implications . Bulletin of Latin American Research , 36 , 160 – 176 . doi:10.1111/blar.12471 Google Scholar CrossRef Search ADS Gil de Zuñiga H. , Puig-I-Abril E. , Rojas H. ( 2009 ). Weblogs, traditional sources online and political participation: An assessment of how the Internet is changing the political environment . New Media and Society , 11 , 553 – 574 . doi: 10.1177/1461444809102960 Google Scholar CrossRef Search ADS Glynn C. J. , Hayes A. F. , Shanahan J. ( 1997 ). Perceived support for one's opinions and willingness to speak out: A meta-analysis of survey studies on the “spiral of silence” . Public Opinion Quarterly , 61 , 452 – 463 . Retrieved from http://www.jstor.org/stable/2749581. Google Scholar CrossRef Search ADS Glynn C. J. , Huge M. E. ( 2014 ). Speaking in spirals: An updated meta-analysis of the spiral of silence. In Donsbach W. , Salmon C. T. , & Tsfati Y. (Eds.), The spiral of silence. New perspectives on communication and public opinion (pp. 65 – 72 ). New York, NY : Routledge . Griffin L. J. ( 2004 ). “Generations and collective memory” revisited: Race, region, and memory of civil rights . American Sociological Review , 69 , 544 – 557 . doi: 10.1177/000312240406900404 Google Scholar CrossRef Search ADS Habermas J. ( 1996 ). Between facts and norms: Contributions to a discourse theory of law and democracy . Cambridge, MA : The MIT Press . Hayes A. F. ( 2007 ). Exploring the forms of self‐censorship: On the spiral of silence and the use of opinion expression avoidance strategies . Journal of Communication , 57 , 785 – 802 . doi: 10.1111/j.1460-2466.2007.00368.x Google Scholar CrossRef Search ADS Hayes A. F. , Glynn C. J. , Shanahan J. ( 2005a ). Validating the willingness to self-censor scale: Individual differences in the effect of the climate of opinion on opinion expression . International Journal of Public Opinion Research , 17 , 443 – 455 . doi: 10.1093/ijpor/edh072 Google Scholar CrossRef Search ADS Hayes A. F. , Glynn C. J. , Shanahan J. ( 2005b ). Willingness to self-censor: A construct and measurement tool for public opinion research . International Journal of Public Opinion Research , 17 , 298 – 323 . doi: 10.1093/ijpor/edh073 Google Scholar CrossRef Search ADS Jennings M. K. ( 1996 ). Political knowledge over time and across generations . Public Opinion Quarterly , 60 , 228 – 252 . doi: 10.1086/297749 Google Scholar CrossRef Search ADS Kim S. H. , Han M. , Shanahan J. , Berdayes V. ( 2004 ). Talking on ‘Sunshine in North Korea’: A test of the spiral of silence as theory of powerful mass media . International Journal of Public Opinion Research , 16 ( 1 ), 39 – 62 . doi: 10.1093/ijpor/16.1.39 Google Scholar CrossRef Search ADS Lasorsa D. L. ( 1992 ). Policymakers and the third-person effect. In Kennamer J. D. (Ed.), Public opinion, the press, and public policy (pp. 163 – 175 ). Westport, CT : Praeger . Lup O. ( 2015 ). Informal political conversation in old and new democracies in a comparative perspective . International Journal of Public Opinion Research , 27 , 497 – 508 . doi: 10.1093/ijpor/edv039 Mannheim K. ( 1952 ). The problem of generations. In Essays on the sociology of knowledge (pp. 276 – 320 ). New York, NY : Routledge & Kegan Paul . Matthes J. , Hayes A. F. , Rojas H. , Shen F. , Min S. J. , Dylko I. B. ( 2012 ). Exemplifying a dispositional approach to cross-cultural spiral of silence research: Fear of social isolation and the inclination to self-censor . International Journal of Public Opinion Research , 24 , 287 – 305 . doi: 10.1093/ijpor/eds015 Google Scholar CrossRef Search ADS Matthes J. , Rios Morrison K. , Schemer C. ( 2010 ). A spiral of silence for some: Attitude certainty and the expression of political minority opinions . Communication Research , 37 , 774 – 800 . doi:10.1177/0093650210362685 Google Scholar CrossRef Search ADS Mill J. S. ( 1859/1989 ). On liberty and other writings . Stefan C. (Ed.) New York, NY : Cambridge University Press . Moy P. , Scheufele D. A. ( 2000 ). Media effects on political and social trust . Journalism & Mass Communication Quarterly , 77 , 744 – 759 . doi: 10.1177/107769900007700403 Google Scholar CrossRef Search ADS Noelle‐Neumann E. ( 1974 ). The spiral of silence a theory of public opinion . Journal of Communication , 24 , 43 – 51 . doi: 10.1111/j.1460-2466.1974.tb00367.x Google Scholar CrossRef Search ADS OECD (Organisation for Economic Co-operation and Development) (2016). Society at a Glance 2016: OECD Social Indicators, OECD Publishing . Paris, doi: dx.doi.org/10.1787/9789264261488-en. Osborne D. , Sears D. O. , Valentino N. A. ( 2011 ). The end of the solidly democratic South: The impressionable‐years hypothesis . Political Psychology , 32 , 81 – 108 . doi: 10.1111/j.1467-9221.2010.00796.x Google Scholar CrossRef Search ADS Oshagan H. ( 1996 ). Reference group influence on opinion expression . International Journal of Public Opinion Research , 8 , 335 – 354 . doi: 10.1093/ijpor/8.4.335 Google Scholar CrossRef Search ADS Paley J. ( 2001 ). Marketing democracy: Power and social movements in post-dictatorship Chile . Berkeley, CA : University of California Press . Rojas H. ( 2012 ) El efecto de la comunicación autoritaria sobre la expresión política en Chile [The effect of authoritarian communication on political expression in Chile]. In Scherman A. (Ed.), Jóvenes, participación y medios 2011 [Youth, participation, and media 2011] (pp. 32 – 39 ). Santiago, Chile : CIP/UDP . Rojas H. ( 2006 ). Communication, participation and democracy . Universitas Humanística , 62 , 109 – 142 . Rojas F. , Hopke J. ( 2010 ). Socializados para la autocensura: Comunicación autoritaria y opinión pública [Socialized for self-censorship: Authoritarian communication and public opinión] . Revista Latinoamericana de Opinión Pública , 1 , 165 – 173 . Scherman A. , Arriagada A. , Valenzuela S. ( 2015 ). Student and environmental protests in Chile: The role of social media . Politics , 35 , 151 – 171 . doi: 10.1111/1467-9256.12072 Google Scholar CrossRef Search ADS Scherman A. , Arriagada A. , Valenzuela S. ( 2016 ). ¿Qué hemos aprendido de la encuesta “Jóvenes, participación y consumo de medios”?: Una revisión entre 2009 y 2012 [What have we learned from the Youth, Participation and Media Consumption surveys? A review from 2009 to 2012]. In Muñiz C. , Saldierna A. R. , Marañón F. J. , & Maldonado L. (Eds.), Estudios sobre comunicación política en Latinoamérica [Studies on political communication in Latin America] (pp. 253 – 276 ). Mexico, DF : Fontamara/UANL . Scherman A. , Arriagada A. ( 2012 ). Jóvenes, postmaterialismo y consumo en medios [Youth, postmaterialism, and media consumption]. In Scherman A. (Ed.), Jóvenes, participación y medios 2011 [Youth, participation, and media 2011] (pp. 8 – 17 ). Santiago, Chile : CIP/UDP . Schuman H. , Corning A. D. ( 2006 ). Comparing Iraq to Vietnam: Recognition, recall, and the nature of cohort effects . Public Opinion Quarterly , 70 , 78 – 87 . doi: 10.1093/poq/nfj010 Google Scholar CrossRef Search ADS Schuman H. , Corning A. D. ( 2017 ). The conversion of generational effects into collective memories . International Journal of Public Opinion Research , 29 , 520 – 532 . doi: 10.1093/ijpor/edw012 Schuman H. , Corning A. , Schwartz B. ( 2012 ). Framing variations and collective memory . Social Science History , 36 , 451 – 472 . doi: 10.1215/01455532-1717127 Tessler M. , Konold C. , Reif M. ( 2004 ). Political generations in developing countries: Evidence and insights from Algeria . Public Opinion Quarterly , 68 , 184 – 216 . doi: 10.1093/poq/nfh011 Google Scholar CrossRef Search ADS Valenzuela S. ( 2011 ). Autocensura y ciudadanía [Self-censorship and citizenship]. In Rojas H. , Orozco M. M. , Gil de Zúñiga H. , & Wojcieszak M. (Eds.), Comunicación y ciudadanía [Communication and citizenship] (pp. 153 – 171 ). Bogotá, Colombia : Universidad Externado de Colombia . Valenzuela S. , Brandão A. S. ( 2015 ). Historical dramas, current political choices: Analyzing partisan selective exposure with a docudrama . Mass Communication and Society , 18 , 449 – 470 . doi: 10.1080/15205436.2015.1016178 Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved. 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)
International Journal of Public Opinion Research – Oxford University Press
Published: Apr 17, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera