Applying the theory of planned behavior to self-report dental attendance in Norwegian adults through structural equation modelling approach

Applying the theory of planned behavior to self-report dental attendance in Norwegian adults... Background: Understanding factors that affect dental attendance behavior helps in constructing effective oral health campaigns. A socio-cognitive model that adequately explains variance in regular dental attendance has yet to be validated among younger adults in Norway. Focusing a representative sample of younger Norwegian adults, this cross-sectional study provided an empirical test of the Theory of Planned Behavior (TPB) augmented with descriptive norm and action planning and estimated direct and indirect effects of attitudes, subjective norms, descriptive norms, perceived behavioral control and action planning on intended and self-reported regular dental attendance. Method: Self-administered questionnaires provided by 2551, 25–35 year olds, randomly selected from the Norwegian national population registry were used to assess socio-demographic factors, dental attendance as well as the constructs of the augmented TPB model (attitudes, subjective norms, descriptive norms, intention, action planning). A two-stage process of structural equation modelling (SEM) was used to test the augmented TPB model. Results: Confirmatory factor analysis, CFA, confirmed the proposed correlated 6-factor measurement model after re-specification. SEM revealed that attitudes, perceived behavioral control, subjective norms and descriptive norms explained intention. The corresponding standardized regression coefficients were respectively (β = 0.70), (β =0.18), (β = − 0.17) and (β =0.11) (p < 0.001). Intention (β =0.46) predicted action planning and action planning (β =0.19) predicted dental attendance behavior (p < 0.001). The model revealed indirect effects of intention and perceived behavioral control on behavior through action planning and through intention and action planning, respectively. The final model explained 64 and 41% of the total variance in intention and dental attendance behavior. Conclusion: The findings support the utility of the TPB, the expanded normative component and action planning in predicting younger adults’ intended- and self-reported dental attendance. Interventions targeting young adults’ dental attendance might usefully focus on positive consequences following this behavior accompanied with modeling and group performance. Keywords: Dental attendance, Young adults, Theory of planned behavior, Structural equation modelling, AMOS * Correspondence: Anne.Aastrom@uib.no Oral health Centre of Expertise in Western Norway, Bergen, Hordaland, Norway Department of Clinical Dentistry, Faculty of Medicine, University of Bergen, PO Box 7804, N-5020 Bergen, Norway © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Åstrøm et al. BMC Oral Health (2018) 18:95 Page 2 of 10 Background between attitudes and behavior mediated by behavioral The importance of dental attendance is a key measure in intentions. Intention directly influences behavior and is health education as regular attendance is associated with shaped by attitudes, subjective norms and perceived good health and well-being. In Norway, the Public behavioral control regarding the behavior. Empirical Dental Service (PDS) is financed by taxes and provides validations of the TPB have revealed that the model free dental care to children and adolescents until 20 years reliably explains 40–50% of the variance in intention and of age [1, 2]. The private dental services provides dental that intention explains between 20 and 40% of the care to the general adult population and is organized variance in actual behavior [14–16]. according to market mechanisms, with dental fees In spite of its predictive success, TPB has been determined by supply and demand and with very limited criticized for its validity and it has been shown that private or public insurance arrangements [3]. Regardless other variables explain considerable proportions of the of the disparities in dental coverage between Norway variance in intention and behavior [17]. Moreover, de- and other Scandinavian countries, dental attendance scriptive norms and action planning have shown residual rates have been high among Norwegian adults. About 80 effects on intention and behaviour after consideration of and 77% of Norwegian adults above 20 years of age re- the original TPB variables [17, 18]. Evidence suggests ported having visited a dentist during the last 12 months that adding action planning would improve the predic- in 2008 and 2013, respectively [4–6]. tion of behavior from the TPB [18, 19]. Thus, formation Nevertheless, as in other Scandinavian countries, the of action plans can be used to promote the realization of prevalence of regular dental care utilization among Nor- desired outcomes [18, 19]. The role of subjective norms wegian adults varies according to age, period and within the TPB has also been considered [20, 21]. socio-economic status, being smallest in the younger Subjective norms have been criticized for its narrow age- and the lower income groups [4–7]. Støle et al. [8] conceptualization, focusing what significant people found that 87% of 23–24 year old Norwegian adults thinks others ought to do, neglecting descriptive visited a dentist every second year in 1983, whereas the norms referring to what significant others themselves corresponding figures in 1994 was 85%. Among 25 year actually do [20–22]. Descriptive norms correlate with old Norwegian adults, 62 and 44% reported dental behavioral intentions and have shown to be among attendance once a year in 1997 and 2007, respectively the strongest correlates of physical activity [23]. The [9]. According to the Official Statistics of Norway, the TPB has received considerable empirical support prevalence of having visited public dental health care across health- and social behaviors, including oral services during last year was highest among 45–66 year hygiene behaviors and health screening [15]. However, olds and lowest among 21–24 year olds in 2016 [10]. to our knowledge, only one previous study has exam- Previous studies have identified enabling factors, such as ined use of public dental services in the context of cost of treatment and dental anxiety, as important TPB [16]. barriers towards regular use of dental care [11, 12]. In a Focusing a representative sample of young Norwe- recent population-based study of Swedish adults, finan- gian adults 25–35 years old, this study provides an cial problems and lack of social support were associated empirical test of the TPB augmented with descriptive with refraining from seeking dental care [11]. Whereas norm and action planning and estimates direct and socio-demographic- and need related factors are import- indirect effects of attitudes, subjective norms, descrip- ant covariates of dental care utilization, relatively few tive norms, perceived control, and action planning on studies have considered modifiable socio-cognitive intended and self-reported regular dental attendance. determinants in the younger adult populations. Based on the conceptualization of the TPB and previ- Influencing younger adults’ adherence to continued ous empirical support it was hypothesized that the re- dental attendance requires understanding of the sponses to 16 observed indicator variables could be socio-cognitive factors underlying their decision to explained by 6 latent factors in terms of attitudes, comply with advices for regular dental care. A subjective norms, descriptive norms, perceived con- socio-cognitive model that adequately explains vari- trol, intention and action planning. Further, it was ance in regular dental attendance has yet to be hypothesized that each indicator would have a stron- validated among younger adults in Norway. ger relation with their corresponding factor than with the competing factors. Finally, it was hypothesized Theoretical approach that attitudes, subjective norm, descriptive norm and The theory of planned behavior, TPB, is a widely applied perceived behavioral control would predict behavioral socio-cognitive model of the attitude–behavior relation- intention and that intention, action planning and per- ship, assuming that most conscious behaviors is rational ceived behavioral control, would predict self-reported and goal directed [13, 14]. TPB proposes a causal link dental attendance. Åstrøm et al. BMC Oral Health (2018) 18:95 Page 3 of 10 Methods measured by two items, e.g. “My friends attend dentist Study design, participants and ethical issues regularly”– with response categories ranging from 1 Thepresent studyuseddata from anelectronic, (strongly disagree) to 5 (strongly agree). Action planning cross-sectional public dental health survey conducted was assessed using the action planning scale adopted in Norway. A representative sample of 9000 adults from Sniehotta et al. [17], including three items, e.g. “I (using individuals as the primary sampling unit) aged have made a detailed plan when to attend, where to 25–35 years was randomly selected from the Norwegian attend and how to attend a dentist regularly in the national population registry in September 2016. Participa- future”. Response categories ranged from 1 (strongly tion was voluntary and anonymous and the return of a disagree) to 5 (strongly agree). Parts of the question- completed questionnaire recognized as the informed naire used in the present study is available in English consent. Ethical permission to carry out the survey was in the Additional file 1. granted by the Ombudsman, Norwegian Center for Research Data (NO.49241). NORSTAT (www.norstat.no) Statistical analyses was responsible for the electronically distributed question- Data were analyzed using SPSS version 22.0 (IBM Corp. naires and for the data collection. An eligible sample of Released 2013, IBM SPSS Statistics for Windows, 9052 adults aged 25–35 years of age received an electronic Armonk NY: IBM Corp). IBM SPSS AMOS 16.0 [24] version of the questionnaire with an introductory letter was used to test the hypothesized augmented TPB explaining the purpose of the study. Total response rate model using a two-step modelling approach whereby the was 29% (2635/9052). Eighty-four respondents were re- measurement model (step1) and the structural model moved due to incomplete questionnaires. All participants (step 2) were constructed separately [25]. First, a who provided complete questionnaires were included in confirmatory factor analysis, CFA, using maximum the present study (n = 2551). likelihood estimation (ML) was conducted to test the adequacy of the measurement model [25]. Modification Measures indices (MI) were used to identify sources of misfit in Dental attendance behavior was measured using one the model. A prerequisite for testing of invariance across question; “How often do you usually visit a dentist?” the structural paths in the full structural model (step 2) is response categories ranged from (1) twice a year or more that the measurement model has configural and metric to (4) more seldom than every second year. Components invariance. Configural invariance was examined by test- of an augmented version of Ajzen’s TPB [13] was mea- ing the fit of the modified correlated measurement sured in terms of attitudes, subjective norms, perceived model separately for males and females and by testing behavioral control, descriptive norm and action planning the fit of an un-constrained multi-group model. Metric in relation to regular dental attendance. In accordance invariance was examined by comparing a multi-group with recommendations from Ajzen [13], each construct model with all factor loadings constrained equal to the was measured considering the four elements of action baseline configural model in which the factor loadings (attending), target (dentist), context (on a regular basis), were free to vary. The models were assumed and time (future) (13) (. Intention to attend a dentist non-invariant if the change in chi square was significant regularly was measured by two items, e.g. “I intend to at- and the decrease in comparative fit index, CFI, was less tend a dentist regularly in the future.” Responses were than 0.001 [26]. indicated on a four-point scale: (1) Strongly disagree, (2) A full structural equation modelling, SEM, (step 2) Disagree, (3) Neither agree nor disagree (4) Agree and examined whether the hypothesized TPB model was ac- (4) Strongly agree. Attitude towards regular dental at- ceptable fit to the present data, testing simultaneously tendance was assessed by four items, e.g. “to attend a the interrelationships specified within the a priori aug- dentist regularly in the future do not make sense to me”. mented TPB model. To assess how adequately the hy- Responses were indicated on a five-point scale ranging pothesized model described the sample data, chi-square from 1 (strongly disagree) to 5 strongly disagree). Sub- test was used together with the following goodness of fit jective norm was measured by three items, e.g. “My par- indices; CFI (Comparative fit index), RMSEA (root mean ents (partner/friend, dentist) want me to attend a dentist square error of approximation) and AIC (Akaike’s regularly in the future”. Responses were indicated on a information criteria). In line with the conventional five-point scale ranging from 1 (strongly disagree) to 5 recommendations of Hu and Bentler [27], a good model (strongly agree). Perceived behavioral control was mea- fit was indicated by a RMSEA less or equal to 0.06, a sured by two items, e.g. “Its totally up to me whether I CFI greater or equal to 0.90 and with a model having attend a dentist regularly in the future”. Responses were lower AIC being the more plausible together with a indicated on a five-point scale ranging from 1 (strongly non-significant Chi square. Statistical significance of disagree) to 5 (strongly agree). Descriptive norm was parameter estimates are the Critical Ration (CR) Åstrøm et al. BMC Oral Health (2018) 18:95 Page 4 of 10 representing the parameter estimate divided by its stand- Table 1 Frequency distribution of participants’ socio-demographic characteristics and dental attendance behavior, (n = 2551) and ard error. Based on a level of 0.05, the test statistics (CR) corresponding percentage figures in the total population needed to be 1.96 before rejection of the null hypothesis. Participants Total population Results Category % (n)% Sample profile Gender Male 43.3 (1105) 51.3 In spite of the relatively low response rate (29%) obtained, the age distribution of the final sample Female 56.7 (1446) 49.0 corresponded with that of the Norwegian population Age 25-29 years 43.7 (1116) 46.3 20–44 years old by December 2016. The age distribution 30–35 years 56.3 (1435) 53.0 of younger (25–29 years) and older (30–35 years) partic- Country of birth Norway 91.5 (2333) ipants were 43.7 and 56.3%, respectively. Corresponding Other Nordic 2.6 (66) figures in the population were respectively 46.3 and 53%. Outside Nordic 6.0 (152) Whereas the gender distribution in the sample was 43% men and 56.7% women, the corresponding population Civil status Single 36.6 (935) distribution was 51.3 and 49.0%. Among the partici- Married 63.4 (1616) pants, 27.3, 38.6 and 34.1% reported respectively, Highest level of Primary/secondary 27.3 (679) 26.5 primary-, bachelor- and college/university level of educa- education Bachelor degree 38.6 (962) 37.8 tion. Corresponding figures in the adult population College/university 34.1 (850) 32.9 16 years and above were 26.5, 37.8 and 32.9%. Among Income (NOK) At least 400.000 19.5 (413) the respondents (n = 2551), 91.5% were of native Norwegian origin. Eight percent confirmed dental 400,001–800,000 41.2 (876) attendance at least twice a year, 47.2% once a year, > 800,000 39.3 (835) 21.2% every second year and 21.2% more seldom than Dental attendance At least twice a 8.0 (203) every second year (Table 1). year Once a year 47.2 (1205) Descriptive statistics of TPB variables Every second year 21.2 (540) Table 2 depicts mean, standard deviation, minimum More seldom 21.2 (540) and maximum scores and theoretical range for each Norwegian population 20–44 years by December 2016 indicator measuring the latent constructs of attitudes, Norwegian population above 20 years by December 2016 subjective norms, perceived control, descriptive norms, intention and action planning. On average the pairs of error terms, resulting from item overlap, or study group demonstrated strong intentions with reflecting biases in responding such as “yea” saying or mean values in the range 4.2–4.3, both positive and “no” saying. Attitude had a non-significant loading to one negative attitudes (mean values 2.4–4.7), moderate to indicator (attend a dentist regularly is intolerable) which strong subjective norms (mean values 3.9–4.4), mod- was removed from the model. Re-estimation of the erate descriptive norms (mean values 3.4–3.9), strong 6-factor model gave acceptable fit (CMIN = (df) 655.666 perceived behavioral control (mean values 4.4–4.5), (69), CFI = .96, RMSEA = .058, AIC = 757.666). As shown and weak action planning (mean values 2.0–2.2). in Table 3, all factor loadings were in the expected direc- tion and had significant regression weights with their re- Evaluation of the measurement model lated latent variables (C.R. > 1.96), indicating convergent The default ML estimation with AMOS requires validity. Most statistically significant items’ standardized continuous multivariate normality of the observed indi- regression weights were above 0.3, and thus in accordance cator variables. As multivariate kurtosis represented by with the threshold proposed [28]. Higher values of Mardia’s coefficient was below the recommended value the indicators were associated with stronger (positive) of 3.0, it was not deemed necessary to bias correct attitudes, stronger subjective norms, descriptive norms, estimates through bootstrapping [24]. According to the perceived behavioral control, intentions and action fit indices (CFI, RMSEA, AIC) employed, CFA indicated planning. The inter-factor correlations (correlations be- that an initially proposed correlated 6-factor model tween the 6 latent variables) were below 0.85 indicating (attitudes, subjective norms, descriptive norms, perceived acceptable discriminant validity (< 0.85). Figure 1 depicts control, intention, action planning) was not an acceptable the modified 6- factor measurement model based on CFA. fit on any of the fit indices employed (CMIN (df) = 1457.6 Gender specific modified correlated factor models (89), CFI = .925, RMSEA =0.058, AIC = 757.666). Inspec- indicated acceptable fit for males (CMIN 306, df 69, tion of modification indices indicated covariation between p < 0.000, CFI = 0.968, RMSEA = 0.056) as well as for Åstrøm et al. BMC Oral Health (2018) 18:95 Page 5 of 10 Table 2 Descriptive statistics of all variables related to the augmented model of planned behavior Mean SD Min Max Theoretical range Intention I intend to attend dentist regularly (Q31_1) 4.3 1.0 1 5 Low-high I have made a decision to attend (Q31_2) 4.2 1.1 1 5 Low-high Attitudes To attend dentist regularly is: —reasonable (Q31_4) 4.7 0.7 1 5 Negative-positive –necessary (Q31_6) 4.3 1.0 1 5 Negative-positive -economic burden (Q31_5) 2.4 1.3 1 5 Negative-positive Intolerable Q31_3 4.2 1.1 1 5 Negative-positive Subjective norm My parents want me to attend regularly (Q31_7) 4.0 1.1 1 5 Low-high My partner want me to attend regularly (Q31_8) 3.9 1.1 1 5 Low-high My dentist want me to attend regularly (Q31_9) 4.4 0.9 1 5 Low-high Descriptive norm My friends attend regularly (Q31_12) 3.4 1.0 1 5 Low-high My parents attend regularly (Q31_13) 3.9 1.1 1 5 Low-high Perceived control It’s up to me to attend regularly (Q31_10) 4.5 0.8 1 5 Low-high I am capable to attend regularly (Q31_11) 4.4 1.0 1 5 Low-high Action planning I have made a detailed plan regarding—————— When attending (Q31_14) 2.0 1.2 1 5 Low-high Where attending (Q31_15) 2.2 1.4 1 5 Low -high How attending (Q31_16) 2.0 1.3 1 5 Low-high Table 3 Standardized regression weights for the different components of the modified correlated 6-factor measure model including intention(INT), attitudes (ATT), subjective norms (SN), descriptive norms (DN), perceived behavioral control (PBC), action planning (AP) Parameters Observed variable (figure label) Parameter estimate (factor loading) INT I intend to attend (Q31_1) 0.927 *** I have decided to attend (Q3_2) 0.890*** ATT To attend is reasonable (Q31_4) 0.695*** To attend is necessary (Q31_6) 0.715*** To attend is an economic burden (Q31_5) 0.117*** SN My parents want me to attend (Q31_7) 0.789*** My friends want me to attend (Q31_8) 0.627*** My dentist want me to attend (Q31_9) 0.792*** DN My friends attend (Q31_12) 0.680*** My parents attend (Q31_13) 0.642*** PBC Its up to me whether to attend (Q31_10) 0.358*** I am capable to attend (Q31_11) 0.916*** AP I have made a detailed plan when (Q31_14) 0.901*** I have made a detailed plan where (Q31_14) 0.895*** I have made a detailed plan how (Q31_15) 0.832*** ***p < 0.001 Åstrøm et al. BMC Oral Health (2018) 18:95 Page 6 of 10 Fig. 1 Modified 6-factor measurement model based on CFA females (CMIN 402,846, df = 69, p < 0.000, CFI = 0.964, Structural equation model RMSEA = 0.058). Multi-group analyses, testing for in- Structural equation modelling, SEM, was conducted to variance across males and females simultaneously, re- estimate the fit of the augmented TPB model and the vealed acceptable fit for the unconstrained model relationships among the latent constructs. The model (CMIN = 709.689, df 138, p < 0.000, CFI = 0.966, with intention (INT), action planning (AP), and dental RMSEA = 0.040) indicating configural invariance attendance predicted by attitudes (ATT), subjective (equivalent factor structure). Compared to the uncon- norms (SN), descriptive norms (DN) and perceived be- strained baseline model, a model with constrained havioral control (PBC) was an acceptable fit to the data; measurement weights were statistically significant (CMIN CMIN 821.234 (85), p < 0.001, CFI = 0.959, RMSEA = 738.00 df147, p < 0.001, CFI = 0.964, RMSEA = 0.040). As 0.058 and AIC = 923,234). Direct paths from attitudes, indicated by the slightly increase in CMIN and decline in subjective norms and descriptive norms on dental CFI values as compared to those in the unconstrained attendance behavior did not improve the fit of the configural model, some variance in factor loadings could model and none of those paths was statistically be expected across males and females. The difference significant. Figure 2 depicts the direct effects for the Δ CMIN =28.319, DF 9 was statistically significant augmented TPB model. at p < 0.001 indicating lack of metric invariance or at As hypothesized by the augmented TPB, stronger best partial invariance for the factor loadings. attitudes β = .70, p < 0.001, perceived behavioral control Åstrøm et al. BMC Oral Health (2018) 18:95 Page 7 of 10 Fig. 2 The augmented Theory of Planned Behavior structural equation model. For ease of interpretation only direct pathways are shown β = 0.18, p < 0.001 and descriptive norm β = .11, p < 0.001 estimated by multiplying the direct effects of the variables were all linked to stronger intentions (Table 4). involved in the total pathway. An indirect path from Subjective norm was negatively related to intention perceived behavioral control to behavior (β = 0.01) was as β = −-17, p < 0.001. Stronger intention was linked to follows; Perceived behavioral control-intention (β =.18), stronger action planning β =.27, p <0.001 and to intention-action planning (β = .27), action planning- more frequent dental attendance β =0.46, p < 0.001. behavior (β = .19). This indicates that the effect of Stronger PBC was also linked to more frequent den- perceived behavioral control on dental attendance was tal attendance, however this path was not statistically primarily through intention and action planning. An indir- significantly β =0.06, 0.03n.s. Stronger action plan- ect path from intention to behavior (β =0.05) was as ning was linked to more frequent dental attendance follows; intention-action planning (β =0.27), action β = .19, p<0.001.Attitudes,subjectivenorms,de- planning-behavior (β =0.19). The effect of intention on scriptive norms and perceived behavioral control behavior was primarily a direct one. accounted for 64% of the variance in intention, intention accounted for 7.6% of the variance in ac- Discussion tion planning and intention, action planning and per- The present study examined, for the first time, the effect ceived behavioral control accounted for 32% variance of motivational (intention) and volitional (action planning) in dental attendance. Specific indirect effects were factors upon regular dental attendance using a Table 4 Significant direct standardized regression weights for the extended theory of planned behavior- Modified SEM model Standardized regression weight % total effect Direct standardized effects Intention-attitudes .76 (.70)*** Intention: 64 Intention-subjective norms .-19 (.-17)*** Action plan: 7.6 Intention-descriptive norm .11 (.11)*** Behavior: 32 Intention-perceived control .16 (.18)*** Intention-Action plan .27 (.27)*** Action plan-behavior .21 (.19)*** Intention -Behavior .51(.46)*** Perceived control-behavior .06 (.03)ns Indirect standardized effects Perceived control-intention-action plan - behavior .01 Intention-action plan -behavior .05 ***p < 0.001 Åstrøm et al. BMC Oral Health (2018) 18:95 Page 8 of 10 cross-sectional design, a structural equation modelling ap- intention. A direct path from descriptive norm to proach (SEM) and a representative sample of Norwegian intention suggests that young adults are guided by what adults 25–35 years of age. The benefit of SEM over other others do regarding their dental visiting behavior. Meta statistical procedures is its ability to test the hypothesized analytical reviews of health related behaviors have also re- relationships among observed and latent variables in the vealed that descriptive norm adds to the prediction of TPB model completely and simultaneously. Structural intention independent of the TPB constructs [30, 31]. equation modelling has gained considerable popularity Nevertheless, attitudes were the strongest motivational de- and whilst modelling the TPB constructs as latent terminant implying that younger adults’ decision to attend variables shows the ability to account for measure- a dentist on a regular basis was almost entirely based on ment errors, which may influence the relationships in anticipated benefits of that behavior but also on social the model [25, 26]. norms (subjective norm and descriptive norm) and con- This study revealed that the proportion of dental at- siderations of potential obstacles (perceived behavioral tendance at least once a year amounted to 47.2% among control), in that order. The finding that attitudes and per- 25–35 year old Norwegian adults. This prevalence rate is ceived behavioral control are predictive of intended dental marginally lower and higher than those reported among attendance is in line with other studies predicting deci- 25-year-old Norwegians in 1997 (62%) and 2007 (44.6%), sions to utilize health care services [1, 16]. These findings respectively, and deviates with figures from 2013 indicat- imply that reduced perceived control due to barriers, such ing that 63% of 20–39 year olds had visited a dentist as for instance dental anxiety and fear, would reduce within the previous year [6, 9]. Nevertheless, dental at- intention and actual use of dental health care services tendance rate is not satisfactory as long as 21% reported among younger Norwegian adults. Unexpectedly, the dir- dental attendance frequency less than every second year. ection of the path from subjective norms to intention was In a first step, a modified correlated factor analytical negative implying that higher perceived social approval for model provided support for the factorial validity of a dental attendance result in lower motivation for that questionnaire supposed to measure intention, action behavior. Although speculative, the construct of psycho- planning, attitudes, subjective norms, perceived control logical reactance may offer an explanation to this uncom- and descriptive norms thus confirming construct validity mon finding as psychological reactance effects in health of a modified 6-factor model (Fig. 1). Although a small related behavior have been observed previously in various and statistically significant p-value for the chi-square sta- domains [20]. In practical terms, however, interventions tistics indicated poor fit of the measurement model, by targeting young adults’ dental attendance behavior might taking sample size into consideration, the comparative usefully focus on informed awareness of the positive oral fit indices fulfilled the criteria for good fit [24, 26]. In health consequences following this behavior accompanied the final model, all inter-factor correlations were below with strategies such as modeling and group performance. the threshold set to indicate poor discriminative validity Educational messages aimed at increasing young adults’ [25–27]. Structural equation modelling in a second step regular dental attendance could highlight the prevalence showed that the augmented TPB model applied was a of dental attendance among the youth in the community. good fit to the data explaining large amounts of vari- If young adults get a sense that everybody at their own ation in intention and attendance behavior. In addition, age is attending on a regular basis, they might be encour- multi-group analysis revealed that the structural part age to abandon their non-attendance. (configural invariance) of the model operated equiva- Intention was by far the strongest predictor of dental lently across males and females, although the factor attendance (β = .46), whereas action planning came sec- loadings did not achieve invariance. ond (β = .21). This supports the hypothesis that action The present finings add to previous findings consider- planning contributes to the prediction of dental attend- ing the ability of the augmented TPB to account for ance over and above the effect of intention whereas the greater variance in intention and behavior than the TPB association between intention and behavior was partly alone [28, 29]. The explained variance in intention (64%) mediated through action planning. Forming action plans and behavior (32%) was higher than that commonly re- considering when, where and how to act facilitates be- ported in meta-analyses of the TPB, being in the range havioral action by setting situational cues that activate of 40–49% for intention and 26–36% regarding actual cognitive processes needed to execute the behavior and behavior [14, 15]. Thus, the results revealed direct highlights that intenders may benefit from formulating statistically significant pathways to intention from at- plans to engage in regular dental attendance [28, 29]. titude (β = .70), subjective norm (β =.-.17), PBC (β =.18), Previous studies have revealed that action planning con- and descriptive norm (β = .11). These findings support the tributes to the prediction of health service utilization, hypothesis that TPB augmented with descriptive norm such as cervical cancer screening [18, 20]. Schutz et al. would predict behavior indirectly through behavioral [28] and Åstrøm [19] examined subsequent flossing in Åstrøm et al. BMC Oral Health (2018) 18:95 Page 9 of 10 the context of social cognition theory and found that ac- Conclusions tion planning was a significant predictor of actual flossing The presents study is the first large nationally represen- alongside intentions and previous flossing. Consistent with tative population based study analyzing younger adults’ those studies, but at odds with others [18, 19, 28], the dental attendance behavior within the context of an aug- present one found action planning to be a significant mented TPB model and using a structural equation predictor of dental attendance. Inconsistent with TPB, modeling approach. The present findings support the perceived behavioral control did not emerge as a signifi- utility of the TPB, the expanded normative component cant predictor of dental attendance. This accords, and the construct of action planning in predicting youn- however, with a meta-analysis by Cooke and French [31], ger adults’ intended and self-reported dental attendance. where perceived control was an unimportant predictor of Interventions targeting young adults’ dental attendance screening behavior. Thus, attending a dentist on regular behavior might usefully focus on positive consequences basis seem to be under complete volitional control by following this behavior accompanied with modeling and younger adults in Norway, who do not require particular group performance. resources, opportunities and technical skills for perform- ance [13]. Additional file This study should be interpreted within the context of Additional file 1: Questionnaire in English language version translated its strengths and limitations. The evidence provided from Norwegian. (DOCX 18 kb) from a large population based study that dental attend- ance is strongly associated with action planning and Abbreviations intention, which in turn is associated with attitudes, AIC: Aikaikes; AP: Action planning; ATT: Attitudes; CFA: Confirmatory factor subjective norms, descriptive norms and perceived analysis; CFI: Comparative fit index; CMIN: minimum discrepancy; CR: Critical ratio; DN: Descriptive norms; INT: Intention; MI: Modification indices; behavioral control identifies targets for informing dental ML: Maximum likelihood; PBC: Perceived behavioral control; RMSEA: root health care interventions among young adults in mean square error of approximation; SEM: Structural Equation Modeling; Norway. A limitation of the present study was the use of SN: subjective norms; SPSS: Statistical Packages for Social Sciences; TPB: Theory of planned behavior self-reported dental attendance that might be biased by social desirability bias resulting in over reporting as Acknowledgements compared with medical records. Evidence suggests that The authors are grateful to the Public Dental Health care Services, Hordaland county for financial support to conduct this study. the validity of self-reported use of dental services ranges from poor to excellent, depending on service type [32]. Funding Moreover, the dental attendance question was adopted This study did receive funding form the Public Dental Health Care Services in Hordaland County, Norway. The funding body did not have any role in the from previously tested measures and it is reasonable to design of the study, nor in the data collection analysis and interpretation of assume that it was sufficiently simple and unambiguous the data when writing the manuscript. to achieve a satisfactory degree of reliability. Another Availability of data and materials weakness associated with the present cross-sectional The dataset generated and analyzed during the current study is not publicly study, as with most population based electronically ad- available as more articles will be based on the data set. Data may be ministered surveys, is the relatively low response rate. available from the corresponding author on reasonable request. Comparison of sex, age and educational level distribu- Authors’ contributions tions among participants with the corresponding figures ANÅ planned the study, designed the questionnaire and co0nducted all in the population provided by official statistics showed a statistical analyses. She was the main contributor in completing the writing of the manuscript. SAL: Contributed to the conduction of the statistical similar age- and educational level distribution but a analyses and provided valuable statistical guidance. FG: Contributed to the moderately different gender distribution that most prob- study design and completion of the manuscript. Provided valuable technical ably did not affect the generalizability of the findings guidance. All authors read and approved the final manuscript presented. A further weakness is the use of a cross Ethics approval and consent to participate –sectional design, thus violating Ajzen’s recommended The Ombudsman, Norwegian Centre for Research Data, NSD, approved and longitudinal design for the original TPB model [13]. registered the survey (NO.49241). NORSTAT (www.norstat.no) was responsible for the distribution of the questionnaires and collection of data. Participants Measuring intention to attend dentists and self-reported were informed about the purpose of the study and that participation was dental attendance in one point in time might have voluntary. Return of a completed questionnaire was recognized as the resulted in an unrealistic high explained variance of informed consent. behavior since the intention behavior gap widens the Competing interests longer the time interval between intention and behavior The authors declare that they have no competing interests. [13]. Further studies should incorporate subsequent measures of behavior or use information from dental Publisher’sNote records to validate the self-report measure utilized in Springer Nature remains neutral with regard to jurisdictional claims in the present study. published maps and institutional affiliations. Åstrøm et al. BMC Oral Health (2018) 18:95 Page 10 of 10 Received: 27 December 2017 Accepted: 22 May 2018 25. Byrne B. Structural equation modelling with AMOS. Basic concepts, applications and programming. London: Lawrence Erlbaum Associates, Publishers; 2001. 26. Brown T. Confirmatory factor analysis for applied research. London: Guilford; References 27. Hu L, Bentler PM. Cut off criteria for fit indices in co-variance structures 1. Holst D. Varieties of oral health care systems. Public Dental services: analysis: conventional criteria versus new alternatives. Str Eq Mod. Organization and Financing of Oral health Care services in the Nordic 1999;6:1–55. countries. In: Pine C, Harris R, editors. Community Oral health. 2nd ed. 28. Schutz B, Sniehotta FF, Wiedemann A, Seemann R. Adherence to a daily London: Quintesse Publishing CO. Ltd; 2007. flossing regimen in university students:effects of planning when, where , 2. Nihtila A. A Nordic project of quality indicators for oral health care. Finland: how and what to do in face of barriers. J Clin Periodontol. 2006;33:612–9. Helsinki; 2010. 29. Gollwitzer PM, Sheeran P. Implementation intention and goal achievement: 3. Grytten J. Payment systems and incentives in dentistry. Community Dent a meta- analysis of effects and processes. Adv Exp Soc Psychol. Oral Epidemiol. 2017;45:1–11. 2006;38:69–119. 4. Vikum E, Krokstad S, Holst D, Westin S. Socio-economic inequalities in 30. Rivis A, Sheeran P. Descriptive norm as an additional predictor in the theory dental service utilization in a Norwegian county: the third Nord-Trondelag of planned behavior: a meta-analysis. Curr Psychol. 2003;22:218–33. health survey. Scand J Public Health. 2012;40:648–55. 31. Cooke R, French DP. How well do the theory of reasoned action and theory 5. Grytten J, Holst D, Skau I. Demand for and utilization of dental services of planned behavior predict intentions and attendance at screening according to household income in the adults population in Norway. programs? A meta- analysis. Psychol Health. 2008;23:745–65. Community Dent Oral Epidemiol. 2012;40:297–305. 32. Gilbert GH, Rose JS, Shelton BJ. A prospective study of the validity of self – 6. Grytten J, Holst D, Skau I. Demand for dental services and dental treatment reported use of specific types of dental services. Public health Rep. patterns among Norwegian adults. Nor Tannlegeforen Tid. 2014;124:276–83. 2003; 118:18–26. 7. Suominen AL, Helminen S, Lahti S, Vehkalahati MM, Knuuttila M, Varsio S, Lindblad A. Use of oral health care services in Finnish adults-results from cross-sectional health 2000 and 2011 surveys. BMC Oral Health. 2017;17:78. 8. Støle AC, Holst D, Schuller AA. Decreasing numbers of young adults seeking dental care on yearly basis. A reason for concern? Nor Tannlegeforen Tid. 1999;109:392–5. (in Norwegian) 9. Åstrøm AN, Skaret E, Haugejorden O. Dental anxiety and dental attendance among 25 –year-olds in Norway: time trends from 1997-2007. BMC Oral Health. 2011;11:10. 10. Official Statistics of Norway, 2016 https://www.ssb.no/statistikkbanken. 11. Berglund E, Westerling R, Lytsy P. Social and health related factors associated with refraining from seeking dental care<: a cross-sectional population study. Community Dent Oral Epidemiol. 2017;45:258–65. 12. Åstrøm AN, Ekback G, Ordell S, Nasir E. Long-term routine dental attendance : influrnce on tooth loss and oral health related quality of life in Swedish older adults. Community Dent Oral Epidemiol. 2014:460–9. 13. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50:179–211. 14. Armitage CJ, Conner M. Efficacy of the theory of planned behavior. A meta analytic review. Br J Soc Psychol. 2001;40:471–99. 15. McEachan BA, Taylor N, Harrison R, Lawton R, Gardner P, Conner M. Meta- analysis of the reasoned action approach to understanding health beaviors. Ann Behav Med. 2016:50–592. 16. Luzzi L, Spencer AJ. Factors influencing the use of public dental services: an application of the theory of planned behavior. BMC Health Serv Res. 2008;8: 17. Sniehotta F, Scholz U, Schwarzer R. Action plans and coping plans for physical exercise: a longitudinal intervention study in cardiac rehabilitation. Bristish J Health Psychol. 2006;11:23–37. 18. Sheeran P, Orbell S. Using implementation intention to increase attendance for cervical cancer. Health Psychol. 2000;19:283–9. 19. Astrom AN. Applicability of action planning and coping planning to dental flossing among Norwegian adults: a confirmatory factor analysis approach. Eur J Oral Sci. 2008;116:250–9. 20. Sieverding M, Ciccarello L, Matterne U. What role do social norms play in the context of men’s cancer screening intention and behavior? Application of an extended theory of planned behavior. Health Psychol. 2010;29:72–81. 21. Lee H. The role of descriptive norm within the theory of planned behavior in predicting Korean American’s exercise behavior. Psychol Rep. 2011;109:208–18. 22. Cialdini RB, Reno RR, Kallgren CA. A focus theory of normative conduct: recycling the concept of norms to reduce littering in public places. J Pers and Soc Osychol. 1990;58:1015–26. 23. Rivis A, Sheeran P. Descriptive norms as an additional predictor in the theory of planned behavior: a meta-analysis. Curr Psychol: Dev, Learn Pers, Soc. 2003;22:218–33. 24. Schumacker RE, Lomax RG. A Beginner’s guide to structural equation modelling. second ed. London: Lawrence Erlbaum Associates Publishers; http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Oral Health Springer Journals

Applying the theory of planned behavior to self-report dental attendance in Norwegian adults through structural equation modelling approach

Free
10 pages

Loading next page...
 
/lp/springer_journal/applying-the-theory-of-planned-behavior-to-self-report-dental-9QQ6oxrdf9
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Dentistry; Dentistry; Oral and Maxillofacial Surgery
eISSN
1472-6831
D.O.I.
10.1186/s12903-018-0558-7
Publisher site
See Article on Publisher Site

Abstract

Background: Understanding factors that affect dental attendance behavior helps in constructing effective oral health campaigns. A socio-cognitive model that adequately explains variance in regular dental attendance has yet to be validated among younger adults in Norway. Focusing a representative sample of younger Norwegian adults, this cross-sectional study provided an empirical test of the Theory of Planned Behavior (TPB) augmented with descriptive norm and action planning and estimated direct and indirect effects of attitudes, subjective norms, descriptive norms, perceived behavioral control and action planning on intended and self-reported regular dental attendance. Method: Self-administered questionnaires provided by 2551, 25–35 year olds, randomly selected from the Norwegian national population registry were used to assess socio-demographic factors, dental attendance as well as the constructs of the augmented TPB model (attitudes, subjective norms, descriptive norms, intention, action planning). A two-stage process of structural equation modelling (SEM) was used to test the augmented TPB model. Results: Confirmatory factor analysis, CFA, confirmed the proposed correlated 6-factor measurement model after re-specification. SEM revealed that attitudes, perceived behavioral control, subjective norms and descriptive norms explained intention. The corresponding standardized regression coefficients were respectively (β = 0.70), (β =0.18), (β = − 0.17) and (β =0.11) (p < 0.001). Intention (β =0.46) predicted action planning and action planning (β =0.19) predicted dental attendance behavior (p < 0.001). The model revealed indirect effects of intention and perceived behavioral control on behavior through action planning and through intention and action planning, respectively. The final model explained 64 and 41% of the total variance in intention and dental attendance behavior. Conclusion: The findings support the utility of the TPB, the expanded normative component and action planning in predicting younger adults’ intended- and self-reported dental attendance. Interventions targeting young adults’ dental attendance might usefully focus on positive consequences following this behavior accompanied with modeling and group performance. Keywords: Dental attendance, Young adults, Theory of planned behavior, Structural equation modelling, AMOS * Correspondence: Anne.Aastrom@uib.no Oral health Centre of Expertise in Western Norway, Bergen, Hordaland, Norway Department of Clinical Dentistry, Faculty of Medicine, University of Bergen, PO Box 7804, N-5020 Bergen, Norway © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Åstrøm et al. BMC Oral Health (2018) 18:95 Page 2 of 10 Background between attitudes and behavior mediated by behavioral The importance of dental attendance is a key measure in intentions. Intention directly influences behavior and is health education as regular attendance is associated with shaped by attitudes, subjective norms and perceived good health and well-being. In Norway, the Public behavioral control regarding the behavior. Empirical Dental Service (PDS) is financed by taxes and provides validations of the TPB have revealed that the model free dental care to children and adolescents until 20 years reliably explains 40–50% of the variance in intention and of age [1, 2]. The private dental services provides dental that intention explains between 20 and 40% of the care to the general adult population and is organized variance in actual behavior [14–16]. according to market mechanisms, with dental fees In spite of its predictive success, TPB has been determined by supply and demand and with very limited criticized for its validity and it has been shown that private or public insurance arrangements [3]. Regardless other variables explain considerable proportions of the of the disparities in dental coverage between Norway variance in intention and behavior [17]. Moreover, de- and other Scandinavian countries, dental attendance scriptive norms and action planning have shown residual rates have been high among Norwegian adults. About 80 effects on intention and behaviour after consideration of and 77% of Norwegian adults above 20 years of age re- the original TPB variables [17, 18]. Evidence suggests ported having visited a dentist during the last 12 months that adding action planning would improve the predic- in 2008 and 2013, respectively [4–6]. tion of behavior from the TPB [18, 19]. Thus, formation Nevertheless, as in other Scandinavian countries, the of action plans can be used to promote the realization of prevalence of regular dental care utilization among Nor- desired outcomes [18, 19]. The role of subjective norms wegian adults varies according to age, period and within the TPB has also been considered [20, 21]. socio-economic status, being smallest in the younger Subjective norms have been criticized for its narrow age- and the lower income groups [4–7]. Støle et al. [8] conceptualization, focusing what significant people found that 87% of 23–24 year old Norwegian adults thinks others ought to do, neglecting descriptive visited a dentist every second year in 1983, whereas the norms referring to what significant others themselves corresponding figures in 1994 was 85%. Among 25 year actually do [20–22]. Descriptive norms correlate with old Norwegian adults, 62 and 44% reported dental behavioral intentions and have shown to be among attendance once a year in 1997 and 2007, respectively the strongest correlates of physical activity [23]. The [9]. According to the Official Statistics of Norway, the TPB has received considerable empirical support prevalence of having visited public dental health care across health- and social behaviors, including oral services during last year was highest among 45–66 year hygiene behaviors and health screening [15]. However, olds and lowest among 21–24 year olds in 2016 [10]. to our knowledge, only one previous study has exam- Previous studies have identified enabling factors, such as ined use of public dental services in the context of cost of treatment and dental anxiety, as important TPB [16]. barriers towards regular use of dental care [11, 12]. In a Focusing a representative sample of young Norwe- recent population-based study of Swedish adults, finan- gian adults 25–35 years old, this study provides an cial problems and lack of social support were associated empirical test of the TPB augmented with descriptive with refraining from seeking dental care [11]. Whereas norm and action planning and estimates direct and socio-demographic- and need related factors are import- indirect effects of attitudes, subjective norms, descrip- ant covariates of dental care utilization, relatively few tive norms, perceived control, and action planning on studies have considered modifiable socio-cognitive intended and self-reported regular dental attendance. determinants in the younger adult populations. Based on the conceptualization of the TPB and previ- Influencing younger adults’ adherence to continued ous empirical support it was hypothesized that the re- dental attendance requires understanding of the sponses to 16 observed indicator variables could be socio-cognitive factors underlying their decision to explained by 6 latent factors in terms of attitudes, comply with advices for regular dental care. A subjective norms, descriptive norms, perceived con- socio-cognitive model that adequately explains vari- trol, intention and action planning. Further, it was ance in regular dental attendance has yet to be hypothesized that each indicator would have a stron- validated among younger adults in Norway. ger relation with their corresponding factor than with the competing factors. Finally, it was hypothesized Theoretical approach that attitudes, subjective norm, descriptive norm and The theory of planned behavior, TPB, is a widely applied perceived behavioral control would predict behavioral socio-cognitive model of the attitude–behavior relation- intention and that intention, action planning and per- ship, assuming that most conscious behaviors is rational ceived behavioral control, would predict self-reported and goal directed [13, 14]. TPB proposes a causal link dental attendance. Åstrøm et al. BMC Oral Health (2018) 18:95 Page 3 of 10 Methods measured by two items, e.g. “My friends attend dentist Study design, participants and ethical issues regularly”– with response categories ranging from 1 Thepresent studyuseddata from anelectronic, (strongly disagree) to 5 (strongly agree). Action planning cross-sectional public dental health survey conducted was assessed using the action planning scale adopted in Norway. A representative sample of 9000 adults from Sniehotta et al. [17], including three items, e.g. “I (using individuals as the primary sampling unit) aged have made a detailed plan when to attend, where to 25–35 years was randomly selected from the Norwegian attend and how to attend a dentist regularly in the national population registry in September 2016. Participa- future”. Response categories ranged from 1 (strongly tion was voluntary and anonymous and the return of a disagree) to 5 (strongly agree). Parts of the question- completed questionnaire recognized as the informed naire used in the present study is available in English consent. Ethical permission to carry out the survey was in the Additional file 1. granted by the Ombudsman, Norwegian Center for Research Data (NO.49241). NORSTAT (www.norstat.no) Statistical analyses was responsible for the electronically distributed question- Data were analyzed using SPSS version 22.0 (IBM Corp. naires and for the data collection. An eligible sample of Released 2013, IBM SPSS Statistics for Windows, 9052 adults aged 25–35 years of age received an electronic Armonk NY: IBM Corp). IBM SPSS AMOS 16.0 [24] version of the questionnaire with an introductory letter was used to test the hypothesized augmented TPB explaining the purpose of the study. Total response rate model using a two-step modelling approach whereby the was 29% (2635/9052). Eighty-four respondents were re- measurement model (step1) and the structural model moved due to incomplete questionnaires. All participants (step 2) were constructed separately [25]. First, a who provided complete questionnaires were included in confirmatory factor analysis, CFA, using maximum the present study (n = 2551). likelihood estimation (ML) was conducted to test the adequacy of the measurement model [25]. Modification Measures indices (MI) were used to identify sources of misfit in Dental attendance behavior was measured using one the model. A prerequisite for testing of invariance across question; “How often do you usually visit a dentist?” the structural paths in the full structural model (step 2) is response categories ranged from (1) twice a year or more that the measurement model has configural and metric to (4) more seldom than every second year. Components invariance. Configural invariance was examined by test- of an augmented version of Ajzen’s TPB [13] was mea- ing the fit of the modified correlated measurement sured in terms of attitudes, subjective norms, perceived model separately for males and females and by testing behavioral control, descriptive norm and action planning the fit of an un-constrained multi-group model. Metric in relation to regular dental attendance. In accordance invariance was examined by comparing a multi-group with recommendations from Ajzen [13], each construct model with all factor loadings constrained equal to the was measured considering the four elements of action baseline configural model in which the factor loadings (attending), target (dentist), context (on a regular basis), were free to vary. The models were assumed and time (future) (13) (. Intention to attend a dentist non-invariant if the change in chi square was significant regularly was measured by two items, e.g. “I intend to at- and the decrease in comparative fit index, CFI, was less tend a dentist regularly in the future.” Responses were than 0.001 [26]. indicated on a four-point scale: (1) Strongly disagree, (2) A full structural equation modelling, SEM, (step 2) Disagree, (3) Neither agree nor disagree (4) Agree and examined whether the hypothesized TPB model was ac- (4) Strongly agree. Attitude towards regular dental at- ceptable fit to the present data, testing simultaneously tendance was assessed by four items, e.g. “to attend a the interrelationships specified within the a priori aug- dentist regularly in the future do not make sense to me”. mented TPB model. To assess how adequately the hy- Responses were indicated on a five-point scale ranging pothesized model described the sample data, chi-square from 1 (strongly disagree) to 5 strongly disagree). Sub- test was used together with the following goodness of fit jective norm was measured by three items, e.g. “My par- indices; CFI (Comparative fit index), RMSEA (root mean ents (partner/friend, dentist) want me to attend a dentist square error of approximation) and AIC (Akaike’s regularly in the future”. Responses were indicated on a information criteria). In line with the conventional five-point scale ranging from 1 (strongly disagree) to 5 recommendations of Hu and Bentler [27], a good model (strongly agree). Perceived behavioral control was mea- fit was indicated by a RMSEA less or equal to 0.06, a sured by two items, e.g. “Its totally up to me whether I CFI greater or equal to 0.90 and with a model having attend a dentist regularly in the future”. Responses were lower AIC being the more plausible together with a indicated on a five-point scale ranging from 1 (strongly non-significant Chi square. Statistical significance of disagree) to 5 (strongly agree). Descriptive norm was parameter estimates are the Critical Ration (CR) Åstrøm et al. BMC Oral Health (2018) 18:95 Page 4 of 10 representing the parameter estimate divided by its stand- Table 1 Frequency distribution of participants’ socio-demographic characteristics and dental attendance behavior, (n = 2551) and ard error. Based on a level of 0.05, the test statistics (CR) corresponding percentage figures in the total population needed to be 1.96 before rejection of the null hypothesis. Participants Total population Results Category % (n)% Sample profile Gender Male 43.3 (1105) 51.3 In spite of the relatively low response rate (29%) obtained, the age distribution of the final sample Female 56.7 (1446) 49.0 corresponded with that of the Norwegian population Age 25-29 years 43.7 (1116) 46.3 20–44 years old by December 2016. The age distribution 30–35 years 56.3 (1435) 53.0 of younger (25–29 years) and older (30–35 years) partic- Country of birth Norway 91.5 (2333) ipants were 43.7 and 56.3%, respectively. Corresponding Other Nordic 2.6 (66) figures in the population were respectively 46.3 and 53%. Outside Nordic 6.0 (152) Whereas the gender distribution in the sample was 43% men and 56.7% women, the corresponding population Civil status Single 36.6 (935) distribution was 51.3 and 49.0%. Among the partici- Married 63.4 (1616) pants, 27.3, 38.6 and 34.1% reported respectively, Highest level of Primary/secondary 27.3 (679) 26.5 primary-, bachelor- and college/university level of educa- education Bachelor degree 38.6 (962) 37.8 tion. Corresponding figures in the adult population College/university 34.1 (850) 32.9 16 years and above were 26.5, 37.8 and 32.9%. Among Income (NOK) At least 400.000 19.5 (413) the respondents (n = 2551), 91.5% were of native Norwegian origin. Eight percent confirmed dental 400,001–800,000 41.2 (876) attendance at least twice a year, 47.2% once a year, > 800,000 39.3 (835) 21.2% every second year and 21.2% more seldom than Dental attendance At least twice a 8.0 (203) every second year (Table 1). year Once a year 47.2 (1205) Descriptive statistics of TPB variables Every second year 21.2 (540) Table 2 depicts mean, standard deviation, minimum More seldom 21.2 (540) and maximum scores and theoretical range for each Norwegian population 20–44 years by December 2016 indicator measuring the latent constructs of attitudes, Norwegian population above 20 years by December 2016 subjective norms, perceived control, descriptive norms, intention and action planning. On average the pairs of error terms, resulting from item overlap, or study group demonstrated strong intentions with reflecting biases in responding such as “yea” saying or mean values in the range 4.2–4.3, both positive and “no” saying. Attitude had a non-significant loading to one negative attitudes (mean values 2.4–4.7), moderate to indicator (attend a dentist regularly is intolerable) which strong subjective norms (mean values 3.9–4.4), mod- was removed from the model. Re-estimation of the erate descriptive norms (mean values 3.4–3.9), strong 6-factor model gave acceptable fit (CMIN = (df) 655.666 perceived behavioral control (mean values 4.4–4.5), (69), CFI = .96, RMSEA = .058, AIC = 757.666). As shown and weak action planning (mean values 2.0–2.2). in Table 3, all factor loadings were in the expected direc- tion and had significant regression weights with their re- Evaluation of the measurement model lated latent variables (C.R. > 1.96), indicating convergent The default ML estimation with AMOS requires validity. Most statistically significant items’ standardized continuous multivariate normality of the observed indi- regression weights were above 0.3, and thus in accordance cator variables. As multivariate kurtosis represented by with the threshold proposed [28]. Higher values of Mardia’s coefficient was below the recommended value the indicators were associated with stronger (positive) of 3.0, it was not deemed necessary to bias correct attitudes, stronger subjective norms, descriptive norms, estimates through bootstrapping [24]. According to the perceived behavioral control, intentions and action fit indices (CFI, RMSEA, AIC) employed, CFA indicated planning. The inter-factor correlations (correlations be- that an initially proposed correlated 6-factor model tween the 6 latent variables) were below 0.85 indicating (attitudes, subjective norms, descriptive norms, perceived acceptable discriminant validity (< 0.85). Figure 1 depicts control, intention, action planning) was not an acceptable the modified 6- factor measurement model based on CFA. fit on any of the fit indices employed (CMIN (df) = 1457.6 Gender specific modified correlated factor models (89), CFI = .925, RMSEA =0.058, AIC = 757.666). Inspec- indicated acceptable fit for males (CMIN 306, df 69, tion of modification indices indicated covariation between p < 0.000, CFI = 0.968, RMSEA = 0.056) as well as for Åstrøm et al. BMC Oral Health (2018) 18:95 Page 5 of 10 Table 2 Descriptive statistics of all variables related to the augmented model of planned behavior Mean SD Min Max Theoretical range Intention I intend to attend dentist regularly (Q31_1) 4.3 1.0 1 5 Low-high I have made a decision to attend (Q31_2) 4.2 1.1 1 5 Low-high Attitudes To attend dentist regularly is: —reasonable (Q31_4) 4.7 0.7 1 5 Negative-positive –necessary (Q31_6) 4.3 1.0 1 5 Negative-positive -economic burden (Q31_5) 2.4 1.3 1 5 Negative-positive Intolerable Q31_3 4.2 1.1 1 5 Negative-positive Subjective norm My parents want me to attend regularly (Q31_7) 4.0 1.1 1 5 Low-high My partner want me to attend regularly (Q31_8) 3.9 1.1 1 5 Low-high My dentist want me to attend regularly (Q31_9) 4.4 0.9 1 5 Low-high Descriptive norm My friends attend regularly (Q31_12) 3.4 1.0 1 5 Low-high My parents attend regularly (Q31_13) 3.9 1.1 1 5 Low-high Perceived control It’s up to me to attend regularly (Q31_10) 4.5 0.8 1 5 Low-high I am capable to attend regularly (Q31_11) 4.4 1.0 1 5 Low-high Action planning I have made a detailed plan regarding—————— When attending (Q31_14) 2.0 1.2 1 5 Low-high Where attending (Q31_15) 2.2 1.4 1 5 Low -high How attending (Q31_16) 2.0 1.3 1 5 Low-high Table 3 Standardized regression weights for the different components of the modified correlated 6-factor measure model including intention(INT), attitudes (ATT), subjective norms (SN), descriptive norms (DN), perceived behavioral control (PBC), action planning (AP) Parameters Observed variable (figure label) Parameter estimate (factor loading) INT I intend to attend (Q31_1) 0.927 *** I have decided to attend (Q3_2) 0.890*** ATT To attend is reasonable (Q31_4) 0.695*** To attend is necessary (Q31_6) 0.715*** To attend is an economic burden (Q31_5) 0.117*** SN My parents want me to attend (Q31_7) 0.789*** My friends want me to attend (Q31_8) 0.627*** My dentist want me to attend (Q31_9) 0.792*** DN My friends attend (Q31_12) 0.680*** My parents attend (Q31_13) 0.642*** PBC Its up to me whether to attend (Q31_10) 0.358*** I am capable to attend (Q31_11) 0.916*** AP I have made a detailed plan when (Q31_14) 0.901*** I have made a detailed plan where (Q31_14) 0.895*** I have made a detailed plan how (Q31_15) 0.832*** ***p < 0.001 Åstrøm et al. BMC Oral Health (2018) 18:95 Page 6 of 10 Fig. 1 Modified 6-factor measurement model based on CFA females (CMIN 402,846, df = 69, p < 0.000, CFI = 0.964, Structural equation model RMSEA = 0.058). Multi-group analyses, testing for in- Structural equation modelling, SEM, was conducted to variance across males and females simultaneously, re- estimate the fit of the augmented TPB model and the vealed acceptable fit for the unconstrained model relationships among the latent constructs. The model (CMIN = 709.689, df 138, p < 0.000, CFI = 0.966, with intention (INT), action planning (AP), and dental RMSEA = 0.040) indicating configural invariance attendance predicted by attitudes (ATT), subjective (equivalent factor structure). Compared to the uncon- norms (SN), descriptive norms (DN) and perceived be- strained baseline model, a model with constrained havioral control (PBC) was an acceptable fit to the data; measurement weights were statistically significant (CMIN CMIN 821.234 (85), p < 0.001, CFI = 0.959, RMSEA = 738.00 df147, p < 0.001, CFI = 0.964, RMSEA = 0.040). As 0.058 and AIC = 923,234). Direct paths from attitudes, indicated by the slightly increase in CMIN and decline in subjective norms and descriptive norms on dental CFI values as compared to those in the unconstrained attendance behavior did not improve the fit of the configural model, some variance in factor loadings could model and none of those paths was statistically be expected across males and females. The difference significant. Figure 2 depicts the direct effects for the Δ CMIN =28.319, DF 9 was statistically significant augmented TPB model. at p < 0.001 indicating lack of metric invariance or at As hypothesized by the augmented TPB, stronger best partial invariance for the factor loadings. attitudes β = .70, p < 0.001, perceived behavioral control Åstrøm et al. BMC Oral Health (2018) 18:95 Page 7 of 10 Fig. 2 The augmented Theory of Planned Behavior structural equation model. For ease of interpretation only direct pathways are shown β = 0.18, p < 0.001 and descriptive norm β = .11, p < 0.001 estimated by multiplying the direct effects of the variables were all linked to stronger intentions (Table 4). involved in the total pathway. An indirect path from Subjective norm was negatively related to intention perceived behavioral control to behavior (β = 0.01) was as β = −-17, p < 0.001. Stronger intention was linked to follows; Perceived behavioral control-intention (β =.18), stronger action planning β =.27, p <0.001 and to intention-action planning (β = .27), action planning- more frequent dental attendance β =0.46, p < 0.001. behavior (β = .19). This indicates that the effect of Stronger PBC was also linked to more frequent den- perceived behavioral control on dental attendance was tal attendance, however this path was not statistically primarily through intention and action planning. An indir- significantly β =0.06, 0.03n.s. Stronger action plan- ect path from intention to behavior (β =0.05) was as ning was linked to more frequent dental attendance follows; intention-action planning (β =0.27), action β = .19, p<0.001.Attitudes,subjectivenorms,de- planning-behavior (β =0.19). The effect of intention on scriptive norms and perceived behavioral control behavior was primarily a direct one. accounted for 64% of the variance in intention, intention accounted for 7.6% of the variance in ac- Discussion tion planning and intention, action planning and per- The present study examined, for the first time, the effect ceived behavioral control accounted for 32% variance of motivational (intention) and volitional (action planning) in dental attendance. Specific indirect effects were factors upon regular dental attendance using a Table 4 Significant direct standardized regression weights for the extended theory of planned behavior- Modified SEM model Standardized regression weight % total effect Direct standardized effects Intention-attitudes .76 (.70)*** Intention: 64 Intention-subjective norms .-19 (.-17)*** Action plan: 7.6 Intention-descriptive norm .11 (.11)*** Behavior: 32 Intention-perceived control .16 (.18)*** Intention-Action plan .27 (.27)*** Action plan-behavior .21 (.19)*** Intention -Behavior .51(.46)*** Perceived control-behavior .06 (.03)ns Indirect standardized effects Perceived control-intention-action plan - behavior .01 Intention-action plan -behavior .05 ***p < 0.001 Åstrøm et al. BMC Oral Health (2018) 18:95 Page 8 of 10 cross-sectional design, a structural equation modelling ap- intention. A direct path from descriptive norm to proach (SEM) and a representative sample of Norwegian intention suggests that young adults are guided by what adults 25–35 years of age. The benefit of SEM over other others do regarding their dental visiting behavior. Meta statistical procedures is its ability to test the hypothesized analytical reviews of health related behaviors have also re- relationships among observed and latent variables in the vealed that descriptive norm adds to the prediction of TPB model completely and simultaneously. Structural intention independent of the TPB constructs [30, 31]. equation modelling has gained considerable popularity Nevertheless, attitudes were the strongest motivational de- and whilst modelling the TPB constructs as latent terminant implying that younger adults’ decision to attend variables shows the ability to account for measure- a dentist on a regular basis was almost entirely based on ment errors, which may influence the relationships in anticipated benefits of that behavior but also on social the model [25, 26]. norms (subjective norm and descriptive norm) and con- This study revealed that the proportion of dental at- siderations of potential obstacles (perceived behavioral tendance at least once a year amounted to 47.2% among control), in that order. The finding that attitudes and per- 25–35 year old Norwegian adults. This prevalence rate is ceived behavioral control are predictive of intended dental marginally lower and higher than those reported among attendance is in line with other studies predicting deci- 25-year-old Norwegians in 1997 (62%) and 2007 (44.6%), sions to utilize health care services [1, 16]. These findings respectively, and deviates with figures from 2013 indicat- imply that reduced perceived control due to barriers, such ing that 63% of 20–39 year olds had visited a dentist as for instance dental anxiety and fear, would reduce within the previous year [6, 9]. Nevertheless, dental at- intention and actual use of dental health care services tendance rate is not satisfactory as long as 21% reported among younger Norwegian adults. Unexpectedly, the dir- dental attendance frequency less than every second year. ection of the path from subjective norms to intention was In a first step, a modified correlated factor analytical negative implying that higher perceived social approval for model provided support for the factorial validity of a dental attendance result in lower motivation for that questionnaire supposed to measure intention, action behavior. Although speculative, the construct of psycho- planning, attitudes, subjective norms, perceived control logical reactance may offer an explanation to this uncom- and descriptive norms thus confirming construct validity mon finding as psychological reactance effects in health of a modified 6-factor model (Fig. 1). Although a small related behavior have been observed previously in various and statistically significant p-value for the chi-square sta- domains [20]. In practical terms, however, interventions tistics indicated poor fit of the measurement model, by targeting young adults’ dental attendance behavior might taking sample size into consideration, the comparative usefully focus on informed awareness of the positive oral fit indices fulfilled the criteria for good fit [24, 26]. In health consequences following this behavior accompanied the final model, all inter-factor correlations were below with strategies such as modeling and group performance. the threshold set to indicate poor discriminative validity Educational messages aimed at increasing young adults’ [25–27]. Structural equation modelling in a second step regular dental attendance could highlight the prevalence showed that the augmented TPB model applied was a of dental attendance among the youth in the community. good fit to the data explaining large amounts of vari- If young adults get a sense that everybody at their own ation in intention and attendance behavior. In addition, age is attending on a regular basis, they might be encour- multi-group analysis revealed that the structural part age to abandon their non-attendance. (configural invariance) of the model operated equiva- Intention was by far the strongest predictor of dental lently across males and females, although the factor attendance (β = .46), whereas action planning came sec- loadings did not achieve invariance. ond (β = .21). This supports the hypothesis that action The present finings add to previous findings consider- planning contributes to the prediction of dental attend- ing the ability of the augmented TPB to account for ance over and above the effect of intention whereas the greater variance in intention and behavior than the TPB association between intention and behavior was partly alone [28, 29]. The explained variance in intention (64%) mediated through action planning. Forming action plans and behavior (32%) was higher than that commonly re- considering when, where and how to act facilitates be- ported in meta-analyses of the TPB, being in the range havioral action by setting situational cues that activate of 40–49% for intention and 26–36% regarding actual cognitive processes needed to execute the behavior and behavior [14, 15]. Thus, the results revealed direct highlights that intenders may benefit from formulating statistically significant pathways to intention from at- plans to engage in regular dental attendance [28, 29]. titude (β = .70), subjective norm (β =.-.17), PBC (β =.18), Previous studies have revealed that action planning con- and descriptive norm (β = .11). These findings support the tributes to the prediction of health service utilization, hypothesis that TPB augmented with descriptive norm such as cervical cancer screening [18, 20]. Schutz et al. would predict behavior indirectly through behavioral [28] and Åstrøm [19] examined subsequent flossing in Åstrøm et al. BMC Oral Health (2018) 18:95 Page 9 of 10 the context of social cognition theory and found that ac- Conclusions tion planning was a significant predictor of actual flossing The presents study is the first large nationally represen- alongside intentions and previous flossing. Consistent with tative population based study analyzing younger adults’ those studies, but at odds with others [18, 19, 28], the dental attendance behavior within the context of an aug- present one found action planning to be a significant mented TPB model and using a structural equation predictor of dental attendance. Inconsistent with TPB, modeling approach. The present findings support the perceived behavioral control did not emerge as a signifi- utility of the TPB, the expanded normative component cant predictor of dental attendance. This accords, and the construct of action planning in predicting youn- however, with a meta-analysis by Cooke and French [31], ger adults’ intended and self-reported dental attendance. where perceived control was an unimportant predictor of Interventions targeting young adults’ dental attendance screening behavior. Thus, attending a dentist on regular behavior might usefully focus on positive consequences basis seem to be under complete volitional control by following this behavior accompanied with modeling and younger adults in Norway, who do not require particular group performance. resources, opportunities and technical skills for perform- ance [13]. Additional file This study should be interpreted within the context of Additional file 1: Questionnaire in English language version translated its strengths and limitations. The evidence provided from Norwegian. (DOCX 18 kb) from a large population based study that dental attend- ance is strongly associated with action planning and Abbreviations intention, which in turn is associated with attitudes, AIC: Aikaikes; AP: Action planning; ATT: Attitudes; CFA: Confirmatory factor subjective norms, descriptive norms and perceived analysis; CFI: Comparative fit index; CMIN: minimum discrepancy; CR: Critical ratio; DN: Descriptive norms; INT: Intention; MI: Modification indices; behavioral control identifies targets for informing dental ML: Maximum likelihood; PBC: Perceived behavioral control; RMSEA: root health care interventions among young adults in mean square error of approximation; SEM: Structural Equation Modeling; Norway. A limitation of the present study was the use of SN: subjective norms; SPSS: Statistical Packages for Social Sciences; TPB: Theory of planned behavior self-reported dental attendance that might be biased by social desirability bias resulting in over reporting as Acknowledgements compared with medical records. Evidence suggests that The authors are grateful to the Public Dental Health care Services, Hordaland county for financial support to conduct this study. the validity of self-reported use of dental services ranges from poor to excellent, depending on service type [32]. Funding Moreover, the dental attendance question was adopted This study did receive funding form the Public Dental Health Care Services in Hordaland County, Norway. The funding body did not have any role in the from previously tested measures and it is reasonable to design of the study, nor in the data collection analysis and interpretation of assume that it was sufficiently simple and unambiguous the data when writing the manuscript. to achieve a satisfactory degree of reliability. Another Availability of data and materials weakness associated with the present cross-sectional The dataset generated and analyzed during the current study is not publicly study, as with most population based electronically ad- available as more articles will be based on the data set. Data may be ministered surveys, is the relatively low response rate. available from the corresponding author on reasonable request. Comparison of sex, age and educational level distribu- Authors’ contributions tions among participants with the corresponding figures ANÅ planned the study, designed the questionnaire and co0nducted all in the population provided by official statistics showed a statistical analyses. She was the main contributor in completing the writing of the manuscript. SAL: Contributed to the conduction of the statistical similar age- and educational level distribution but a analyses and provided valuable statistical guidance. FG: Contributed to the moderately different gender distribution that most prob- study design and completion of the manuscript. Provided valuable technical ably did not affect the generalizability of the findings guidance. All authors read and approved the final manuscript presented. A further weakness is the use of a cross Ethics approval and consent to participate –sectional design, thus violating Ajzen’s recommended The Ombudsman, Norwegian Centre for Research Data, NSD, approved and longitudinal design for the original TPB model [13]. registered the survey (NO.49241). NORSTAT (www.norstat.no) was responsible for the distribution of the questionnaires and collection of data. Participants Measuring intention to attend dentists and self-reported were informed about the purpose of the study and that participation was dental attendance in one point in time might have voluntary. Return of a completed questionnaire was recognized as the resulted in an unrealistic high explained variance of informed consent. behavior since the intention behavior gap widens the Competing interests longer the time interval between intention and behavior The authors declare that they have no competing interests. [13]. Further studies should incorporate subsequent measures of behavior or use information from dental Publisher’sNote records to validate the self-report measure utilized in Springer Nature remains neutral with regard to jurisdictional claims in the present study. published maps and institutional affiliations. Åstrøm et al. BMC Oral Health (2018) 18:95 Page 10 of 10 Received: 27 December 2017 Accepted: 22 May 2018 25. Byrne B. Structural equation modelling with AMOS. Basic concepts, applications and programming. London: Lawrence Erlbaum Associates, Publishers; 2001. 26. Brown T. Confirmatory factor analysis for applied research. London: Guilford; References 27. Hu L, Bentler PM. Cut off criteria for fit indices in co-variance structures 1. Holst D. Varieties of oral health care systems. Public Dental services: analysis: conventional criteria versus new alternatives. Str Eq Mod. Organization and Financing of Oral health Care services in the Nordic 1999;6:1–55. countries. In: Pine C, Harris R, editors. Community Oral health. 2nd ed. 28. Schutz B, Sniehotta FF, Wiedemann A, Seemann R. Adherence to a daily London: Quintesse Publishing CO. Ltd; 2007. flossing regimen in university students:effects of planning when, where , 2. Nihtila A. A Nordic project of quality indicators for oral health care. Finland: how and what to do in face of barriers. J Clin Periodontol. 2006;33:612–9. Helsinki; 2010. 29. Gollwitzer PM, Sheeran P. Implementation intention and goal achievement: 3. Grytten J. Payment systems and incentives in dentistry. Community Dent a meta- analysis of effects and processes. Adv Exp Soc Psychol. Oral Epidemiol. 2017;45:1–11. 2006;38:69–119. 4. Vikum E, Krokstad S, Holst D, Westin S. Socio-economic inequalities in 30. Rivis A, Sheeran P. Descriptive norm as an additional predictor in the theory dental service utilization in a Norwegian county: the third Nord-Trondelag of planned behavior: a meta-analysis. Curr Psychol. 2003;22:218–33. health survey. Scand J Public Health. 2012;40:648–55. 31. Cooke R, French DP. How well do the theory of reasoned action and theory 5. Grytten J, Holst D, Skau I. Demand for and utilization of dental services of planned behavior predict intentions and attendance at screening according to household income in the adults population in Norway. programs? A meta- analysis. Psychol Health. 2008;23:745–65. Community Dent Oral Epidemiol. 2012;40:297–305. 32. Gilbert GH, Rose JS, Shelton BJ. A prospective study of the validity of self – 6. Grytten J, Holst D, Skau I. Demand for dental services and dental treatment reported use of specific types of dental services. Public health Rep. patterns among Norwegian adults. Nor Tannlegeforen Tid. 2014;124:276–83. 2003; 118:18–26. 7. Suominen AL, Helminen S, Lahti S, Vehkalahati MM, Knuuttila M, Varsio S, Lindblad A. Use of oral health care services in Finnish adults-results from cross-sectional health 2000 and 2011 surveys. BMC Oral Health. 2017;17:78. 8. Støle AC, Holst D, Schuller AA. Decreasing numbers of young adults seeking dental care on yearly basis. A reason for concern? Nor Tannlegeforen Tid. 1999;109:392–5. (in Norwegian) 9. Åstrøm AN, Skaret E, Haugejorden O. Dental anxiety and dental attendance among 25 –year-olds in Norway: time trends from 1997-2007. BMC Oral Health. 2011;11:10. 10. Official Statistics of Norway, 2016 https://www.ssb.no/statistikkbanken. 11. Berglund E, Westerling R, Lytsy P. Social and health related factors associated with refraining from seeking dental care<: a cross-sectional population study. Community Dent Oral Epidemiol. 2017;45:258–65. 12. Åstrøm AN, Ekback G, Ordell S, Nasir E. Long-term routine dental attendance : influrnce on tooth loss and oral health related quality of life in Swedish older adults. Community Dent Oral Epidemiol. 2014:460–9. 13. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50:179–211. 14. Armitage CJ, Conner M. Efficacy of the theory of planned behavior. A meta analytic review. Br J Soc Psychol. 2001;40:471–99. 15. McEachan BA, Taylor N, Harrison R, Lawton R, Gardner P, Conner M. Meta- analysis of the reasoned action approach to understanding health beaviors. Ann Behav Med. 2016:50–592. 16. Luzzi L, Spencer AJ. Factors influencing the use of public dental services: an application of the theory of planned behavior. BMC Health Serv Res. 2008;8: 17. Sniehotta F, Scholz U, Schwarzer R. Action plans and coping plans for physical exercise: a longitudinal intervention study in cardiac rehabilitation. Bristish J Health Psychol. 2006;11:23–37. 18. Sheeran P, Orbell S. Using implementation intention to increase attendance for cervical cancer. Health Psychol. 2000;19:283–9. 19. Astrom AN. Applicability of action planning and coping planning to dental flossing among Norwegian adults: a confirmatory factor analysis approach. Eur J Oral Sci. 2008;116:250–9. 20. Sieverding M, Ciccarello L, Matterne U. What role do social norms play in the context of men’s cancer screening intention and behavior? Application of an extended theory of planned behavior. Health Psychol. 2010;29:72–81. 21. Lee H. The role of descriptive norm within the theory of planned behavior in predicting Korean American’s exercise behavior. Psychol Rep. 2011;109:208–18. 22. Cialdini RB, Reno RR, Kallgren CA. A focus theory of normative conduct: recycling the concept of norms to reduce littering in public places. J Pers and Soc Osychol. 1990;58:1015–26. 23. Rivis A, Sheeran P. Descriptive norms as an additional predictor in the theory of planned behavior: a meta-analysis. Curr Psychol: Dev, Learn Pers, Soc. 2003;22:218–33. 24. Schumacker RE, Lomax RG. A Beginner’s guide to structural equation modelling. second ed. London: Lawrence Erlbaum Associates Publishers;

Journal

BMC Oral HealthSpringer Journals

Published: May 31, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

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

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

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.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off