The school climate is the element that differentiates one school from another, it has the power to influence the behavior of students and teachers (Rudasill et al., 2018; Syahril and Hadiyanto, 2018). While being a multidimensional concept (Grazia and Molinari, 2020), which defines school climate in four ways: academic, community, safety, and institutional environment (Wang and Degol, 2016), school climate can be understood in terms of the feelings and attitudes that students and teachers have due to the school environment in which they carry out their activities (Loukas, 2007; Thapa et al., 2013). Over time, many scales and surveys were developed in order to measure school climate (Kohl et al., 2013; Grazia and Molinari, 2020). One of these surveys is the What is Happening in This School Questionnaire—which comprises five dimensions: teacher support, peer connectedness, school connectedness, affirming diversity, rule clarity, reporting irregularities, and seeking help (Aldridge and Ala’I, 2013).

School climate reflects peoples’ experiences in the school environment, and it refers to a wide range of elements such as norms, values, interpersonal relationships, or learning and teaching practices (Cohen et al., 2009; Thapa et al., 2013). It has a major role in the life of students both from an educational and social perspective (Rudasill et al., 2018). Previous studies that focused on the connection between school climate and students’ socio-emotional health revealed that students who perceive their schools as supportive and well-structured environments in which teachers are treated with respect by other students have better socio-emotional outcomes (Wang and Degol, 2016; Larson et al., 2020; Wong et al., 2021).

The individual’s social behavior can be considered the result of its interpretation of the event or social situation, the situation to which the individual attributes values and meanings (Lubsky et al., 2016). Also, it can be discussed in terms of prosocial and antisocial behaviors. Prosocial behavior can be understood as the type of actions a person carries out which benefit other people (Pfattheicher et al., 2022). In order to develop such actions, individuals must pay attention to the needs, desires, or goals of other people (Staub, 1978). Antisocial behaviors involve actions and attitudes that are regarded as dysfunctional and which can have negative consequences at the individual and societal level (Byrd et al., 2014). Such actions include acts of bullying, domestic violence, or discrimination (Hashmani and Jonason, 2021).

Referring to students’ social behavior in relationship with the school climate, previous studies have shown that a negative school climate can determine students to have negative behavior, and even engage in actions that involve the victimization of other students (Giovazolias et al., 2010; Wang et al., 2013). The influence of school climate on the social behavior of students was also studied in the context of the phenomenon of bullying and delinquent behaviors. Researchers discovered significant correlations between school climate and bullying or victimization behaviors, arguing that the occurrence of such a phenomenon could be diminished by improving the school climate (Wang et al., 2013; Aldridge et al., 2018).

Religion is considered a unitary system that comprises practices and beliefs that are related to sacred things and that `unite into one single moral community called a Church (Durkheim, 1995). Religiosity is a multidimensional concept that implies religious affiliation, religious actions—going to church, praying, and religious beliefs and faith in divinity (Bjarnason, 2007). Religious beliefs, together with other elements such as culture or science influence the way individuals carry out their daily lives (Johnson et al., 2011). Considering the American context, in today’s society, religion and religiosity tend to be described more through the actions carried out by people at a certain time and place, such as going to church on Sundays (Williams, 2015). In this regard, the idea that religion shapes the behavior of people became a subject of interest for many researchers.

A previous study that focused on the matter of religiosity and peoples’ tendency to help others revealed that religious people were involved more in volunteering activities and were more willing to donate to religious organizations (Jackson et al., 1995). Another study emphasizes the role of religiosity in peoples’ decision to volunteer, showing that individuals who go to church regularly tend to volunteer more than people who do not usually go to church (Ruiter and De Graaf, 2006).

Religiosity has also been studied in relation to subjective well-being. Even though happiness and life satisfaction can be affected by the individual’s health or the individual’s relationships with other people (Goian, 2014), religiosity can influence positively the social motivation. Researchers show that religiosity was associated with behaviors that encouraged social affiliation (Van Cappellen et al., 2017). The ability of students to socially affiliate and to socialize is also influenced by their relationship with their parents (Grusec, 2011). If parents get more involved in the child’s activities, the child will develop better communication abilities (Goian, 2019).

Data were collected from 353 high-school students through a questionnaire. The questionnaire was applied in classrooms during courses between January and February 2020, which corresponds to the second semester of the 2019–2020 school year. The questionnaire was self-administrated and the average time needed to complete it was 30 min. The research population included students from high schools in Timișoara, Romania. Timișoara is part of the Banat region of Romania, a well-developed region from the perspective of educational institutions, and the perspective of institutions involved in social work or public organizations (Goian, 2013). For sampling, we used data based on the ranking of the high schools according to the results obtained at the baccalaureate exam in the June–July 2019 session. At the top of the high schools, we were primarily interested in the confessional high schools in order to be able to select non-confessional high schools as well.

To choose the appropriate sample, the following criteria were applied: (a) The four confessional high schools in Timișoara, with the mention that at the time of the research, one of these high schools was not confessional but included high school classes specializing in theology; (b) Four non-confessional high schools were selected, which had to be theoretical, not vocational (e.g., sports, artistic, pedagogical, and bilingual), state high schools, and located in the municipality of Timișoara. These schools were chosen according to the ranking of the confessional high schools, ensuring similarity in the baccalaureate exam results; (c) Due to the purpose of the research, the sample included students from the 10th to 11th grades, excluding those absent on the day the questionnaire was applied. The consideration underlying this selection takes into account, on the one hand, the fact that the students in the selected classes are in the middle of adolescence (16–17 years old), and on the other hand, the fact that they spent at least 1 year (9th grade) in the confessional school, so they can differentiate and perceive the specific school culture. Taking into account the aspects mentioned above, we included in the study four confessional schools and four non-confessional schools.

The participants consisted of 353 high school students, including 166 from confessional schools and 187 from non-confessional schools. Most students are aged between 15 and 18 years old, with an average age of 16.5 years old. Among the participants, 21 are 15 years old, 149 are 16 years old, 165 are 17 years old, and 18 are 18 years old. The data indicate that there are 161 boys (male—45.60%) and 192 girls (female—54.40%) (Table 1).

The questionnaire, which can be found in Appendix A, comprises four sections named A, B, C, and D, each section measuring a certain concept. The three concepts measured are: school climate, social behavior, and religiosity.

Section A includes items specific to the school climate, which was measured through an instrument elaborated by Aldridge and Ala’I (2013), called WHITS—What’s Happening In This School Questionnaire. The instrument was translated and adapted to the Romanian cultural context and it comprises six dimensions through which school climate is measured.

Section B includes items specific to social behavior: prosocial and antisocial behavior. The items were adapted from a scale used in previous studies (Padilla-Walker et al., 2018) that measures the general prosocial behavior of students toward individuals they do not know. Antisocial behavior is measured through two dimensions: aggressivity and delinquency.

Section C includes items specific to religiosity. Religiosity was measured by taking into account five dimensions. The first dimension—religious faith is measured through two items (C1 and C2 in the questionnaire), that were previously used in the World Values Survey in 2012 for the Romanian population (Inglehart et al., 2014). The other four dimensions: private religious practices, organizational religiousness, overall self-ranking, and religious preference/affiliation were taken and adapted from the instrument Brief Multi-dimensional Measure of Religiousness/Spirituality—BMMRS (Fetzer Institute, 2003).

Section D includes items referring to the respondents’ socio-demographic characteristics: high school, grade, age, and gender.

The data were analyzed using IBM SPSS Statistical Package for the Social Sciences, version 20. To measure the reliability of the scales, a Cronbach test was performed for each of the dimensions of the measured concepts. The results of all the Cronbach tests showed values above 0.7, proving the reliability of the scales. Additionally, to validate and adapt the scales to the Romanian cultural context, an exploratory factor analysis was conducted for each of the analyzed concepts. According to the exploratory factorial analysis, school climate is measured through six factors: relationship with teachers, relationship with peers, affirming diversity, clarity of rules, reporting irregularities, and seeking help. Social behaviors are measured through three factors: prosocial behavior, antisocial behavior—aggression, and antisocial behavior—delinquency. In the case of religiosity, only one factor was extracted, which was named religiosity.

In this study, results with p < 0.05 are considered statistically significant, but those with p < 0.01 (highly significant statistical differences) or p < 0.005 (very significant statistical differences) are deemed to provide stronger evidence against the null hypothesis. Results with p < 0.001 (very highly significant statistical differences) demonstrate the highest level of statistical certainty, suggesting the effects are highly unlikely to be due to chance.

The research results revealed differences in school climate between the two types of schools, particularly regarding the relationship with teachers. In the case of students from confessional schools, the results showed more positive experiences concerning the support they receive from teachers. Compared to students from non-confessional schools, a higher percentage of students from confessional schools consider that teachers have a positive attitude toward the problems of the students and make an effort to understand them—t (353) = 4.442, p < 0.001—very highly significant statistical differences (Table 2).

Regarding the relationship with school in the context of students’ feeling of being part of that school, the results presented in Table 2, t (353) = 2.407, p < 0.05, reveal significant statistical differences between students of confessional and non-confessional schools in terms of feeling part of their school. Specifically, students from confessional schools declared to a higher extent than students from non-confessional schools, that they have a stronger feeling of being part of their school.

Concerning the fourth dimension of school climate—affirmation of diversity—the results from Table 2, t (353) = 3.523, p < 0.001, revealing very highly significant statistical differences, show a higher level of positive answers in the case of students from confessional schools, concerning their integration and appreciation of their culture in the school environment. In this regard, students from confessional schools, compared to students from non-confessional schools, consider to a higher extent that the school climate is integrative.

While referring to the dimension of rule clarity, the research revealed a greater tendency toward knowledge and appreciation of the rules among students from confessional schools (134 students out of 166). Table 3 shows that, for most of the time, students are aware of their school’s rules.

Regarding the prosocial behavior of students, our findings indicate that students from confessional schools have a higher level of prosocial behavior manifested through actions involving offering help. Table 4 presents the attitude of students toward offering help to people they do not know. The results in Table 4 [t (353) = 2.144, p < 0.05—significant statistical differences] revealed that students from confessional schools, compared to students from non-confessional schools, registered a higher tendency to help people they do not know, even if doing so is not always easy for them.

In order to understand better the differences between students’ social behavior depending on the type of school they attend, we performed a t-test. The results revealed that there are significant differences in the way students behave depending on the two types of schools (Table 4). Specifically, very highly significant statistical differences can be seen in the context of factor 2—antisocial behavior—delinquency: t (351) = −4.76, p < 0.001, and significant statistical differences in the context of factor 3—prosocial behavior: t (351) = 2.17, p = 0.03 (p < 0.05). Our findings revealed that in the case of confessional schools delinquent behavior registers lower levels, and with an error of 5%, the result shows that in confessional schools, students tend to engage more in prosocial behaviors.

To establish the level of association between students’ faith in God, or other divinity depending on the type of school they study in, a Chi-Square test was performed. Since the value of the test was χ² = 36.457, df (3) with p = 0,000 (p < 0.001), it can be affirmed that the respondents from the two types of schools present very highly significant differences in terms of their faith in God, and these differences are not a result of the random sampling variation (Table 5).

As expected, the results point toward high levels of religious private and public behavior for students from confessional high schools. The existence of the differences between students’ level of religiosity is also highlighted by the results of the t-test. Thus, since t (351) = −15.57, p < 0.001—very highly significant statistical differences, with an error of 1%, we can affirm that compared to students from non-confessional schools, those who study in confessional schools have higher levels of religiosity.

The analysis of the correlations of the three factors confirms the influence that religiosity has on school climate in both types of schools (Table 6). Even though the correlation can be considered weak (r < 0.30), higher values, at the significance level of 1% were registered in the case of confessional schools due to the high level of students’ religiosity. The strongest correlation can be seen in the case of students’ relationship with their school [r (164) = −0.29, p < 0.01—highly significant statistical differences], a result which strengthens the idea that students in confessional schools feel they are part of the community and they are respected, because of the intersection of the religious values of each student with the values of the school’s organizational culture. In the non-confessional school environment, a higher level of religiosity determines a higher level of students’ actions regarding affirming their diversity: r (185) = −0.23, p < 0.01—highly significant statistical differences (Table 6).

Our findings reveal correlations between religiosity and the social behavior of students. An increase in the level of religiosity is associated with a decrease in the tendency to engage in antisocial behavior and with an increase in the tendency to adopt prosocial behavior (Table 6). In confessional schools the correlation has a medium level of intensity between religiosity and students’ behavior manifested through aggressivity [r (164) = 0.39] and delinquency [r (164) = 0.34], at a level of highly significance p < 0.01. The results show that as students’ religiosity increases, their level of antisocial behavior decreases.

We used binary logistic regression modeling to find out what factors influence the perceived religiosity of the students. To answer this question, we defined the following variables: (i) Dependent variable: religiosity (with values 0—Little and not at all religious and 1—Very or moderately religious) obtained from item 10 in the questionnaire and (ii) Independent variables: school integration (or “school integration index” is a summative index obtained from variables i1–i48) and sociability (or “sociability index” is a summative index obtained from variables B1–B10 after the recodification of the variables in which B6–B10 were recoded). These two indexes have the descriptive values that are presented in Table 7.

Other categorical variables used in this model were: high school (with values 1—Confessional and 2—Non-confessional), gender (1—Male, 2—Female), and age (1–15 years old, 2–16 years old, 3–17 years old, 4–18 years old).

For logistic regression, the basic assumptions are: “independence of errors, linearity in the logit for continuous variables, absence of multicollinearity, and lack of strongly influential outliers” (Stoltzfus, 2011). All the assumptions were fulfilled. The multicollinearity among these two predictors from Table 7 was avoided (Spearman rho = 0.361, p < 0.001—very highly significant statistical differences). In addition, in all cases VIF < 10 guarantees the absence of multicollinearity. With the procedure Casewise List of extreme values we verified that there are no extreme values in the model. Finally, it is important to note that the observations were independent of each other.

In our binary logistic model with many independent variables, the formula can be: lny = bo+b1*school integration+b2*sociability+b3*high school+b4*gender+b5*age.

As is known, binary logistic regression operates with data in two blocks: (i) In Block 0 none of the predictors are in the model. From the classification table, we observed that the percentage of prediction rate is 64.3%; (ii) In Block 1, we observed the Omnibus Tests for Model Coefficients table. This indicates that the model is better than Block 0 (Chi-Square = 66.458, df = 7, p < 0.001—very highly significant statistical differences).

In the table Model Summary in SPSS Output, we tested the goodness of fit of the model. Using Cox & Snell R Square and Nagelkerke R Square, we decide that 23.6% of the variation in religiosity is explained by all variables included.

With Hosmer and Lemeshow test, we can conclude that the data proposed by the model are close to the real data [Chi-Square = 11.295, df = 8 and p = 0.186 (>0.05)].

Finally, how the table looks with all the variables included in the formula is presented in Table 8.

Binary logistic regression indicates that school integration, sociability, and high school are significant predictors for religiosity [Chi-Square = 66.458, df = 7 and p = 0.000 (<0.05)]. The other predictors gender and age are not significant. All the predictors “explain” 23.6% of the variability of religiosity. School integration was very highly significant at a confidence level of 95% (Wald = 12.92, p < 0.001); Sociability was very highly significant at a 95% confidence level (Wald = 11.13, p < 0.001) and high school was significant at a 95% confidence level. The odds ratio for school integration was 1.017 (95% CI 1.008–1.027); for sociability, the odds ratio was 1.067 (95% CI 1.027–1.108) and the odds ratio for high school was 2.765 (95% CI 1.676–4.563). The model correctly predicted 44.4% of cases where religiosity is low and 84.1% of cases where religiosity is high or moderate. Overall, the correctness of the prediction was 70%.