Education and Religion: The Second Demographic Transition in Utah, - - PDF document

Education and Religion: The Second Demographic Transition in Utah, 1850-2000 Ashley Larsen Gibby, Pennsylvania State University*, agl132@psu.edu Jane Lankes, Pennsylvania State University*, jlankes@psu.edu David Baker, Ph.D., Pennsylvania State University * indicates equal contribution

Abstract: Education and religion are often conceptualized as competing forces in the second

demographic transition (SDT), with education driving the trends and religiosity slowing them. However, this assumption has not been effectively tested. Utah, with high levels of education and religiosity, offers a useful case to examine this assumption. Using census data across 150 years, we find that the rates of fertility and marital status for Utah parallel national trends, but are consistently higher for fertility and marriage and lower for divorce and proportion never married. Therefore, the repeated explanation that Utah is merely lagged in the SDT, meaning, that while it has not participated in the SDT yet, it will in the future, is incomplete. Rather, our findings suggest that Utah has been an active participate in the SDT. Further, neither religiosity nor education dominated demographic trends in Utah. Instead, our findings support the conclusion that religion and education are not necessarily in conflict but, rather, may work together to provide values, frameworks, and resources that contribute to family processes. DRAFT PAPER – DO NOT CITE OR CIRCUALTE WITHOUT THE AUTHORS’ PERMISSION Acknowledgements: The authors acknowledge support from the Population Research Institute at Penn State, which receives funding from the National Center of Child Health and Human Development (R24-DH041025) as well as the NICHD Pre-Doctoral Training Grant in Family Demography (T-32HD007514).

1

INTRODUCTION Both education and religion are highly cited influences on the second demographic transition (SDT); however, empirical tests of their influence remain minimal. Highly religious populations generally exhibit lower rates of participation in the second demographic transition (SDT) (Surkyn and Lesthaeghe 2004; Bystrov 2012). Therefore, religion has often been cited as a deterrent to the SDT. Conversely, education is regarded as a catalyst in demographic transitions, increasing economic resources (Lesthaeghe and van de Kaa 1986; Lesthaeghe and Lopez-Gay 2013; Lesthaeghe and Neidert 2006) and serving as a source of ideation change (Hernandez et al. forthcoming). As such, religion and education are often pitted against each other as competing forces in the SDT, with religion slowing the trends and education accelerating them. Furthermore, researchers have often explained exceptions to the SDT trends as “lags,” meaning, while these populations have not undergone the SDT yet, they will follow a similar trajectory at some time in the future. Religion, or “traditional ideation” is often cited as an explanation for this perceived lag in participation (Kirk 1996; Lesthaeghe 1983, Carlson 2005; Surkyn and Lesthaeghe 2004). Utah is one of these exceptions. It appears that Utah has yet to participate in the SDT according to cross-sectional data that demonstrate Utah's high fertility and marriage rates as well as low divorce and cohabitation rates (Surkyn and Lesthaeghe 2004). While Utah has been identified as a lagging state, no formal test of this lag or historical examination of Utah's progress in the SDT has been conducted. Using census data from 1850-2000, we examine trends in the second demographic transition in Utah longitudinally. Utah has high rates of both religiosity and education, making it a particularly apt case to examine the influence of both education and religion on the SDT. Further, Utah is highly comparable to the United States at large in terms of wealth disparities, health differences, etc., eliminating competing explanations and sources of endogeneity. We find that Utah has not been lagging in the SDT but, instead, followed the trends of the United States at large at a higher or lower mean (depending on the indicator). We also postulate that religion does not necessarily slow the SDT while education propels it. Rather, religion and education can act synergistically to help individuals negotiate choices regarding family formation. These findings, as well as this discussion, are important to consider. While previous studies have examined the relationships between religion and education with the second demographic transition, these studies are largely cross-sectional. To our knowledge, none have considered religion and education in conjunction longitudinally. By ignoring or oversimplifying religion’s influence, or by explaining away exceptions to the SDT without testing these explanations, we may miss important nuance and incorrectly estimate future trends. Further, these explanations have implications beyond those of Utah, and can be applicable internationally to populations, especially highly religious areas, who are currently not participating in the SDT. BACKGROUND AND CONTEXT The second demographic transition (SDT) includes increased cohabitation and divorce along with decreased fertility, marriage, non-marital childbearing and the postponement of marriage and childbearing (Lesthaeghe and van de Kaa 1986). Evidence of this transition is found in all but a small group of countries in the world and every state in the United States (Lesthaeghe and Neidert 2006b). Essentially, the SDT involves decisions regarding family formation and

2

family structure that are deeply personal and moral. Therefore, it is no surprise that these types of decisions are influenced heavily by an individual’s religion and educational attainment. An in-depth body of literature exists on the significant relationships between religion and demographic processes. In general, religion is associated with more traditional family forms such as higher fertility (Frejka and Westoff 2008), higher likelihood of marriage (Thornton, Axinn, and Hill 1992), and lower rates of divorce (Larson and Golz 1989). Lower religiosity, on the other hand, predicts non-traditional union formation (Thornton, Axinn, and Hill 1992) and lower fertility (Surkyn and Lesthaeghe 2004). While not all religious groups are associated with these trends, religion as a whole is painted as a deterrent to the SDT. Lesthaeghe and Neidert (2006a) find that percent Evangelical/Mormon is associated lower SDT score, and reiterate Lesthaeghe and van de Kaa's 1986 SDT paper stating that new living arrangements, cohabitation in particular, are a reflection of secular, anti-authoritarian, and better educated new birth cohorts. This is not surprising, as conservative religion discourages several behaviors associated with the SDT such as divorce, postponing marriage and fertility, and non-marital childbearing. Further, Bystrov (2012) found that more highly religious areas in Israel were less likely to participate in the SDT with certain aspects of the SDT such as childlessness, non-marital cohabitation, and non-marital childbearing, almost nonexistent. Religion is often cited as a primary distinguishing factor between American and European SDT progression. It is the United States’ high religious adherence, Carlson argues (2005), which is a partial explanation for the United States’ fertility levels never falling to below replacement. To explain these types of community differences, Lesthaeghe attributes differences in fertility levels and speeds of change to a shift from community needs toward individual wants, calling it an “emancipation process” from religious control (1983). He posits that demographic change is a result of ideation shift intertwined with religious institutions. Therefore, societies are progressing in the SDT due to the distance they are gaining from religious control. But religion is still active in certain populations. Religious institutions shape attitudes about both fertility (McQuillan 2004) and family formation (Ellison, Wolfinger, and Ramos-Wada 2012), creating demographic divides between the religious and the secular. In general, however, this literature on religion and demographic transitions tends to be cross-sectional, limiting our scope of the second demographic

• transition. While some studies have examined trends longitudinally (Kane 2013; Glass 2014;

Lesthaeghe 2010), by examining this over a much longer period, we hope to gain a clearer picture

• f how these trends have evolved.

While religion has been viewed as a deterrent, education has been cited as a catalyst or driver of the second demographic transition worldwide (Hernandez et al. forthcoming). The highly educated have historically been on the vanguard of demographic behaviors associated with the second demographic transition. For example, highly educated women were the first to experience vast increases in divorce (Norton and Moorman 1987). Most literature considering education's impact on the SDT considers education's economic advantages. First, education increases economic resources, providing individuals with more power to negotiate their preferred family form (Lesthaeghe and van de Kaa 1986; Lesthaeghe and Lopez-Gay 2013; Lesthaeghe and Neidert 2006). Second, education improves the labor market for women, providing them more bargaining power and more opportunity costs (Lesthaeghe and Lopez-Gay 2013; Lesthaeghe and Neidert 2006). However, a new line of research argues that education, like religion, provides more than just economic advantage. Instead, education provides a framework from which to view the world, values, and ideologies. Thus, education is similar to religion in its impact not only on resources, but on human thoughts and values (Baker forthcoming). Therefore, education’s large association

3

with the second demographic transition in the United States could be due to its impact on individuals’ priorities while forming and navigating families (Hernandez et al. forthcoming). The Education vs. Religion Assumption Religion and education have long been considered in conflict. Many scholars incorrectly predicted that the death of religion would be at the hands of the scientific and education revolution (e.g., Inkles and Smith 1974; Douglas 1982). These predictions emerged from the assumption that the values and ideations taught through education were different from those exuding from conservative religion. For example, while education promotes rationality and viewing society as a human construct, religion promotes “faith, revelation, and the existence of a supernatural being” (Baker forthcoming). However, religion did not vanish during the education revolution. Instead, it endured and, in certain areas, even thrived during the impressive spread of education (for a review: Gorski and Altinordu 2008). In some places, education seemed to destroy religion. For example, in Iceland only two percent of the population attends weekly religious services. However, upon further examination, only two percent of adults in Iceland reported they were "convinced atheists" with 80 percent or more believing in some supernatural being, life after death, and the human soul. Many even reported to praying to some supernatural being (Stark and Finke 2000). It seems rational to assume that the spread of education would undermine religion. Yet, they seem to coexist. Several explanations have been proposed for this oddity. First, religious beliefs differ on the base of education. Religious individuals with at least a Bachelors' degree are twice as likely to believe in a "distant,” abstract, forgiving God. Further, more educated Christians are less likely to believe the bible is literal (Schwadel 2011; McFarland et al. 2011). These results can be interpreted as a shift in religious ideology to lessen cognitive dissonance – an attempt, within an individual, to hold complementary ideologies, so that their religious beliefs and understanding of science are not in conflict. Another explanation is that those who hold such beliefs are more likely to seek education. Due to limited longitudinal studies, the directionality of this relationship is unclear (Baker forthcoming). The existing longitudinal studies seem to suggest that education does not lead to a rejection of religious-thought but rather strengthens an individual's spirituality (Graham and Donaldson 1996; Lee 2002; Mayrl and Uecker 2011; Pascarella and Terenzini 2005). Second, the institutions of religion and education may promote similar ideologies and pursue similar goals, thus leading to their cooperation. This is especially pertinent to the United States which experienced a protestant revivalism. The revivalism discouraged beliefs that human nature is fallen or that salvation was out of an individual's control and supported beliefs that individuals have free choice, all are created equal, and that we could understand God’s world, (Thomas 1989). Education was (and is) clearly compatible with these goals. This same logic applies to the SDT. While religion has goals of family formation, stability, and reproduction through marriage, it also has goals of human agency and worth. Thus, while religion and education may hold different ideals and goals, their means and logic are not as vastly different as they may be portrayed. Thus, it is possible that claims of the competing impacts of education and religion

• n the SDT could be over-estimated.

4

The Unique Case of Utah An area with high rates of both education and religion was needed in order to fully examine the relative influence of religion and education on the SDT. Utah was ideally suited for this goal. Utah has among the highest education rates in the United States (Hernandez et al. Forthcoming) and is one of the most highly religious states. Utah contains one of the only high concentrations of a high-commitment religion in the United States, the Church of Jesus Christ of Latter-day Saints,

• r, the LDS Church. While other conservative religions exhibit similar influences on demographic

trends, none are as highly concentrated as the LDS Church in Utah, which encourages marriage and having children as well as obtaining high levels of education (Pew Research Center 2012). Others have examined areas of religious concentration in the Middle East (Bystrov 2012). However, Utah provides one of few conservative, religious populations consolidated into one area in the United States. In particular, Utah is a particularly useful case because of its unique history. The Church

• f Jesus Christ of Latter-day Saints (LDS) was founded in the mid-1800s in New York. Due to

persecution, church members were forced to migrate across the United States. They subsequently settled the state of Utah and were somewhat isolated from the rest of the country for several

• decades. Still today, about 70 percent of Utah is LDS and the church headquarters resides in the

state capital. The LDS Church has clear, strict rules regarding family formation and processes. Its core doctrine prioritizes family as not only important, but the very fabric of heaven (Pew Research Center 2012). Thus, the church encourages marriage and childbearing highly and heavily discourages divorce (except in specific cases such as domestic abuse) and pre-marital sex (ibid). Somewhat unique to the LDS Church is its emphasis on education. LDS members are asked to gain as much education as possible in their situation and to continually learn from various sources (Pew Research Center 2012). Thus, the LDS population has high levels of education but is motivated to not fully participate in the SDT. Past literature has noticed Utah’s peculiarity in its high rates of education as well as fertility and marriage (Lesthaeghe and Neibert 2006a, Hernandez et al. forthcoming). Utah has been conceptualized as lagged in the SDT without extensive empirical testing. Lesthaeghe and Neibert categorize Utah as “resisting the SDT” and note that it is conservative, with low numbers of teenage mothers and non-marital childbearing (2006b). Yet, they only examine this at one point in time, unable to determine how these trends have progressed over time. Therefore, this study seeks to examine if Utah was and is indeed lagged in the second demographic transition across a variety of measures by examining this assumption longitudinally. Here, we take a step back from “reading history sideways” (Thornton 2005), and observe the second demographic transition over 150 years. We observe this progression across several states: one with high rates of both education and religion (Utah), two with high rates of education but low levels

• f religion (California and Colorado), and one with high rates of religion but low levels of

education (Idaho) in order to estimate the relative influence of education and religion on how states experience the SDT.

5

RESEARCH QUESTIONS We sought to answer the following questions: (1) What happens in a context of high religiosity and high education? Which aspect more strongly influences demographic behaviors? And, more broadly, (2) What contributes to, or hinders, the second demographic transition? HYPOTHESES Figure 1 shows a hypothetical example of what we may find when observing the historical

• data. Of course, this represents an ideal type of our hypotheses, as it is much more simple and

straightforward than real-world data are. The y-axis represents participation in the second demographic transition. Increasing divorce and cohabitation, and decreasing fertility and marriage rates would indicate higher rates of participation in the second demographic transition, for

• example. The x-axis represents time. The light green bar represents the United States (again,

represented in a simplified fashion). Through time, the United States has increasingly participated in the second demographic transition. For example, rates of non-marital childbearing have been increasing since the 1950s (Ventura 2009). If education is driving Utah’s trends, its trajectory would look like this line (H1). If Utah were lagged in its participation in the SDT, we would observe something similar to the light blue line. First, there would be a period of non-participation, followed by a period of accelerated participation. This is what past research has postulated as an explanation for Utah’s higher rates of fertility and marriage and lower rates of divorce (e.g, Lesthaeghe and Neibert 2006a). Utah’s high rates of religiosity are cited as a possible explanation for this (ibid). Therefore, this represents our second hypothesis (H2). However, it may be that Utah will not participate at

• all. If this were the case, Utah’s trends would appear like the dark green bar with minimal change

in their demographic patterns. This represents our third hypothesis (H3). Figure 1: Conceptual Model DATA AND METHODS We utilized the decennial census data from 1850-2000. For each year, we used the largest sample available through IPUMS USA. This was the 1% sample (a random sample of 1 in every

6

100 census respondents) for 1850, 1860, 1870, 1910, 1920, 1940, 1950, and 1970; the 5% sample for 1900, 1930, 1960, 1980, 1990, and 2000; and the 10% sample for 1880. Historical trends were estimated for Utah as the focal state for all available years and California, Colorado, and Idaho for comparison states. Colorado and California represent highly educated states that exhibit low levels

• f religiosity (see Figure 2). Therefore, they serve as our education comparison. If education is the

driving force for Utah’s progression in the second demographic transition, its trajectory should look similar to these states. Idaho was chosen as the opposing comparison. Idaho has low levels

• f education and high levels of religiosity, and, high levels of Mormon adherents (see Figure 4).

From 1900 on, Idaho had relatively high rates of Mormon adherents while California and Colorado had rates close to zero. Therefore, if religion is deterring Utah’s SDT, its trajectory should look like Idaho’s. To observe these trajectories, we first utilized the individual-level data and estimated individuals’ participation in the second demographic transition (i.e., whether they were divorced, their age at first birth, etc.). We then used this information to create rates of multiple demographic behaviors by state. Then, we collapsed this information into state-years, the unit of our analysis. Figure 2: Religiosity (% of adults who are “highly religious”) and Educational Attainment by State, 2014

Sources: U.S. Religious Landscape Study, 2014; 2011-2015 American Community Survey 5-Year Estimates

7

Figure 3: LDS Members per 1000 Population by State, 1890-2000

Sources: authors’ calculations; data from the Department of the Interior, U.S. Census (1890), Department

• f Commerce and Labor, U.S. Census (1906-1936), National Council of Churches (1952), and Association of

Statisticians of American Religious Bodies (1971-2000); downloaded from the Association of Religion Data Archives, www.TheARDA.com Note: Data collection switched from “members” to “adherents” in 1971. Members in the LDS church are anyone aged 8+. Adherents include members and their children. This presents a small change in measurement over

• time. Therefore, 1971-2000 numbers are inflated.

Measures We measured several aspects of the second demographic transition relating to fertility, marriage, and education. We used Lesthaeghe and Neidert (2006a) as a guide for these measurements.

• Fertility. We measured decreased fertility rates by estimating the average number of children born

to women over the age of 18 for each state, each year (number of children ever born to women

• ver 18/total number of women over the age of 18 * 100).
• Marriage. To measure the rates of marriage through time, we estimated the number of women

who were married from all women 18 years or older. Further, we estimated the rate of women over 18 who respond that they have never been married. This was done for all available years.

• Divorce. We estimated the divorce rate for women over the age of 25 (women over 25 who

indicated that they were currently divorced/all women who were 25 or older * 100), for each state in all available years.

• Education. Levels of education were measured by the proportion of women aged 18-25 currently

attending college as well as the proportion of the adult population (18 and over) population, both men and women, who completed college by the age of 25. This was done for all available years.

8

RESULTS Figure 4: Proportion of 18-25 year olds attending college

• Education. Through time, Utah has had consistently higher rates of enrolled young adults than all
• ther states in the analyses. Further, the education revolution is clearly observed for all states,

regardless of religiosity rates. Therefore showing that education did not slow the education revolution and, in Utah’s case, may have supported it (Baker 2012, Baker forthcoming). Figure 5: Average Children ever born for Women over 18

• Fertility. Utah’s fertility rate is consistently higher from those of Colorado and California.

However, the general trends are synchronized. Utah experienced the same drastic decrease and slight increase, at the same time, as both Colorado and California; however, these trends occurred

20 40 60 80 1900 1920 1940 1960 1980 2000 Year Utah Colorado California Idaho

Proportion of all 18-25 year olds attending college by year

1 2 3 4 5 1900 1920 1940 1960 1980 2000 Year Utah Colorado California Idaho

Average children ever born by year for Women over 18

9

at a higher fertility rate that remained higher throughout the entire time period. In fertility trends, Utah appears more similar to Idaho, the state with high religion and low education. Figure 6: Divorce rates for women over 25

• Divorce. Once again, it appears that Utah experiences the same trends of increasing and then

stabilized divorce rates as California and Colorado (see Figure 6). Utah consistently experiences a lower incidence of divorce while following the same timing and transitions of Colorado and

• California. Idaho experiences these trends in unison with other states as well; however, generally

experiencing lower divorce rates. Figure 7: Marriage rates for women over 18 Marriage (see Figure 7). Here again, the same trends occurred at the same time for all four states. Idaho has the highest rates of marriage for women over the age of 18, consistently, with Utah close

• behind. In the early 1900s, Utah appears much more similar to Colorado and California, our high

5 10 15 1850 1900 1950 2000 year_1 Utah Colorado California Idaho

Divorce rates by year for Women over 25

55 60 65 70 75 80 1900 1920 1940 1960 1980 2000 Year Utah Colorado California Idaho

Marriage rates by year for Women over 18

10

education states. However, with time, Utah becomes more similar to Idaho, the state with low education rates and high religiosity. Figure 8: Rates of never married for women over 25 The incidence of “never marriage” (see Figure 8) decreased with time up until 1950 when women

• ver 18 who had never married stabilized at about ten percent. After the 1950s, the incidence

increased again and appears to still be increasing. California is the clear leader in this trend. Utah appears more similar to highly educated states until the 1950s when it once again joins Idaho with low rates of never married adult women. We did not find support for any of our proposed hypotheses. Utah did not participate in the second demographic transition to the same degree as Colorado or California (H1 represented by the light green line in Figure 9). However, Utah did participate in the second demographic transition (rejecting H3 and the dark green line in Figure 9), and did so in unison with California and Colorado (rejecting H2 and the light blue line in Figure 9). Instead, we find what we represent by the purple line in Figure 9. Utah experienced the second demographic transition with similar timing as other states in the United States. Therefore, explaining its higher rates of marriage and fertility and lower rates of divorce and cohabitation as a lag seems incomplete. However, Utah did not experience the same dramatic rates of these demographic behaviors. Often, Utah appeared more similar to Idaho, the religious comparison, meaning that religion likely had an impact. Utah’s marriage and fertility rates were consistently higher, and Utah’s divorce and never married rates are consistently lower than those of California and Colorado.

10 20 30 40 1900 1920 1940 1960 1980 2000 Year Utah Colorado California Idaho

Rates of Never Married by year for Women over 25

11

Figure 9: Conceptual Model DISCUSSION This study, to our knowledge, is the first examination of the influence of religion and education on the second demographic transition, longitudinally, rather than cross-sectionally. We employed census data spanning 150 years to determine how areas with differing degrees of education and religiosity experienced the SDT. We focused on Utah as a place with high levels of conservative religiosity and education and found that neither education nor religion dominated Utah's experience with the SDT. Utah experienced increased divorce and postponements in marriage as well as increased fertility and divorce in parallel with other highly educated states. However, Utah's rates of fertility and marriage always stayed considerably higher and divorce and postponement of marriage stayed considerably lower than other states with high levels of education and low levels of religiosity. This type of examination provides several contributions to the current literature. First, a historical analysis illustrated the timing of these transitions across several states. Therefore, we were able to conclude that describing religious Utah as “lagged” was an incomplete assessment (e.g., Kirk 1996; Lesthaeghe 1983). Second, this analysis allowed us to more clearly assess the influence of religion and education by observing the differences in these transitions across different

• states. And, third, this analysis began an exploration into the "exceptions" of the SDT, allowing us

to refine our thinking and provide more empirical explanations of why different areas experience the SDT differently. It also leads to a more thorough consideration of religion and its possible impacts on demographic behaviors. While literature often postulates that demographic transitions will accompany declines in religiosity (Kirk 1996), the Pew Research Center projects that the global religious population will actually increase in percentage between now and 2050 (2015). We hypothesize that, if accompanied by education, religious areas will not lag in second demographic transition in terms of timing, but will undergo transitions instead at different levels than their non- religious counterparts. Our study faces several limitations. Most notably, we use states as proxy for differing education and religious levels rather than testing individual education and religiosity directly. There are several reasons for this. First, data for religiosity across such a long time-span is

• unavailable. Further, cross-sectional or more contemporary longitudinal studies did not cover the

12

time-span necessary to observe the actual transition. This limitation constrains our abilities to make conclusions about causality or direct association, and does not account for other co-variates such as the racial/ethnic composition of states through time. However, despite these limitations, this study contributes meaningfully to the current

• literature. Our work supports the conclusions of Baker (forthcoming) who argues that the

education revolution did not come at the expense of religion (Gorski and Altinordu 2008). In an age of expanding education, Baker notes, spirituality and religion endure and, in certain areas like Utah, they thrive. This presents demographers and sociologists with a paradox: how can two seemingly contradictory phenomena coexist? Baker argues that these two institutions, religion and education, do not always compete but can exist symbiotically. Both education and religion provide frameworks from which individuals view the world. Further, they both provide ideologies and values as well as resources individuals can use to negotiate their family formation and processes. Future research should be done to more formally and rigorously test these assumptions. Data including both religiosity and education through time would be ideal for such tests. Such a data set could more directly test the influence of religion and education on the SDT. Further, a larger sample size would allow for more sophisticated modeling (i.e., growth curve models). Estimating counties and accounting for differing denominations, for example, would be one fruitful future

• direction. Overall, our findings encourage a more open dialogue of both education’s and religion’s

role in the demographic transitions.

13

REFERENCES Baker, David. Forthcoming. "The Great Antagonism that Never Was: Unexpected Affinities between Religion and Education in Postindustrial Society." Working Paper. Bystrov, Evgenia. 2012. “The Second Demographic Transition in Israel: One for All?” Demographic Research 27(10):261-298. Carlson, A. 2005. "The fertility gap: Recrafting American population." Family Policy Lectures, Family Research Council. Douglas, M. 1982. "The effects of modernization on religious change." Daedalus 117(3): 457- 484. Ellison, Christopher, Nicholas H. Wolfinger, and Aida I. Ramos-Wada. 2012. “Attitudes Toward Marriage, Divorce, Cohabitation, and Casual Sex Among Working-Age Latinos: Does Religion Matter?” Journal of Family Issues 34(3):295-322. Frejka, Tomas and Charles F. Westoff. 2008. “Religion, Religiousness and Fertility in the US and in Europe.” European Journal of Population 24:5-31. Glass, Jennifer and Philip Levchak. 2014. “Red States, Blue States, and Divorce: Understanding the Impact of Conservative Protestantism on Regional Variation in Divorce Rates.” American Journal of Sociology 119(4):1002-46. Gorski P S and Altinordu A (2008) After secularization? Annual Review of Sociology 34:55–85. Graham, S. and J. Donaldson. 1996. "Assessing personal growth for adults enrolled in higher education." Journal of Continuing Higher Education 44:7-22. Hernandez, Erik L., Emily Smith-Greenaway, Sarah M. Ludwig-Dehm, David P. Baker.

• Forthcoming. "Education and the Second Demographic Transition in the United States:

Population and Individual Trends." Inkeles, A. and Smith, D. 1974. Becoming modern: individual change in six developing countries, Cambridge: Harvard University Press. Kane, Jennifer B. 2013. “A Closer Look at the Second Demographic Transition in the U.S.: Evidence of Bidirectionality from a Cohort Perspective (1982-2006).” Population Research Policy Review 32(1):47-80. Kirk, Dudley. 1996. “Demographic Transition Theory.” Population Studies 50:361-387. Larson, Lyle E. and J. Walter Goltz. 1989. “Religious Participation and Marital Commitment.” Review of Religious Research 30(4):387-400. Lee, J.J. 2002. "Religion and college attendance: Change among students." The Review of Higher Education 25:369-384. Lesthaeghe, Ron. 1983. “A Century of Demographic and Cultural Change in Western Europe: An Exploration of Underlying Dimensions.” Population and Development Review 9(3):411-435. Lesthaeghe, Ron. 2010. “The Unfolding Story of the Second Demographic Transition." Population and Development Review 36(2):211-251. Lesthaeghe, Ron and Antonio Lopez-Gay. 2013. “Spatial continuities and discontinuities in two successive demographic transitions: Spain and Belgium, 1880-2010.” Demographic Research 28(4):77-136. Lesthaeghe, Ron J. and Lisa Neidert. 2006a. “The Second Demographic Transition in the United States: Exception or Textbook Example?” Population and Development Review 32(4):669-698. Lesthaeghe, Ron and Lisa Neidert. 2006b. “The ‘Second Demographic Transition’ in the US:

14

Spatial Patterns and Correlates.” Population Studies Center. Ann Arbor, MI: September 7, 2016 (http://www.psc.isr.umich.edu/pubs/pdf/rr06-592.pdf). Lesthaeghe, R. and D. J. van de Kaa. 1986. "Twee demografische transities?," Pp 9-24 in R. Lesthaeghe and D. J. van de Kaa (eds.), Groei of Krimp. Annual book issue of "Mens en Maatschappij." Deventer (Netherlands): Van Loghum-Slaterus. Mayrl, D., and J. Uecker. "Higher education and religious liberalization among young adults." Social Forces 90.1 (2011):181-208. McFarland, M., Wright, D., and Weakliem, M. (2011) “Educational Attainment and Religiosity: Exploring Variations by Religious Tradition.” Sociology of Religion 2011, 72:2 166-188. McQuillan, Kevin. 2004. “When Does Religion Influence Fertility?” Population and Development Review 30(1):25-56. Mosher, William D., Linda B. Williams, and David P. Johnson. 1992. “Religion and Fertility in the United States: New Patterns.” Demography 29(2):199-214. Norton, Arthur J. and Jeanne E. Moorman. 1987. "Current Trends in Marriage and Divorce among American Women." Journal of Marriage and Family 49:3-14. Pascarella, E. and P. Terenzini. 2005. "How college affects students: A third decade of research." San Francisco: Jossey-Bass. Pew Research Center. 2015. “The Changing Global Religious Landscape.” The Pew Forum on Religion & Public Life. Washington, D.C.: April 5, 2017 (http://assets.pewresearch.org/wp-content/uploads/sites/11/2017/04/07092755/FULL- REPORT-WITH-APPENDIXES-A-AND-B-APRIL-3.pdf). Pew Research Center. 2012. "Mormons in America: Certain in Their Beliefs, Uncertain of Their Place in Society." The Pew Forum on Religion & Public Life. Washington, D.C.: September 24, 2016 (http://www.pewforum.org/files/2012/01/Mormons-in-America.pdf). Schwadel, P. 2011. “The effects of education on Americans’ religious practices, beliefs, and affiliations.” Review of religious research 53(2):161-182. Stark, R. and R. Finke. 2000. Acts of faith: Explaining the human side of religion: Univ of California Press. Surkyn, Johan and Ron Lesthaeghe. 2004. “Value Orientations and the Second Demographic Transition (SDT) in Northern, Western and Southern Europe: An Update.” Demographic Research 3(3):45-86. Thomas, G. 1989. Revivalism and cultural change : Christianity, nation building, and the market in the nineteenth-century United States. Chicago: University of Chicago Press. Thornton, Arland, William G. Axinn, and Daniel H. Hill. 1992. “Reciprocal effects of religiosity on cohabitation and marriage.” American Journal of Sociology 98(3):628-651. Ventura, Stephanie. 2009. Changing patterns of nonmarital childbearing in the United States. NCHS Data Brief no. 18. Washington, DC: Centers for Disease Control and Prevention, National Center for Health Statistics. Available from www.cdc.gov/nchs/data/databriefs/db18.pdf.