SEARCH

SEARCH BY CITATION

Keywords:

  • higher education;
  • super-empirical beliefs;
  • university faculty;
  • identity development

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

This study examines the impact of educational enrollment and attainment on several measures of religious belief using nationally representative panel data. Although college does not appear to substantially alter the religious beliefs of most emerging adults, findings do reveal a modest increase in skepticism toward super-empirical religious beliefs among college students and graduates compared to those who have never attended any form of postsecondary education. This effect is dependent on college type, with students attending elite universities exhibiting the greatest increase in skepticism. Apart from changes in super-empirical belief, graduating from college modestly increases preferences for institutionalized religion while simultaneously reducing adherence to exclusivist religious belief. Faculty commitment to secularism, the degree of student academic engagement, and developing social identities may play a role in religious belief change, particularly at elite universities.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

Recent years have seen a steady proliferation of scholarly research devoted to the religious and spiritual practices of college students (Astin, Astin, and Lindholm 2010; Bryant, Choi, and Yasuno 2003; Clydesdale 2007a; Hill 2009; Lee 2002; Reimer 2010; Saenz and Barrera 2007; Scheitle 2011; Uecker, Regnerus, and Vaaler 2007). The results paint a mixed picture of higher education's influence on the religious and spiritual life of students (Mayrl and Oeur 2009). Initial results suggested that college had little overall impact on religious practice (Uecker, Regnerus, and Vaaler 2007) and identity (Clydesdale 2007b); approximately half of students reported that their religious beliefs and convictions remained unchanged during college (Hurtado et al. 2007; Lee 2002). However, the influence of higher education on beliefs may be more complex. There is evidence that the college experience may liberalize the theological beliefs of some religious students (Reimer 2010) while simultaneously enhancing students’“inner” spiritual life (Astin, Astin, and Lindholm 2010). There is also increasing evidence that the influence of college is conditional on the subpopulation being studied as well as institutional context. For example, Hill (2008) finds that religious affiliation and salience of faith increases among African Americans who attend and graduate from college, but that affiliation and salience of faith declines among college-going Catholics. The impact of educational attainment on religious practice and belief also varies significantly by religious tradition (McFarland, Wright, and Weakliem 2010) and the religious affiliation of the college or university (Astin, Astin, and Lindholm 2010; Hill 2009).

Despite progress in specifying the impact of higher education on the religious and spiritual lives of young people, many of the most important dimensions of religiosity remain understudied. The research reported here focuses specifically on the impact of college attendance and graduation on the content of religious belief, arguably one of the most neglected topics in the recent literature (Mayrl and Oeur 2009; but see Reimer 2010 for a recent exception). The neglect of content is not due to lack of interest; many accounts of the secularizing influence of college primarily posit its impact in terms of religious orthodoxy, a composite of religious beliefs (see Feldman and Newcomb 1969; Hunsberger 1978; Hunter 1987; Madsen and Vernon 1983). The most likely reason for neglect is the unavailability of national data that contain multiple indicators of religious belief. The present study utilizes panel data available in the National Study of Youth and Religion (NSYR) to assess the impact of attending and graduating from college on change in religious belief from adolescence to emerging adulthood.

Religious Beliefand HigherEducation: Previous Research

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

Until recently, most accounts of religious belief in college assumed that college was a “breeding ground for apostasy” (Caplovitz and Sherrow 1977). Prior to the decade of the 1970s, numerous studies were accumulated by social scientists exploring this particular relationship (for a review, see Feldman and Newcomb 1969:23–28). The conclusions were tentative at best and often plagued by inadequate methodologies and limited samples, making causality claims and inference to the larger population difficult. An examination of the studies reviewed by Feldman and Newcomb (1969) reveals small samples selected from single colleges and universities with only a few having longitudinal data. Most of these studies use a college population to study cross-sectional differences between freshmen and seniors, finding seniors less orthodox in their belief, more skeptical about God, and less trustful in the institution of the church as compared to freshmen. Despite the lack of generalizability, the studies were numerous and consistent enough to establish what was considered a modest secularizing effect of college on students.

In an update to Feldman and Newcomb's summary, Pascarella and Terenzini (2005:284) claim recent research has countered these findings and suggest that religious beliefs “may not so much increase or decrease as become reexamined, refined, and incorporated in subtle ways with other beliefs and philosophical dispositions.” Ethnographic research on college campuses suggests college students are moving away from institutional expressions of religious faith and toward noninstitutional expressions of spirituality (Cherry, DeBerg, and Porterfield 2001). Recent research from the Spirituality in Higher Education Project corroborates this finding by documenting small to moderate growth in most measures of spirituality from freshman to junior year, very little change in religious commitment, and a slight increase in the percentage of the student body who report struggling with their faith (Astin, Astin, and Lindholm 2010). In a survey of college seniors, Lee (2002) reports that a substantial minority (38 percent) report the strength of their religious convictions increasing, while only 14 percent report a decline. Forty-eight percent of students report no change in their religious beliefs and convictions. Similarly, Hurtado et al. (2007) find that 56 percent of sophomores report no change in their religious beliefs since the beginning of college.

Most recently, based on data from a sample of church-going Protestants, Reimer (2010) found that theological liberalism is correlated with higher education. The type of institution (religious vs. secular) and exposure to secular theories and philosophies at college were the strongest predictors of theological liberalism. Other recent cross-sectional studies find negative relationships between education and the following: belief in heaven, hell, Satan, and demons (Baker 2008; Rice 2003), a literal interpretation of the Bible (McFarland, Wright, and Weakliem 2010), paranormal beliefs such as ESP, psychic powers, and communicating with the dead (Baker and Draper 2010; Orenstein 2002), and theistic certainty (Sherkat 2008).

Despite some progress in our understanding of the impact of higher education on religious belief, several shortcomings need to be addressed. First, measures of religious beliefs are often simplistic (e.g., belief in God) or unrelated to traditional religious beliefs (e.g., measures of spirituality). Beliefs about the ultimate validity of religious truth claims, the afterlife, reincarnation, angels, and similar questions are mostly absent. Secondly, a noncollege sample is not incorporated for comparison. Maryl and Oeur (2009) highlight the importance of not relying solely on samples of college students. Without a sample of emerging adults who do not attend college it is impossible to specify the impact of college on religious belief. Lastly, most studies of religious belief (with the exception of the Spirituality in Higher Education Project) are neither longitudinal nor nationally representative. Longitudinal studies allow for establishing a baseline belief measure before the college experience and assessment again at a later time. The reliance on self-reported changes in religious beliefs is subject to memory distortion and social desirability bias. A national sample is important to confidently infer the overall impact of attending college on religious beliefs. This research addresses these concerns by using nationally representative panel data with multiple measures of religious belief replicated over time.

Hypothesizing Religious Changein College

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

Before hypothesizing about the effect of college on specific dimensions of religious belief, two distinct types of religious belief are considered: super-empirical beliefs and beliefs about institutionalized religion and religious exclusivity.

Super-Empirical Beliefs

Many religious beliefs encompass belief in a realm that is not immediately accessible to the human senses (Robertson 1970; Smith 2003b:98–99; Stark and Finke 2000:89–90). Belief in the existence of God, angels, demons, and an afterlife are central components of conventional Christian belief (although not all are necessarily exclusive to Christian doctrine). Other super-empirical beliefs are associated with non-Christian traditions (e.g., reincarnation in Hinduism or the belief in kami for Shinto practitioners), and still others fall outside of most major world religious traditions (e.g., ESP and astrology). The relationship between Christian super-empirical beliefs and other beliefs has been the subject of debate in the social scientific literature (Rice 2003). Those who strictly adhere to conventional Christian beliefs are less likely than the general population to believe in non-Christian super-empirical beliefs (Baker and Draper 2010; Stark 2008). However, among those whose adherence to conventional Christian belief is less strict, there is a positive association between Christian and other super-empirical beliefs, perhaps due to a preference for or against super-empirical explanations (Baker and Draper 2010; Weeks, Weeks, and Daniel 2008).

Among U.S. emerging adults, super-empirical beliefs tend to cluster into Christian and non-Christian groupings. Analyzing these concepts separately helps to assess whether attending and graduating from college influences super-empirical beliefs more generally, or whether higher education is more likely to influence beliefs typically associated with the Christian tradition.

Beliefs About Institutionalized Religion and Religious Exclusivity

Beliefs about the commensurability (or incommensurability) of existing religious systems, as well as beliefs about the legitimate carriers and arbiters of religious truth claims (e.g., institutions vs. individual conscience), should be differentiated from the super-empirical content of religious belief. Such beliefs are not propositions about the nature of ultimate reality, but rather stances toward how the individual should properly relate to the existing cosmology of belief systems. These are, in some sense, beliefs about religion as opposed to religious beliefs.

These beliefs are analyzed as a distinct set, separate in kind from super-empirical belief. Although there are methodological reasons for doing so, separating out such beliefs also makes sense in light of the existing literature. Although Americans frequently profess belief in God, an afterlife, the divinity of Christ, the authority of scripture, and other beliefs specific to conventional Christian doctrine (Gallup and Lindsay 1999:21–42; Stark 2008), most do so in a nonsectarian manner. Survey research has repeatedly confirmed that most Americans do not hold exclusivist conceptions of their own faith traditions (Putnam and Campbell 2010; Stark 2008; Woolever and Bruce 2010), choosing to believe that most major world religions are essentially compatible with one another and provide legitimate, and many believe equal, access to the divine. In addition to this, the majority of Americans believe that, in matters of faith, individual conscience trumps the authority of religious institutions (Bellah et al. 1985; Dillon and Wink 2007; Smith and Snell 2009). Beliefs about the essential commensurability of all religion and the epistemic grounds of faith are often inconsistent with the official doctrine and leadership of particular faith traditions (see Putnam and Campbell 2010:538–40 for an account of the gap between clergy and lay adherents in matters of religious exclusivity). Smith and Snell (2009) document similar themes in their interviews with 18- to 23-year-old emerging adults. Although many of their interviewees confirm traditional religious beliefs, most also affirm that most religions are essentially good and that the truth claims and practices of any religion should ultimately be chosen by the individual, including the right to pick and choose what aspects of a faith tradition to adopt. Adhering to the tenets of a particular faith tradition is not necessarily part of the same package as religious exclusivity and authority, as suggested by the analysis that follows. Beliefs concerning the unique authority of religious institutions as well as beliefs about religious exclusivity are analyzed separately from belief in super-empirical entities and occurrences.

Social Influences from Higher Education on Religious Beliefs

Perhaps the most commonly assumed source of secular influence can be traced to the halls of academia itself. Exposure to secular theories and methods in the classroom, and the “expanded horizons” (Hadaway and Roof 1988) associated with liberal learning, has long been assumed to be at the heart of the secularizing influence of college in both popular (see Budziszewski 1999; Nye 2005; Wheaton 2005) and scholarly accounts (see Hadaway and Roof 1988; Hunter 1987). Research on religious liberalization of church-going Protestants lends some credence to this account; respondents who report exposure to “secular theories and philosophies” also score higher on religious liberalism (Reimer 2010). Moreover, research on the religious beliefs and practices of university scientists, often seen as the leading edge of the academy, find that they are considerably less religious than the general public (Ecklund 2010; Gross and Simmons 2007, 2009). In her study of elite university scientists, Ecklund (2010:93) concludes that the single largest group of university scientists in her study “endorsed Enlightenment thinking, arguing that reason is the primary authority and tending to privilege science over other forms of knowledge.” Exposure to this epistemological narrative in the classroom, from the trusted authority of college faculty, might affect the plausibility of certain religious beliefs, particularly belief in the super-empirical realm not directly accessible through reason or empirical methods.

An alternative source of secular influence documented in the literature focuses less on the pedagogy of the classroom and more on the worldview pluralism of faculty and students on college campuses (Becker 1977; Hammond and Hunter 1984; Hill 2009; Madsen and Vernon 1983). As students move from their own religious subculture to a religiously pluralistic campus, the “plausibility structures” (Berger 1967) that gird religious belief may be undermined. If religious belief is largely a “background assumption” (Berger and Zijderveld 2009), then the social interaction on a college campus may force these religious beliefs into the foreground. This does not necessarily suppose a decline in traditional beliefs. Moving religious assumptions into the foreground may actually strengthen particular religious beliefs, although most secularization accounts focus on the faith-undermining influence of pluralism. Students influenced by these forces on college campuses might be more likely to embrace nonexclusivist religious claims and eclectic expressions of spirituality over institutional expressions of religious faith to ease the cognitive dissonance associated with worldview pluralism.

However, it is also possible that the previous two social sources of religious influence have little real-world influence given the nature of religious belief and identity for many young emerging adults. The cultural worlds of adolescents may produce a set of religious beliefs that are disjointed and individualistic (see Smith and Denton 2005:172–92). As students enter college, these beliefs are compartmentalized and never engaged within the classroom or in other social interactions. Consistent with this account, most college students see no conflict between religion and science (Scheitle 2011). Religious beliefs belong to an “identity lockbox” (Clydesdale 2007b) and remain protected during the college years precisely because they are fragmented and private. This alternative account posits that religion is peripheral to the identity of most students and not subject to revision during the transition to college life.

Religiously Affiliated Colleges and Universities

The hypothesized sources of religious change outlined above operate within an institutional context, and this context can be expected to heighten or temper the impact of the social processes. Recent work on religion and higher education has confirmed that institutional identity is an important factor in understanding religion among college students (Bryant and Astin 2008; Cherry, DeBerg, and Porterfield 2001; Freitas 2008; Hill 2009). Religiously affiliated institutions (particularly evangelical Protestant colleges) are more deliberate in their incorporation of religious practice and belief into the curriculum and social activities of students (Kuh 2000; Kuh and Umbach 2004; Wolfe 2006). Hill (2009) claims that some conservative Protestant institutions can be conceived of as “moral communities” (borrowing the term from Stark [Stark 1996; Stark, Doyle, and Kent 1980]), noting that their collective religious identity and religiously homogenous student population helps to legitimize individual-level religious practices. Although institutions that maintain a religious identity will vary in the centrality of this identity, across all such institutions the intensity and frequency with which the social mechanisms outlined above are operating would be lessened. For example, the majority of faculty at religiously affiliated institutions are either religious themselves, or have a more accommodating stance toward religious belief and practice than members of their guilds at nonreligious institutions (Gross and Simmons 2007; Higher Education Research Institute 2006). The religious composition of students and faculty will almost certainly be more homogenous on these campuses when compared to similar secular institutions, thus lessening the impact of religious and worldview pluralism. Moreover, because attending a religious college is voluntary, the religious beliefs of entering students will likely be less fragmented and private than the typical student attending a nonreligious college or university.

Elite Universities

Historically, many of the leading American universities were the first to adopt the model of the secular German research university toward the close of the 19th century (Marsden 1994; Smith 2003a). The faculty, previously under the authority of the church, increasingly conceptualized their role as knowledge experts within specialized fields of studies. As ecclesial authority waned, and the modern notion of academic freedom came to dominate the ethos of these institutions, these historical accounts contend that the general orientation of faculty became critical and skeptical of institutional authority, particularly ecclesial authority (Burtchaell 1998; Marsden 1994). Although these disputes with religious authorities have long-since passed for the modern research university, recent research has confirmed that faculty at elite universities are still more committed to establishing secularism in public institutions and are privately less religious than their counterparts at other postsecondary institutions (Ecklund 2010; Gross and Simmons 2007).

The enduring secular orientation of faculty at elite institutions may be partially due to their role as representatives of their guilds to the public and other political and religious elites in society. Wuthnow (1985) explains that irreligion may function as a “boundary-posturing mechanism” for many university scientists, helping to maintain the distinction of their disciplines, and allowing them to symbolically separate themselves from the general public. The scientists who actively participate in this “boundary-posturing” disproportionately reside in Ph.D.-granting research universities. Consistent with this expectation, Gross and Simmon's (2009) recent analysis of the religiosity of the professoriate finds that faculty members with a “research orientation” are, on average, less religious across multiple measures of practice and belief. These high-profile, high-research faculty are the public face of their guilds, and consequently have the highest stakes in guarding the meaning of their respective disciplines.

Thus, classroom exposure to theories and ideas that are at odds with conventional faith beliefs may be a more salient feature of education for students attending elite institutions. Not only are the faculty at these universities more secular in orientation, students themselves are more academically engaged. Arum and Roksa (2011) report that students at selective postsecondary institutions are more engaged with faculty, read more, write more, and spend more hours per week studying. The combination of engaged students and faculty with a secular orientation may lead to greater exposure to ideas that threaten some religious beliefs.

The elite status of these postsecondary institutions may also contribute to identity work among students that has consequences for religious belief. Students at elite institutions work harder to differentiate themselves from the general public. If attending a well-regarded academic institution confers high social status, students may be more receptive to the academic socialization process at these universities. Faculty at elite institutions are more secular than their counterparts elsewhere, and, in addition, the students attending elite institutions are more willing to be resocialized as knowledge-class professionals with little attachment to conventional conceptions of religious faith. This is unlikely to be detrimental to public religious participation, which is often associated with middle-class civic participation (Hill 2009; Uecker, Regnerus, and Vaaler 2007), but may be accompanied by a religious skepticism evidenced in measures of conventional religious belief.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

Data

Data come from the first and third waves of the NSYR and from the Integrated Postsecondary Educational Data System (IPEDS). The NSYR's longitudinal telephone survey began as a nationally representative telephone survey of 3,290 English- and Spanish-speaking teenagers between the ages of 13 and 17. The baseline survey was conducted, with the teen respondents and one of their parents, between July 2002 and April 2003 by researchers at the University of North Carolina at Chapel Hill. A random-digit dial telephone method was employed to generate numbers representative of all household telephones in the 50 states of the United States. Also included were 80 oversampled Jewish households, not nationally representative, bringing the total number of completed cases in the first wave of NSYR to 3,370. The second and third waves of the NSYR are resurveys of the Wave 1 English-speaking teen respondents. All waves of the survey were conducted by telephone using a computer assisted telephone interviewing system. Wave 3 of the survey was fielded from September 2007 through April 2008 when the respondents were between 18 and 23 years old. Every effort was made to contact and survey all original NSYR respondents, whether they completed the Wave 2 survey or not, including those out of the country and in the military. In Wave 3, 2,532 original youth respondents participated in the survey for an overall Wave 1 to Wave 3 retention rate of 77.1 percent. The main source of attrition in the third wave was nonlocated respondents. The Wave 3 refusal rate, calculated as the number of eligible respondents (3,282) who refused, was 6 percent.

Diagnostic analyses comparing NSYR data with U.S. Census data on comparable households and with comparable adolescent surveys—such as Monitoring the Future, the National Household Education Survey, and the National Longitudinal Study of Adolescent Health—confirm that the NSYR provides a nationally representative sample without identifiable sampling and nonresponse biases of U.S. teenagers ages 13 to 17 and their parents living in households (for details, see Smith and Denton 2003). For descriptive purposes, a weight was created to adjust for number of teenagers in household, number of household telephone numbers, census region of residence, and household income. A separate weight is used in multivariate analyses that control for census region and household income; this weight adjusts only for number of teenagers in household and number of household telephone numbers.

College and university institutional characteristics were merged into the NSYR data from the IPEDS. IPEDS is an annual census of postsecondary educational institutions in the United States conducted by the National Center for Educational Statistics. The IPEDS codes were used to identify the religious affiliation of postsecondary institutions in the NSYR data. Missing data are minimal and handled through listwise deletion.

Dependent Variables

The NSYR contains a number of variables that measure the content and structure of religious belief. Using data-reduction methods similar measures of belief were grouped to avoid redundancy of content as well as reduce the overall number of coefficients in the models and thus the number of type I errors. Some dependent variables are composite measures in the final models, but additional models (not shown) of each dependent variable were also examined.

Twelve indicators of religious belief are measured identically at Wave 1 and Wave 3. The original wording for these survey items can be found in Appendix A (available online at wileyonlinelibrary.com). Beliefs about a judgment day were recoded to account for those who were legitimately skipped because they do not believe in the existence of God. Views about religion were recoded into a dichotomous variable (belief that only one religion is true vs. other beliefs about religion). Using 12 indicators, a factor analysis was performed on a matrix of polychoric correlations. After oblique promax rotation,1 six of the 12 variables had significant loadings of .5 or greater on the first factor. An additional two variables loaded highly on another factor. The remaining four variables were each associated with separate factors.

The six variables that loaded onto a single factor consist of the following: belief in God, belief in a judgment day, belief in an afterlife, belief in angels, belief in demons, and belief in the possibility of miracles. This group of variables measures conventional Christian super-empirical belief. Belief in reincarnation and belief in astrology loaded onto a single factor, measuring non-Christian super-empirical belief. An additive index was constructed for both sets of Wave 3 variables.2 Change score variables were constructed by creating an additive index using Wave 1 data, subtracting Wave 1 scores from Wave 3 scores, and then dividing change scores by their standard deviation to make interpretation in the full models more meaningful.3

The factor analysis indicated that the remaining four variables should not be reduced. The remaining items are: identification as spiritual but not religious, belief in the necessity of a religious congregation to be truly religious, belief that only one religion is true, and the acceptability of picking and choosing specific beliefs without accepting religious teachings as a whole. While there is not a single construct measured by these items, they can be thematically grouped together under the label of attitudes toward institutionalized religion and religious exclusivity. These align most closely with the second category of beliefs previously outlined.

Independent Variables

The primary independent variables are measures of educational enrollment and attainment. Respondents were first asked whether they were currently enrolled in school. Those who were enrolled in a postsecondary institution (vocational or technical school, community/junior college, or a college/university) in Wave 3 were assigned IPEDS codes (N= 1,335). Twenty-five of these respondents had obtained a bachelor's degree and an additional 32 respondents reported being currently enrolled in a postsecondary institution but could not be matched to existing IPEDS codes (or the IPEDS code did not match an institution), for a total of 1,342 respondents currently enrolled who had not yet obtained a bachelor's degree. Among these respondents dichotomous variables were generated for those who were attending a religiously affiliated institution (N= 144) and those attending an elite university (N= 93). Following Gross and Simmons (2007), elite institution is defined as being among the top-ranked 51 universities in the 2007 US News and World Report's college rankings. For the remaining respondents, the Carnegie classification was used to categorize institutional type where possible. Forty-nine cases were unable to be matched to the Carnegie classification system or were among the 32 respondents who did not have an IPEDS code or could not match their IPEDS code to an existing institution. In these cases, the respondent's self-report was used to classify the institutional type. The remaining currently enrolled respondents were then coded as attending a postsecondary institution that primarily grants associates degrees (N= 422), a nonelite doctorate or masters university (N= 585), or a baccalaureate institution (N= 82). An additional 16 cases were classified as special-focus colleges and are included in the model, but the coefficient is not shown due to the small number of cases. All of the classification categories are mutually exclusive, with the exception of religiously affiliated institutions and elite institutions (there are eight cases that belong to both categories).

Wave 3 respondents who were not currently enrolled in a postsecondary institution, had not attained a bachelor's degree, but reported attending a postsecondary institution in any wave of the NSYR were coded as attending some college (N= 274). Those that reported obtaining a bachelor's degree in either Wave 2 or Wave 3 were coded in a separate dichotomous variable (N= 89). The remaining Wave 3 respondents had never attended any postsecondary educational institution (N= 835), and function as the reference category for all of the dichotomous educational enrollment and attainment variables.

The lagged indexes of super-empirical beliefs from Wave 1 and measures of religious identity and practice from Wave 1 are also included. Using the lagged initial values is common practice in change score models and can be considered a version of the conditional change panel model (Finkel 1995; Plewis 1985). The estimated coefficient for these lagged endogenous variables is referred to as the “stability coefficient” and is used as a control for the effect of initial religious beliefs on change in belief over time. The additional religious indicators are measures of religious service attendance, self-rated importance of faith, frequency of prayer, frequency of scripture reading, and a measures of religious tradition (using the RELTRAD method [Steensland et al. 2000]). These variables estimate whether religious identity and practice during adolescence predict change in religious belief over time. All models include controls for gender, race/ethnicity, age, region of residence, and relational closeness to parents from Wave 1 and a measure of current residence from Wave 3. Table 1 presents descriptive statistics.

Table 1.  Descriptive statistics, unweighted
 NMeanSDMinMax
  1. Source: National Study of Youth and Religion 2002–2003, 2007–2008.

Dependent variables
  Christian supernaturalism scale (change score)2,440−.131−3.943.29
  Non-Christian supernaturalism scale (change score)2,488−.011−2.492.49
  Spiritual but not religious (W3)2,5031.77  .7013
  Do not need congregation to be religious (W3)2,513.75 01
  Only one religion is true (W3)2,482.28 01
  Okay to pick beliefs without accepting whole faith (W3)2,500.52 01
Currently enrolled
College/university type
  Religiously affiliated2,532.06 01
  Elite2,532.04 01
  Associates2,532.17 01
  Doctorate or masters university2,532.23 01
  Baccalaureate2,532.03 01
Currently not enrolled
  Never enrolled2,532.33 01
  Some college2,532.11 01
  Bachelor's degree2,532.04 01
Lagged variables
  Christian super-empiricism (z-score)2,469  01−2.791.07
  Non-Christian super-empiricism (z-score)2,495  01 −.972.70
  Spiritual but not religious (W1)2,4821.64  .6413
  Do not need congregation to be religious (W1)2,506.69 01
  Only one religion is true (W1)2,480.30 01
  Okay to pick beliefs without accepting whole faith (W1)2,483.47 01
Additional W1 religious variables
  Attended religious school2,532.08 01
Religious tradition
  Evangelical Protestant2,532.31 01
  Mainline Protestant2,532.12 01
  Black Protestant2,532.11 01
  Catholic2,532.24 01
  Jewish2,532.04 01
  LDS2,532.03 01
  Not religious2,532.11 01
  Other religion2,532.02 01
  Indeterminate2,532.02 01
Importance of faith2,5313.451.1315
Frequency of prayer2,5284.352.0017
Frequency of Bible reading2,5252.561.7117
Frequency of church attendance2,5304.242.1817

Plan of Analysis

For the two composite measures of super-empirical beliefs, models estimate the influence of educational status and attainment on the outcome variable net of standard demographic controls and the lagged endogenous variable. This approach provides a reliable estimate of the effect of attending various postsecondary institutions as well as graduating from college on change in religious beliefs. The second model for each outcome includes additional measures of religious identification and practice at Wave 1 to assess whether certain religious groups or religious practices are predictive of changes in religious belief. These variables also act as controls for possible spurious initial conditions that predict both educational outcomes and change over time in religious belief. If educational effects hold up in the second model, we can be fairly confident attributing differences in changing beliefs to the educational experience of students. Table 3 reports the proportion of emerging adults predicted to decline in conventional super-empirical beliefs by educational attainment and type of institution, holding all other factors at their means.

Table 3.  Predicted probabilities based on logistic regression predicting decline in supernatural belief between Wave 1 and Wave 3, weighteda
 (1)(2)
ChristianNon-Christian
Super-EmpiricismSuper-Empiricism
  1. Source: National Study of Youth and Religion 2002–2003, 2007–2008.

  2. aModels include all variables from Model 2 and Model 4 in Table 2 (not shown). Variables not shown are fixed at their mean.

  3. +p  <  .10, *p  <  .05, **p <  .01, ***p  <  .001.

Currently enrolled
College/university type
  Religiously affiliated.253.140
  Elite.585.198
  Associates.386.199
  Doctorate or masters university.379.209
  Baccalaureate.416.110
Currently not enrolled
  Never enrolled.335.169
  Some college.386.134
  Bachelor's degree.380.287

The second set of analyses examines the variables related to institutional religion and religious exclusivity. Individual Wave 3 measures are used as outcome variables instead of change scores and the models control for the lagged endogenous variable (i.e., the same measure at Wave 1). Because the majority of these four outcomes are dichotomous (one is ordinal), it becomes somewhat more complicated to convert the outcomes to meaningful change scores and still present parsimonious models (Menard 2010:265–67).4 All models include the same covariates as the full models in the previous analyses.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

Table 2 presents the results from an ordinary least squares regression of the composite measure of change in super-empirical Christian beliefs on key educational indicators and control variables. Higher values on the dependent variable indicate increased belief, while lower values indicate decreased belief. Model 1 measures the effect of college attendance and graduation. For all of the educational coefficients, the reference category is those who have never been enrolled in any form of higher education. These coefficients indicate more decline in super-empirical Christian beliefs among emerging adults who are currently attending or have graduated from college compared to those who have never enrolled. The largest effect is found for those who are currently attending an elite university. These students become nearly one half of a standard deviation less super-empiricist in their belief change than those who have never attended college. We cannot be statistically confident that those who attend religiously affiliated institutions and those who have attended some college, but are not currently enrolled, are different from those who have never enrolled in terms of changes in Christian super-empirical beliefs.

Table 2.  Ordinary least squares regression predicting change in super-empirical beliefs, weighteda
 Christian Super-EmpiricismNon-Christian Super-Empiricism
(1)(2)(3)(4)
  1. Source: National Study of Youth and Religion 2002–2003, 2007–2008.

  2. aAll models include control variables for race/ethnicity, gender, age, relational closeness to parents, census region, and type of residence.

  3. bReference category is never attended college.

  4. cReference category is evangelical Protestant.

  5. +p <  .10, *p  <  .05, **p  <  .01, ***p  <  .001.

Currently enrolled
College/university typeb
  Religiously affiliated.138.0540−.113−.0982
  Elite−.448***−.383***−.238*−.263*
  Associates−.122*−.124*−.0946−.0845
  Doctorate or masters university−.160**−.156**−.0188−.0321
  Baccalaureate−.205+−.217+.0884.0976
Currently not enrolledb
  Some college−.0982−.120+.0656.0783
  Bachelor's degree−.249*−.238*−.208+−.220+
Exogenous lagged variable
  Christian super-empiricism−.341***−.550***  
  Non-Christian super-empiricism  −.251***−.266***
Additional W1 religious variables
  Attended religious school −.0573 −.116
  Religious traditionc
   Mainline Protestant −.392*** .0700
   Black Protestant −.278** .0645
   Catholic −.142* .0770
   Jewish −.514** .0814
   LDS .0650 −.171
   Not religious −.136 −.155+
   Other religion −.469*** −.0897
   Indeterminate −.494*** .00377
  Importance of faith .119*** −.00809
  Frequency of prayer .0614*** .0352**
  Frequency of Bible reading −.0125 −.0215
  Frequency of church attendance .0454*** −.0284*
N2,4232,4232,4572,457
R2.149.213.090.101

Model 2, Table 2 includes additional Wave 1 religious indicators. These variables indicate that the religious tradition of the respondent as an adolescent is an important predictor of belief change. Those raised in most nonevangelical religious traditions, with the exception of Mormons, are more likely to decline in their super-empirical Christian beliefs compared to evangelical Protestants. Moreover, importance of faith, frequency of private prayer, and frequency of church attendance as an adolescent are all predictors of positive relative change in Christian super-empiricism; frequency of Bible reading is the only nonsignificant religious practice. The inclusion of these additional variables does little to alter the interpretation of the educational effects. The effect of attending an elite college is modestly decreased, while the effect of having attended some college (but not being currently enrolled) is statistically significant at the .10 level. All effects from college, with the exception of being currently enrolled in a religiously affiliated college, are negative and at least marginally statistically significant.5

Models 3 and 4, Table 2 repeat the same analyses for change in non-Christian supernatural beliefs. The coefficients in Model 3 indicate that we can be at least moderately confident that the beliefs of students enrolled in elite institutions (p < .05), as well as those who are college graduates (p < .10), are more likely to decline compared to those who have never enrolled in any form of higher education. The inclusion of religious indicators from Wave 1 (Model 4) does little to alter the educational coefficients. Moreover, religious tradition is not a robust predictor of change in non-Christian super-empirical belief. Private prayer, however, increases belief in non-Christian super-empiricism, while attending church appears to be related to declining belief over time. Nevertheless, it is important to note that the overall R2 for non-Christian super-empiricism (.101) is considerably lower than the R2 for Christian super-empiricism (.213), suggesting that change in Christian super-empiricism is better predicted by educational and religious variables.6

Table 3 reports the predicted probabilities estimated with a dependent variable coded one if a respondent declined in belief between Wave 1 and Wave 3, and zero otherwise. All noneducational variables (not shown) are set to their mean. Model 1, Table 3 indicates that those who attend a religiously affiliated college have an adjusted probability of .253. In other words, holding constant all other influences, we can expect that approximately 25.3 percent of emerging adults who attend this type of institution will experience a decline in conventional Christian super-empirical belief. On the other end of the distribution, those that attend an elite university have an adjusted probability of .585, or 58.5 percent. All other educational possibilities fall between these two (with the probability for those who have never enrolled in any type of college being .335). Model 2 is equivalent to Model 1, except the outcome is now decline in non-Christian super-empiricism. The probabilities indicate that there is less variation by educational institution, with those attending baccalaureate colleges having a .110 adjusted probability of decline, and college graduates having a .287 probability of decline. The probability of decline from attending an elite institution (.198) is roughly similar to those attending associates, masters, and other doctoral granting institutions. When compared to Model 4 of Table 2, this suggests that the magnitude of decline is greater on average for those who attend elite institutions, while the absolute proportion of those experiencing religious decline at elite universities is about the same as at nonelite institutions.

Table 4 presents analyses for four different outcomes; all independent variables are included in each model. Model 1, Table 4 displays the results of an ordered logit regression predicting agreement with the label of “spiritual but not religious.”7 Coefficients are presented as odds ratios. The results suggest that there is no discernable effect of attending college on the likelihood of identifying as spiritual but not religious. Model 2, Table 4 presents the results of a logistic regression indicating belief that one does not need a religious congregation to be truly religious. The coefficients indicate that the effect of attending a religiously affiliated school is to encourage students to agree that congregations are an important part of being truly religious (compared to those who have never enrolled). Somewhat surprisingly, those who graduate from college with a bachelor's degree exhibit a similar effect, affirming the importance of congregations when compared to those who have never enrolled.

Table 4.  Logit and ordered logit lagged dependent variable regressions predicting beliefs about institutionalized religion and religious exclusivity (odds ratios), weighteda
 (1)(2)(3)(4)
Spiritual but not religiousbDo not need congregation to be religiouscOnly one religion is truecOkay to pick beliefs without accepting whole faithc
  1. Source: National Study of Youth and Religion 2002–2003, 2007–2008.

  2. aAll models include control variables for race/ethnicity, gender, age, relational closeness to parents, census region, and type of residence.

  3. bOrdered logit model.

  4. cLogit model.

  5. dReference category is never attended college.

  6. eReference category is evangelical Protestant.

  7. +p <  .10, *p <  .05, **p  <  .01, ***p  <  .001.

Currently enrolled
College/university typed
  Religiously affiliated.943.453**1.330.973
  Elite.744.998.8432.071+
  Associates1.000.786.636*1.425*
  Doctorate or masters university.953.810.8221.198
  Baccalaureate1.2831.228.5851.807+
Currently not enrolledd
  Some college1.063.757.863.882
  Bachelor's degree.959.410*.9192.091*
Lagged dependent variable1.874***3.312***4.909***1.777***
Additional W1 religious variables
  Attended religious school1.057.9241.556*1.166
  Religious traditione
   Mainline Protestant.9812.311***.599*1.640**
   Black Protestant1.273.852.6591.810*
   Catholic.9631.394+.379***1.723***
   Jewish1.371.950.6591.235
   LDS.441***.7681.766.886
   Not religious.8181.304.8201.711*
   Other religion1.1711.172.6031.443
   Indeterminate.8282.704*.6071.388
  Importance of faith1.021.736***1.413***.809**
  Frequency of prayer.9891.0151.125*.989
  Frequency of Bible reading.983.897*1.075.901*
  Frequency of church attendance.933**.871***1.087+.989
N2,4432,4552,4162,436
McFadden's pseudo R2.042.179.275.099

Model 3, Table 4 presents the results of a logistic regression on the belief that only one religion is true. While the coefficients indicate movement away from this view, they are not statistically significant, with the exception of those attending associates colleges. Model 4, Table 4 presents the results of a logistic regression on agreement that it is okay to pick and choose religious beliefs without accepting religious teachings as a whole. Here the coefficients indicate increased agreement for those who attend elite universities, baccalaureate institutions, and associates institutions. Those who graduate with a bachelor's degree are also more likely to agree that it is okay to pick and choose beliefs.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

There is no straightforward effect of college on religious beliefs. However, attention to what types of beliefs are most influenced by college, along with the influence of college type and college graduation, make it possible to piece together plausible interpretations of the impact of college on the structure and content of religious belief.

To begin, it should be noted that, for the vast majority of emerging adults, college only has a minor impact on religious beliefs. For example, if we use a simple dichotomous predictor of belief decline (see Table 3), we will find very little difference in the proportion who decline based on educational attainment and institutional type (with the clear exception of the few students attending elite universities). However, this does not mean that college has no impact on belief. The results do suggest that belief in super-empirical entities and occurrences are altered for some students as a result of attending and graduating from college. The effect is particularly pronounced for those attending elite universities. There is a detectable decline in both conventional Christian super-empirical beliefs as well as super-empirical beliefs related to non-Christian traditions (although Christian super-empiricism exhibits the greatest decline because of higher education). Students on America's college campuses are not abandoning Christian beliefs and adopting other religious beliefs. Rather, there appears to be a modest tendency to become skeptical of the existence of super-empirical entities and occurrences generally. Although we cannot directly link this skepticism to social processes occurring on college campuses, exposure to secular philosophies and the methods of the social and natural sciences, which de-emphasize the role of super-empirical explanation, is a likely influence. Knowledge rooted in rational discourse and empirical methods may simply make super-empirical belief less plausible for some students.

Alternatively, the social status associated with earning a bachelor's degree, and the social status achieved by attending an elite institution of education, may lead to identity work among emerging adults that involves distinguishing themselves from the religious views of those with less education. This explanation would help to explain why the effect from graduating from college was often significant, but the effect from attending some college was mostly nonsignificant. In other words, it may not simply be about exposure to certain philosophies and scientific knowledge, but rather it may be related to the identity work of differentiating oneself from the less educated. Eschewing super-empirical explanation may simply be one way that young people come to identify themselves as knowledge-class professionals. Further research is needed to explore this possibility.

In contrast, college students and graduates do not seem to be rejecting institutionalized religion as a result of their education. There is no consistent pattern showing any negative effect of college on identifying as “spiritual but not religious” or on thinking about the necessity of a religious congregation for religion. College graduates are more favorably disposed toward religious congregations than those who never attend college. This finding does not imply that there is no skepticism toward authoritative institutions among college students and graduates. It may be, as recent surveys have suggested (Pew Research Center 2010), that a growing distrust of institutions has become a part of our common culture, and students come to campus already fully embracing epistemological skepticism and cultural pluralism. For the current generation, then, college graduates may actually become slightly un-alienated from social institutions. Socialization processes in college may provide the knowledge and cultural tools for students to plug back into civic life (Knox, Lindsay, and Kolb 1993). In fact, most accounts of the positive correlation between educational attainment and church attendance note the civic benefits associated with church attendance among the educated (Brown and Taylor 2007; Johnson 1997; Petersen 1994; Uecker, Regnerus, and Vaaler 2007)

Still, the relative lack of institutional alienation does not translate into a lack of ecumenicism and inclusivity among college students and graduates. Attending college appears to be mildly associated with holding less exclusivist views of religious truth claims and a greater acceptance of idiosyncratic beliefs (see Table 4, Models 2 and 3). What might explain the seeming contradiction between the lack of evidence for institutional alienation coupled with the generalized affirmation of individualist beliefs? One possible solution lies in interpreting students’ attitudes toward the institutional church as primarily a marker of civic engagement and not as a declaration of submission to religious authority. College graduates are probably not giving assent to religious dogma simply because it is in line with church teaching. Rather, students and graduates are most likely participating in an anticipatory socializing process that prepares them for middle-class civic life—a life that often includes religious participation.

Institutional Identity

The institutional type of college or university is given special attention in this research, with the expectation that certain institutional identities may alter the processes that lead to change or stability in religious beliefs. I hypothesized that religiously affiliated institutions would purposefully neutralize many of the social mechanisms that might lead to a move away from church-sanctioned beliefs and that elite institutions might accentuate some of the mechanisms (particularly those related to classroom learning processes). Although the findings generally support these hypotheses, it is instructive to assess which beliefs were affected.

Very few statistically significant differences are found between those who attended a religiously affiliated college or university and those who never attended college of any type. Religious institutions seem to keep emerging adults on the belief trajectory they likely would have followed if they had not attended college at all. These institutions have somehow counteracted the social mechanisms that are operational on most college campuses and classrooms. It may be, as argued elsewhere (Hill 2009), that some religious campuses function as “moral communities” (Regnerus 2003; Stark 1996; Stark, Doyle, and Kent 1980) that legitimate religious practices and beliefs by preserving a shared moral order and collective religious identity. Legitimating religion publicly may counter the dominant narratives that question religious authority or find religion to be an irrational source of knowledge about the world. Ideally, these findings could be disaggregated across strength and type of religious affiliation. Unfortunately, there are too few respondents in the NSYR who are attending religious college to assess the effect of religious affiliation in greater depth.

Attending an elite educational institution has a different effect on religious belief than attending other postsecondary institutions. The most robust differences are evident in measures of super-empirical beliefs. Disbelief in the existence of supernatural entities and occurrences appears to be strongly linked with attending an elite university. This change is most plausibly related to exposure to secular ideas, faculty, and possibly to identity work associated with the elite social status of associating with these institutions. It is unlikely that a pluralistic environment (demographically or religiously) is responsible for the elite university effect, as most nonreligiously affiliated institutions display this type of pluralism. Further research is necessary to move beyond plausible ad hoc accounts to a confident understanding of the social processes that lead students to alter their religious beliefs.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

The current literature on religion and higher education only limitedly addresses the influence of higher education on the content of religious belief. The research in the article corrects for this limitation by systematically examining the influence of attending different institutional forms of higher education, as well as graduating with a bachelor's degree, on several measures of religious beliefs. The results indicate that some, but not all, religious beliefs are altered by higher education. Most notably, respondents become slightly more skeptical of the super-empirical if they attend and graduate from college. The effects are often exaggerated for college graduates and those attending elite universities, while those that attend religiously affiliated schools are typically no different from those that do not attend college of any type.

Future research would greatly benefit from an analysis of college student interviews in which respondents articulate their current and former beliefs about religion and how the college experience may have altered their belief trajectories. Interviewers should probe into the effect of classroom instruction, social interactions with faculty, formal extracurricular activities, and informal peer activities and interactions. Such research would greatly aid the search for more accurate and powerful social mechanisms. Methodologically, future research should be more precise in measurement of college identity. Unfortunately, because very small numbers of the overall emerging adult population attend private, religiously affiliated colleges, the present research was unable to more precisely identify the affiliation of the institution or measure how central the religious identity is in the academic and social programs of the institution. Future research would do well to oversample such institutions to understand how students’ religious trajectories are embedded within institutional contexts.

Footnotes
  • 1

    Oblique promax rotation is more appropriate than orthogonal varimax rotation because resultant factors may well be correlated with one another.

  • 2

    Additional models were run for each set of variables using a dependent variable constructed as a single principal components factor. A single component was extracted that accounted for 75 percent of the total variance consisting of the first group of six variables. For the second group of two variables, the single component extracted accounted for 77 percent of the total variance. Models performed similarly to a simple additive index of scores, but have the disadvantage of incommensurability with the factor scores of similar indexes extracted from the same measures at Wave 1. For this reason additive indexes are used for both Wave 1 and Wave 3 in the final models.

  • 3

    Dividing the dependent variable by a constant does nothing to alter the underlying statistical relationship between the independent and dependent variables. Rather, altering the metric of the dependent variables allows us to roughly gauge effect size across models with dependent variables originating in different metrics. For example, this method allows us to compare the effect of attending an elite university on both Christian and non-Christian change in super-empiricism, despite the original metric of both of these indexes being different. Moreover, the original metric of these indexes is unlikely to be inherently meaningful for most readers.

  • 4

    In a typical “agree/disagree” dichotomous variable, measured at two points in time, there are four meaningful categories for a change score: “disagree to agree,”“agree to disagree,”“agree to agree,” and “disagree to disagree.” The latter two stable categories are problematic for a theory positing underlying directional change. At the very least, such an analysis would require twice the number of coefficients as a lagged dependent variable model (both the “no change” categories are meaningful base categories in a multinomial logistic regression). After running these more complex models, I concluded that the final interpretations were the same as the lagged dependent variable models.

  • 5

    Separate regression analyses (not shown) for each composite measure indicate similar patterns, but with differing coefficient sizes dependent on the measure. The effects attributed to higher education, particularly for students enrolled in elite universities and those who have earned a BA or more, are strongly negative for belief in angels. Belief in God, demons, miracles, and a future judgment day are similarly negative, but less strong. Belief in the afterlife shows similar patterns but is not significant.

  • 6

    Once again, separate regression analyses were conducted for each dependent variable (not shown). The effects are similar to the overall composite measure, with a stronger negative effect from attending elite universities on belief in astrology. The effect from earning a bachelor's degree is roughly equivalent for both outcomes.

  • 7

    A Brant test confirms that ordered logit is the appropriate statistical method over a multinomial logit model.

Acknowledgments: 

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

The author wishes to express his thanks to Calvin College for the fellowship that provided the time to work on the research presented in this article. Additional thanks go to the editor, Marie Cornwall, and the anonymous reviewers for their helpful comments and sound advice on earlier drafts. The National Study of Youth and Religion, http://www.youthandreligion.org, whose data were used by permission here, was generously funded by Lilly Endowment Inc., under the direction of Christian Smith, of the Department of Sociology at the University of Notre Dame and Lisa Pearce, of the Department of Sociology at the University of North Carolina at Chapel Hill.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information
  • Arum, Richard and Josipa Roksa. 2011. Academically adrift: Limited learning on college campuses. Chicago , IL : University of Chicago Press.
  • Astin, Alexander W., Helen S. Astin, and Jennifer A. Lindholm. 2010. Cultivating the spirit: How college can enhance students’ inner life. San Francisco , CA : Jossey-Bass.
  • Baker, Joseph O. 2008. Who believes in religious evil? An investigation of sociological patterns of belief in Satan, hell, and demons. Review of Religious Research 50(2):20620.
  • Baker, Joseph O. and Scott Draper. 2010. Diverse supernatural portfolios: Certitude, exclusivity, and the curvilinear relationship between religiosity and paranormal beliefs. Journal for the Scientific Study of Religion 49(3):41324.
  • Becker, Lee B. 1977. Predictors of change in religious beliefs and behaviors during college. Sociological Analysis 38(1):6574.
  • Bellah, Robert N., Richard Madsen, William M. Sullivan, Ann Swidler, and Steven M. Tipton. 1985. Habits of the heart: Individualism and commitment in American life. Berkeley , CA : University of California Press.
  • Berger, Peter L. 1967. The sacred canopy. New York : Doubleday.
  • Berger, Peter and Anton Zijderveld. 2009. In praise of doubt: How to have convictions without becoming a fanatic. New York : HarperOne.
  • Brown, Sarah, and Karl Taylor. 2007. Religion and education: Evidence from the National Child Development Study. Journal of Economic Behavior and Organization 63(3):43960.
  • Bryant, Alyssa N. and Helen S. Astin. 2008. The correlates of spiritual struggle during the college years. Journal of Higher Education 79(1):127.
  • Bryant, Alyssa N., Jeung Yun Choi, and Maiko Yasuno. 2003. Understanding the religious and spiritual dimensions of students’ lives in the first year of college. Journal of College Student Development 44(6):72345.
  • Budziszewski, J. 1999. How to stay Christian in college: An interactive guide to keeping the faith. Colorado Springs , CO : NavPress.
  • Burtchaell, James T. 1998. The dying of the light: The disengagement of colleges and universities from their Christian churches. Grand Rapids , MI : Eerdmans.
  • Caplovitz, David and Fred Sherrow. 1977. The religious drop-outs: Apostasy among college graduates. Beverly Hills, CA : Sage Publications.
  • Cherry, Conrad, Betty A. DeBerg, and Amanda Porterfield. 2001. Religion on campus. Chapel Hill , NC : University of North Carolina Press.
  • Clydesdale, Timothy T. 2007a. Abandoned, pursued, or safely stowed? Web forum on the religious engagements of American undergraduates. Retrieved June 7, 2010. Available at http://religion.ssrc.org/reforum/Clydesdale.pdf
  • Clydesdale, Timothy T. 2007b. The first year out: Understanding American teens after high school. Chicago , IL : University of Chicago Press.
  • Dillon, Michelle and Paul Wink. 2007. In the course of a lifetime: Tracing religious belief, practice, and change. Berkeley , CA : University of California Press.
  • Ecklund, Elaine Howard. 2010. Science vs religion: What scientists really think. New York : Oxford University Press.
  • Feldman, Kenneth A. and Theodore M. Newcomb. 1969. The impact of college on students. San Francisco , CA : Jossey-Bass.
  • Finkel, Steven E. 1995. Causal analysis with panel data. Thousand Oaks , CA : Sage.
  • Freitas, Donna. 2008. Sex and the soul: Juggling sexuality, spirituality, romance, and religion on America's college campuses. New York : Oxford University Press.
  • Gallup, George, Jr. and D. Michael Lindsay. 1999. Surveying the religious landscape: Trends in U.S. beliefs. Harrisburg , PA : Morehouse Publishing.
  • Gross, Neil and Solon Simmons. 2007. How religious are America's college and university professors? SSRC web forum on the religious engagements of American undergraduates. Retrieved June 7, 2010. Available at http://religion.ssrc.org/reforum/Gross_Simmons.pdf
  • Gross, Neil and Solon Simmons. 2009. The religiosity of American college and university professors. Sociology of Religion 70(2):10129.
  • Hadaway, C. Kirk and Wade Clark Roof. 1988. Apostasy in American churches: Evidence from national survey data. In Falling from the faith: Causes and consequences of religious apostasy, edited by David G. Bromley, pp. 2946. Newbury Park , CA : Sage.
  • Hammond, Phillip E. and James Davison Hunter. 1984. On maintaining plausibility: The worldview of evangelical college students. Journal for the Scientific Study of Religion 23(3):22138.
  • Higher Education Research Institute. 2006. Spirituality and the professoriate: A national study of faculty beliefs, attitudes, and behaviors. Los Angeles , CA : Higher Education Research Institute at University of California, Los Angeles.
  • Hill, Jonathan P. 2008. Religious pathways during the transition to adulthood: A life course approach. Ph.D. thesis, University of Notre Dame.
  • Hill, Jonathan P. 2009. Higher education as moral community: Institutional influences on religious participation during college. Journal for the Scientific Study of Religion 48(3):51534.
  • Hunsberger, Bruce. 1978. The religiosity of college students: Stability and change over years at university. Journal for the Scientific Study of Religion 17(2):15964.
  • Hunter, James Davison. 1987. Evangelicalism: The coming generation. Chicago , IL : University of Chicago Press.
  • Hurtado, Sylvia, Linda J. Sax, Victor Saenz, Casandra E. Harper, Leticia Oseguera, Jennifer Curley, Lina Lopez, De'Sha Wolf, and Lucy Arellano. 2007. Findings from the 2005 administration of Your First College Year (YFCY): National aggregates. Los Angeles , CA : Higher Education Research Institute.
  • Johnson, Daniel Carson. 1997. Formal education vs. religious belief: Soliciting new evidence with multinomial logit modeling. Journal for the Scientific Study of Religion 36(2):23146.
  • Knox, William E., Paul Lindsay, and Mary N. Kolb. 1993. Higher education, college characteristics, and student experiences: Long-term effects on educational satisfactions and perceptions. Journal of Higher Education 63(3):30328.
  • Kuh, George D. 2000. Do environments matter? A comparative analysis of the impress of different types of colleges and universities on character. Journal of College and Character. Published online. Available at http://journals.naspa.org/cgi/viewcontent.cgi?article=1277&context=jcc, last accessed April 22, 2011.
  • Kuh, George D. and Paul D. Umbach. 2004. College and character: Insights from the National Survey of Student Engagement. In Assessing character outcomes in college: New directions in institutional research, edited by Jon C.Dalton, Terrance R.Russell, and SallyKline, pp. 3754. San Francisco , CA : Jossey-Bass.
  • Lee, Jenny J. 2002. Religion and college attendance: Change among students. Review of Higher Education 25(4):36984.
  • Madsen, Gary E. and Glenn M. Vernon. 1983. Maintaining the faith during college: A study of campus religious group participation. Review of Religious Research 25(2):12741.
  • Marsden, George M. 1994. The soul of the American university: From Protestant establishment to established nonbelief. New York : Oxford University Press.
  • Mayrl, Damon and Freeden Oeur. 2009. Religion and higher education: Current knowledge and directions for future research. Journal for the Scientific Study of Religion 48(2):26075.
  • McFarland, Michael J., Bradley R. E. Wright, and David L. Weakliem. 2010. Educational attainment and religiosity: Exploring variations by religious tradition. Sociology of Religion. Published online August 31, 2010, doi: 10.1093/socrel/srq065.
  • Menard, Scott W. 2010. Logistic regression: From introductory to advanced concepts and applications. Thousand Oaks , CA : Sage.
  • Nye, Abby. 2005. Fish out of water: How to survive as a Christian on a secular campus. Green Forest , AZ : New Leaf Press.
  • Orenstein, Alan. 2002. Religion and paranormal belief. Journal for the Scientific Study of Religion 41(2):30111.
  • Pascarella, Ernest T. and Patrick T. Terenzini. 2005. How college affects students: A third decade of research. San Francisco , CA : Jossey-Bass.
  • Petersen, Larry R. 1994. Education, homogamy, and religious commitment. Journal for the Scientific Study of Religion 33(2):12235.
  • Pew Research Center. 2010. Distrust, discontent, anger and partisan rancor: The people and their government. Washington , DC : Pew Resarch Center. Retrieved June 7, 2010. Available at http://people-press.org/report/606/trust-in-government.
  • Plewis, Ian. 1985. Analysing change: Measurement and explanation using longitudinal data. New York : J. Wiley.
  • Putnam, Robert D. and David E. Campbell. 2010. American grace: How religion divides and unites us. New York : Simon & Schuster.
  • Regnerus, Mark D. 2003. Moral communities and adolescent delinquency: Religious contexts and community social control. Sociological Quarterly 44(4):52354.
  • Reimer, Sam. 2010. Higher education and theological liberalism: Revisiting the old issue. Sociology of Religion 71(4):393408.
  • Rice, Tom W. 2003. Believe it or not: Religious and other paranormal beliefs in the United States. Journal for the Scientific Study of Religion 42(1):95106.
  • Robertson, Roland. 1970. The sociological interpretation of religion. New York : Schocken Books.
  • Saenz, Victor B. and Douglas S. Barrera. 2007. Findings from the 2005 College Student Survey (CSS): National aggregates. Los Angeles , CA : Higher Education Research Institute.
  • Scheitle, Christopher P. 2011. U.S. college students’ perception of religion and science: Conflict, collaboration, or independence? A research note. Journal for the Scientific Study of Religion 50(1):17586.
  • Sherkat, Darren E. 2008. Beyond belief: Atheism, agnosticism, and theistic certainty in the United States. Sociological Spectrum 28(5):43859.
  • Smith, Christian. 2003a. Secularizing American higher education: The case of early American sociology. In The secular revolution: Power, interests, and conflict in the secularization of American public life, edited by ChristianSmith, pp. 97159. Berkeley , CA : University of California Press.
  • Smith, Christian. 2003b. Moral, believing animals: Human personhood and culture. New York : Oxford University Press.
  • Smith, Christian and Melinda L. Denton. 2003. Methodological design and procedures for the National Survey of Youth and Religion (NSYR). Chapel Hill , NC : National Study of Youth and Religion.
  • Smith, Christian and Melinda L. Denton. 2005. Soul searching: The religious and spiritual lives of American teenagers. New York : Oxford University Press.
  • Smith, Christian and Patricia Snell. 2009. Souls in transition: The religious & spiritual lives of emerging adults. New York : Oxford University Press.
  • Stark, Rodney. 1996. Religion as context: Hellfire and delinquency one more time. Sociology of Religion 57(2):16373.
  • Stark, Rodney. 2008. What Americans really believe. Waco , TX : Baylor University Press.
  • Stark, Rodney, Daniel P. Doyle, and Lori Kent. 1980. Rediscovering moral communities: Church membership and crime. In Understanding crime: Current theory and research, edited by TravisHirachi and MichaelGottfredson, pp. 4352. Beverly Hills , CA : Sage.
  • Stark, Rodney and Roger Finke. 2000. Acts of faith: Explaining the human side of religion. Berkeley , CA : University of California Press.
  • Steensland, Brian, Jerry Z. Park, Mark D. Regnerus, Lynn D. Robinson, W. Bradford Wilcox, and Robert D. Woodberry. 2000. The measure of American religion: Toward improving the state of the art. Social Forces 79(1):291318.
  • Uecker, Jeremy E., Mark D. Regnerus, and Margaret L. Vaaler. 2007. Losing my religion: The social sources of religious decline in early adulthood. Social Forces 85(4):166792.
  • Weeks, Matthew, Kelly P. Weeks, and Mary R. Daniel. 2008. The implicit relationship between religion and paranormal constructs. Journal for the Scientific Study of Religion 47(4):599611.
  • Wheaton, David. 2005. University of destruction: Your game plan for spiritual victory on campus. Grand Rapids , MI : Bethany House.
  • Wolfe, Alan. 2006. The evangelical mind revisited. Change: The Magazine of Higher Learning 32(2):913.
  • Woolever, Cynthia and Deborah Bruce. 2010. A field guide to U.S. congregations: Who's going where and why, second edition. Louisville , KY : Westminster John Knox Press.
  • Wuthnow, Robert. 1985. Science and the sacred. In The sacred in a secular age, edited by Phillip E.Hammond, pp. 197203. Berkeley , CA : University of California Press.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Religious Beliefand HigherEducation: Previous Research
  5. Hypothesizing Religious Changein College
  6. Methods
  7. Results
  8. Discussion
  9. Conclusion
  10. Acknowledgments: 
  11. References
  12. Supporting Information

Appendix S1. Religious Belief Variables

FilenameFormatSizeDescription
JSSR_1587_sm_AppendixS1.pdf38KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.