Irreligious Socialization ? The Adult Religious Preferences of Individuals Raised with No Religion

Recent birth cohorts of Americans are more likely than previous cohorts to be raised outside of a religious tradition. In addition, those raised with no religion are increasingly likely to have no religion as adults. Despite their growing numbers, individuals raised with no religion have received little attention from scholars. The adult religious preferences of these individuals provide researchers with a unique opportunity to test theories of religion and social change. Using General Social Survey data, I examine the adult religious preferences and beliefs of individuals raised with no religion. I provide evidence of a shift in socialization and social influences experienced by those who report growing up with no religion. Compared with earlier cohorts raised with no religion, more recent cohorts have had more secular upbringings and tend to be more secular, liberal, and wary of organized religion as adults. They are also more likely to have a religiously unaffiliated spouse, if they marry at all. Results from a logistic regression analysis indicate that these trends explain much of the cohort differences in the likelihood of remaining unaffiliated as an adult.


Introduction
According to the nationally representative General Social Survey (GSS), the proportion of American adults claiming no religion remained near 7 percent from the early 1970s to the beginning of the 1990s.By the end of the century, however, that number had doubled to 14 percent.In 2010, roughly 18 percent of all GSS respondents expressed no religious preference.This increase has been corroborated by several other major national surveys and represents a substantial and relatively rapid social change. 2esearch on the growth of religious "nones" has focused primarily on individuals who disaffiliate from religion as adults, rather than those raised outside of a religious tradition.Recent studies have noted two significant trends regarding the latter group that have contributed to the overall growth of a preference for no religion (Hout & Fischer, 2002;Schwadel, 2010;Sherkat, 2001).First, recent birth cohorts are much more likely than previous cohorts to report being raised with no religion.Second, compared with earlier cohorts raised with no religion, those from more recent cohorts are much more likely to have no religion as adults.Growing numbers of Americans raised with no religion prompted the Pew Forum to devote a special section to the topic in a recent follow-up study to the 2007 U.S. Religious Landscape Survey (Pew Forum on Religion & Public Life, 2009).The release of the Pew study even led to major news outlets taking notice of this growing group (Blow, 2009;Gilgoff, 2009;Lampman, 2009).Despite some recent attention, however, we still know little about Americans raised without religion and why they increasingly remain unaffiliated as adults.Examining this group is critical for gaining a better understanding of the growth of religious non-affiliation and for uncovering social change in American society.
It is only for more recent birth cohorts that being raised with no religion is likely to result in having no religion as an adult (Hout & Fischer, 2002;Schwadel, 2010;Sherkat, 2001).A high rate of religious switching among earlier cohorts has generally been attributed to the absence of religious capital -individuals lacking religious capital have little to lose by switching (Iannaccone, 1990;Stark & Finke, 2000).This argument, however, fails to explain why switching has become far less common among later birth cohorts.Furthermore, since individuals raised with no religion presumably undergo little religious socialization, the religious choices they make as adults provide researchers with a unique test of theories of religious behavior and social change.Theories of religious socialization and social influence may help explain their religious choices, as well as uncover social change in American society.Using GSS data from 1973 to 2010, this study seeks to address this gap in the literature by examining individuals raised with no religion and their adult religious preferences.Particular attention is paid to differences between birth cohorts in early religious socialization, marital patterns, and religious and political views.Finally, this study employs logistic regression analysis to examine how these cohort differences are related to the likelihood of preferring no religion as an adult.

Religious Socialization
Research on the formation and expression of religious preferences distinguishes between socialization and social influences (Sherkat, 2003).Early religious socialization has a strong formative influence on MERINO individuals' religious preferences, while social influences affect how individuals act upon those preferences.The family has always been a crucial site of religious socialization.The theory of adaptive preferences posits that prior religious involvement conditions individuals' desires for familiar religious goods (Elster, 1983;Sherkat, 1998;Sherkat & Wilson, 1995).Parental religiosity is an important determinant of one's later religiosity, even as an adult (Myers, 1996;Sherkat, 1998).From a human capital perspective, religious involvement builds individuals' stock of religious human capital.This stock of religious capital better enables individuals to produce religious value from their experiences.Individuals generally prefer religion that best takes advantage of previously acquired religious capital (Iannaccone, 1990).
While parental religiosity has been shown to influence religious preferences throughout the lifecourse (Myers, 1996;Sherkat, 1998), it is unclear whether to expect being raised with no religion to result in a preference for no religion as an adult.The possibility of irreligious socialization has not received consideration in the literature on religious socialization.Individuals raised with no religion have presumably undergone minimal religious socialization and possess little or no religious capital.This lack of religious capital has been cited as an explanation for religious switching by those raised outside of a faith (Iannaccone, 1990).Yet, if adults' religious preferences generally reflect early socialization and their parents' preferences, we should expect that if subsequent social influences on religiosity were minimal, individuals raised with no religion would tend to prefer no religion as adults.Most individuals either remain in their denomination of origin or switch to a similar one (Hadaway & Marler, 1993).Furthermore, many individuals raised in relatively secular homes likely reach adulthood with apathetic or even negative attitudes toward religion.Individuals claiming to have been raised with no religion, however, have experienced varying levels of exposure to religion.Many acquire some religious capital during childhood or adolescence.Parental religious preferences and frequency of attendance at religious services during childhood vary among those reporting a nonreligious upbringing and may help explain which individuals express a religious preference as adults.Having parents with even marginal attachments to religion or attending church occasionally may make acquiring a religious preference as an adult more likely.Differences in religious socialization could help to explain the growing tendency for those raised with no religion to prefer no religion as adults -those from recent cohorts may have experienced more secular upbringings compared with previous cohorts.As a result, they may also hold more secular views and be more wary of religion.

Social Influences on Religiosity
Religiosity is shaped not only by early socialization, but by later social influences as well.Individuals' choices are embedded in sets of social relations that influence decision-making.Religious choices are especially prone to such influences.While rational choice models of religious behavior have enhanced our understanding of how individuals make religious decisions (Stark & Finke, 2000), there is much to be gained by considering how religious decision-making is influenced by social embeddedness (Ellison, 1995;Sherkat, 1997).Religious choices have social consequences and are thus subject to social influences.Influences such as sympathy, example-setting, and sanctions can affect religious decision-making (Sherkat, 2003).Normative constraints and access to nonreligious resources influence religious choices as well (Ellison & Sherkat, 1995;Phillips, 1998;Sherkat & Cunningham, 1998).
As a result of relatively high rates of religious belief and participation in the United States, individuals raised with no religion may face normative pressure to adopt or express a religious preference as adults.Some scholars have suggested, for example, that the over-reporting of church attendance in the United States reflects in part the social desirability of public expression of religiosity (Caplow, 1998;Hadaway, Marler, & Chaves, 1993, 1998;Hout & Greeley, 1998).Notably, Americans express serious unease about the trustworthiness of atheists (Edgell, Gerteis, & Hartmann, 2006;Farkas, Johnson, Foleno, Duffett, & Foley, 2001).While only a minority of those with no religious affiliation are atheists (Baker and Smith, 2009a;Hout & Fischer, 2002), individuals raised with no religion may attempt for social reasons to distance themselves from the ranks of nonbelievers by expressing a religious preference.Indeed, there is survey evidence of such a social desirability bias.In a meta-analytic review, Smith (1991) found that religious preference questions that explicitly mention 'no religion' as a category get more such mentions than those that did not.As Kellstedt and Green (1993, p. 59) suggest, "since it is still normative in American society to be religious, many respondents who have no behavioral or cognitive links to a religious group will nonetheless express a preference, particularly if they are not given the option to express the lack of such a preference".In a highly religious society like the United States, expressing a religious preference, even nominally, may be a convenient way of signaling to others that one is moral and trustworthy.
The growing tendency for individuals raised with no religion to prefer no religion as adults may be an indication that social desirability biases for religious affiliation are declining (Sherkat, 2001(Sherkat, , 2010)).One piece of evidence in support of this hypothesis is a set of GSS items indicating that public support for the civil liberties of anti-religious individuals has steadily and significantly increased since the 1970s (Reimer & Park, 2001).This shift in attitudes may make nonreligious or weakly religious individuals feel increasingly comfortable simply saying they have no religion.In addition, some scholars have argued that, in the latter half of the twentieth century, religious affiliation became increasingly a matter of personal choice and that non-affiliation has grown more acceptable (Hammond, 1992;Putnam & Campbell, 2010;Roof & McKinney, 1987).Noting that all birth cohorts maturing into adulthood in the 1960s or later are disproportionately more likely to disaffiliate from religion, Hout and Fischer (2002) propose a "1960s legacy" effect.An increase in the proportion of Americans raised with no religion is part of that 1960s legacy (Schwadel, 2010).
Several other social factors may influence the adult religious preferences of individuals raised outside of a religious tradition.Marriage and childrearing increase religious identification and participation (Myers, 1996;Stolzenberg, Blair-Loy, & Waite, 1994).Individuals raised with no religion are likely no exception.In addition, those who do acquire a religious preference may view marriage as a desirable and appropriate behavior.However, later birth cohorts are delaying marriage and childrearing, which may be lessening the impact of family variables on religious preference (Hout & Fischer, 2002).In addition, growing numbers of Americans with no religion means that individuals raised with no religion will be more likely to marry similarly unaffiliated partners, which could result in less switching to religion.Indeed, one of the strongest predictors of having no religion is marriage to a similarly unaffiliated partner (Baker & Smith, 2009b).Since marriage often prompts reaffiliation or conversion (Wilson & Sherkat, 1994), and switching tends to be in the direction of the more religious spouse (Stark & Finke, 2000), the increased availability of like-minded partners should further contribute to the growing stability of a preference for no religion.
What it means to express a religious preference may have changed as well.The religiously tinged political climate of the 1980s and 1990s turned off many liberal and moderate Americans to organized religion, contributing to the growth of religious "nones" (Hout & Fischer, 2002).Nonaffiliated individuals tend to be more politically and socially liberal than individuals with a religious preference (Hadaway & Roof, 1979;Hout & Fischer, 2002).It is possible that many individuals raised with no religion, particularly from later cohorts, have been similarly wary of expressing a religious prefer-MERINO ence due to a perceived association between religion and political and social conservatism.Americans who reached adulthood during the 1990s are particularly distinguished by their liberal stance on homosexuality, which is tightly linked to religious non-affiliation (Putnam & Campbell, 2010;Vargas, 2011).In short, the comparatively liberal views of those raised with no religion could discourage them from religious affiliation, further contributing to the stability of a preference for no religion.
Finally, region is likely to be an important factor in the religious choices of individuals raised outside religion.There is substantial regional variation in both religious composition and levels of religious participation in the United States.The cultural significance and meaning of religious involvement varies by region (Hammond, 1992).Studies typically point to a relatively devout South and a relatively irreligious West.Significantly, migration to less committed regions diminishes religious devotion, while moving to more pious areas enhances religious participation and the importance of faith (C.Smith, Sikkink, & Bailey, 1998).Personal religious devotion is significantly less predictive of church attendance among whites in the South than for those in other regions of the country (Stump, 1986).Normative constraints on religiosity thus vary by region of the country.Among those raised with no religion, it is likely, for example, that those living in the South are more likely to switch to a religion than those living elsewhere.

Hypotheses
This study will test a set of hypotheses that derive from the preceding discussion of religious socialization and social influences on religiosity.First, I hypothesize that more recent birth cohorts raised outside of a religious tradition have had comparatively more secular upbringings, as measured by parental religious affiliation and religious service attendance during childhood.Second, individuals from more recent cohorts are less likely to marry and, if they do marry, are more likely to have religiously unaffiliated spouses.Third, those from more recent birth cohorts are more likely to self-identify as atheist or agnostic, hold liberal political views, and lack confidence in organized religion.Finally, these cohort differences will explain much of the cohort differences in the likelihood of remaining unaffiliated as an adult.

Variables Used in the Analysis
Since 1972, the GSS has asked respondents about their current religious preference (RELIG: "What is your current religious preference?Is it Protestant, Catholic, Jewish, some other religion, or no religion?"),and, since 1973, their religion of origin (RELIG16: "In what religion were you raised?").This study treats respondents coded as "none" on RELIG16 as having been raised with no religion.The current study utilizes a number of survey items that appeared on the GSS in 1991GSS in , 1998GSS in , and 2008 as part of a special religion module.Three items measure early religious socialization.Frequency of church attendance as a child is measured by the following survey item: "What about when you were around eleven or twelve, how often did you attend religious services then?"Responses included "never," "less than once a year," "about once or twice a year," "several times a year," "about once a month," "two to three times a month," "nearly every week," "every week," and "several times a week".Two items measure parental religious preferences: "What was your [father's/mother's] religious preference when you were a child?Was it Protestant, Catholic, Jewish, some other religion, or no religion?"A final item from the 1991, 1998, and 2008 religion modules measures respondents' level of confidence in organized religion: "How much confidence do you have in churches and religious organizations?" Responses included "complete confidence," "a great deal of confidence," "some confidence," "very little confidence," or "no confidence at all".
Several items that appear more regularly on the GSS are used to examine individuals raised with no religion.Marital status is determined by an item asking respondents the following: "Are you currently married, widowed, divorced, separated, or have you never been married?"Spousal religious preference is measured by the following item: "What is your (spouse's) religious preference?Is it Protestant, Catholic, Jewish, some other religion, or no religion?"3Political orientation is measured by an item in which respondents were asked to place themselves on a seven-point scale ranging from "extremely liberal" to "extremely conservative".For simplicity, respondents self-identifying as "extremely liberal," "liberal," or "slightly liberal" are collapsed into a single "liberal" category.Belief in God is measured by an item in which respondents were asked: "Please look at this card and tell me which statement comes closest to expressing what you believe in God". 4 Respondents who selected "I don't believe in God" or "I don't know whether there is any way to find out" are considered to be atheist or agnostic, respectively.
The use of items asked only on the 1991, 1998, and 2008 surveys necessitates that the logistic regression analysis be limited to respondents interviewed in those years.Control variables used in the multivariate analyses include survey year, sex, educational attainment, region of residence, and size of the local population.Dummy variables are used to indicate survey year, with 1991 serving as the reference category.Sex is a dichotomous variable (1 = female, 0 = male).Educational attainment is measured as years of schooling.Region is measured by a dichotomous variable indicating whether the respondent resides in the South (1 = South, 0 = non-South).The variable XNORCSIZ is used to indicate the size of respondents' communities.It has been reverse-coded such that values range from 1 (= open country) to 10 (= large central city).

Results
A growing proportion of Americans report being raised with no religion, and birth cohort differences are large.Figure 1 illustrates this trend.For Americans born before 1956, being raised with no religion was quite uncommon.Less than 4 percent of individuals born before 1956 report being raised outside a faith.However, of those born between 1956 and 1970, nearly 7 percent indicate that they were raised with no religion.This number increases to nearly 11 percent for those born after 1970.

MERINO
Even more notable is the growing tendency for those raised with no religion to have no religion as adults.Figure 2 displays this pattern.Among cohorts born after 1955, a clear majority of those raised with no religion have no religion as adults.Among the most recent cohorts, over 70 percent have remained unaffiliated.While some have cautioned that recent cohorts have had less time to acquire a religious preference, particularly given that religious switching is often a function of life course events, the trend across cohorts is clear and pronounced.Moreover, it is not limited to the most recent birth cohorts.The 1944-1955 cohorts, comprised of baby boomers, show a much greater propensity than earlier cohorts to remain outside of a religious tradition after being raised with none.
Figure 3 displays cohort differences in early religious socialization among Americans raised with no religion.Among those born before 1944, the vast majority had religiously affiliated parents as children.Subsequent cohorts, particularly those born after 1970, are far more likely to report that their parents had no religion.Among the most recent cohorts, 42 percent report that their mother had no religion, while 56 percent report that their father had no preference.In addition, more recent cohorts have been less likely to attend religious services as children.Of those born after 1956, over 60 percent report that they attended services "never" or "less than once a year," compared with 42 percent for the 1944-1955 cohorts and 50 percent for the pre-1944 cohorts.Together, these results suggest that more recent cohorts raised with no religion have experienced somewhat more secular upbringings. 5Figure 4 demonstrates the sizable cohort differences in marriage patterns among Americans raised with no religion.While recent cohorts have had less time to marry, the trend toward lower rates of marriage is dramatic.Also notable is the growing tendency for those raised with no religion to marry partners with 5 These trends are also observable for Americans who report being in raised in a religious tradition and have likely contributed to the increase in religious disaffiliation.For example, nearly 11 percent of respondents born after 1970 and raised in a religion report having attended services "never" or "less than once a year".In contrast, only about 6 percent of those in the 1956-1970 cohorts and 3 percent in the pre-1956 cohorts do so.GSS, 1973GSS, -2010 (N = 54,890) (N = 54,890).no religion.For example, only 14 percent of married individuals from the pre-1944 cohorts have an unaffiliated spouse.Among married respondents born since 1971, however, nearly 30 percent have a spouse with no religion.
Compared with earlier birth cohorts raised with no religion, recent cohorts have had more secular upbringings and are more likely to marry a nonreligious spouse, if they marry at all.What are some of the consequences for their religious and political views as adults?Figure 5 offers some insights.It reports survey findings on their beliefs about God, political orientation, and views of churches and religious organizations.The results suggest that those raised with no religion are increasingly secular, liberal, and wary of organized religion.Among the pre-1944 cohorts, roughly 11 percent are atheist or agnostic.Successive cohorts raised with no religion are more likely to hold secular beliefs.Among the 1944-1955 and 1956-1970 cohorts, between 15 and 20 percent of respondents are atheist or agnostic.The most recent cohorts are even more secular -24 percent are atheist or agnostic.6Religious beliefs can change over time and may be affected by aging.Thus, these cohort differences could potentially erode somewhat.Nonetheless, the trend toward increasing secularism is notable.
Figure 2. Percent of those raised with no religion that had no religion at time of survey, by birth cohort and decade, GSS, 1973GSS, -2010 (N = 2,610) (N = 2,610).
Figure 5 also demonstrates that recent cohorts raised with no religion are more liberal and increasingly wary of churches and other religious organizations.The earliest cohorts are unlikely to self-identify as liberal -only 17 percent of those interviewed do so.Later cohorts, however, are more  GSS, 1991GSS, , 1998GSS, , 2008 (N = 286) (N = 286).
Figure 4. Marital patterns among individuals raised with no religion, by birth cohort, GSS, 1973GSS, -2010 (N = 2,610) (N = 2,610).MERINO likely to be politically liberal.For those born between 1956 and 1970, roughly a third self-identify as liberal.The most recent cohorts raised with no religion are even more liberal -38 percent, compared with roughly 20 percent who self-identify as conservative. 7Likewise, more recent cohorts raised with no religion are much more likely to express a lack of confidence in churches and religious organizations.Among the most recent cohorts, roughly 43 percent have "very little" or "no confidence at all," while only about 10 percent have "a great deal".In contrast, roughly a third of those born between 1944 and 1970 and only 15 percent of those born before 1944 express little to no confidence in religion.8Again, aging could play a role in these cohort differences.In addition, earlier cohorts of Americans tend to be more politically conservative.Nonetheless, the clear trend across cohorts suggests that a nonreligious upbringing is associated with political liberalism and a lack of confidence in churches and other religious organizations.
Data collected on the 1991, 1998, and 2008 General Social Surveys allow for an examination of factors that influence the adult preferences of individuals raised with no religion.Table 1 shows the results from a logistic regression predicting a preference for no religion among individuals who report being raised with no religion.Because of the small sample size, cohorts born before 1956 have been  GSS, 1974GSS, -2010 (N = 2,164) (N = 2,164).
IRRELIGIOUS SOCIALIZATION MERINO collapsed into a single group.Model 1 illustrates the large cohort differences in religious switching.Compared with cohorts born before 1956, the odds of having no religion for the most recent cohorts are nearly eleven times greater, while the odds for the 1956-1970 cohorts are roughly 2.4 times greater.Women are significantly less likely to stay unaffiliated.Compared with men, women have .35times the odds of remaining outside a religious tradition as adults.As expected, region plays a significant role.Southerners have roughly half the odds of remaining unaffiliated compared with non-Southerners.Model 2 adds measures of early religious socialization.As hypothesized, parental religious affiliation influences the likelihood of remaining unaffiliated.Compared with individuals whose parents both had a religious preference, those with at least one unaffiliated parent have 1.7 times the odds of expressing no religion as an adult.In contrast, those who more often attended religious services during childhood are significantly less likely to stay unaffiliated.
Model 3 demonstrates that marital outcomes play an important role in the adult preferences of those raised with no religion.As expected, marriage is significantly associated with lower odds of staying outside of a religious tradition.However, having a religiously unaffiliated spouse greatly increases the likelihood of expressing no religious preference.Model 4 adds a measure of respondents' self-placement on a political conservatism-liberalism scale.More liberal individuals are significantly more likely to have no religion at the time of the survey.Finally, Model 5 includes a dichotomous variable indicating whether respondents reported having either "very little" or "no confidence at all" in organized religion.Respondents holding such a view have roughly three times the odds of remaining unaffiliated as adults.
Finally, Table 1 demonstrates that much of the difference between cohorts in the likelihood of remaining unaffiliated is explained by these measures of religious socialization, marital status, and political and religious views.In the final model, the difference between the 1956-1970 cohorts and the pre-1956 cohorts is insignificant, while the odds ratio for the most recent cohorts is nearly cut in half compared with Model 1.These results suggest that more recent cohorts raised with no religion are more likely to remain unaffiliated at least in part because of their relatively secular upbringings, increased likelihood of marrying nonreligious partners (if they marry at all), higher levels of political liberalism, and wariness toward religion.

Discussion
The current study contributes to a growing literature examining individuals with no religion.It is unique, however, in that it focuses on the growing number of Americans who report a nonreligious upbringing.Acquiring a religious preference after being raised with no religion remains common, but is a far less frequent occurrence for later birth cohorts.While parental religiosity and early religious socialization are known to have strong and lasting effects on religiosity, even as an adult (Myers, 1996;Sandomirsky & Wilson, 1990;Sherkat, 1998), scholars have not raised the possibility that being raised with no religion could constitute a kind of irreligious socialization.Some scholars have even attributed a switch to religion among this group to the absence of religious capital, arguing that individuals lacking religious capital have little to lose by switching (Iannaccone, 1990;Stark & Finke, 2000).In contrast, the theory of adaptive preferences (Sherkat, 2003) would predict that individuals raised with no religion would generally prefer as adults that with which they are most familiar: no religion.
This study finds that variation in exposure to religion during childhood is an important factor among individuals raised with no religion.However, it is not the lack of religious capital that predicts a switch to religion in adulthood, rather the acquisition of it.Those with religiously unaffiliated parents as children are significantly less likely to express a religious preference as adults, while those who would sometimes attend religious services as children are significantly more likely to do so.These results support other studies' findings that parental religiosity and early acquisition of religious capital have long and lasting effects on religious behavior.Moreover, the current study provides evidence of a shift in socialization and social influences experienced by those who report growing up with no religion.As Table 3  during their youth, reflecting the "1960s legacy" effect proposed by Hout and Fischer (2002).In other words, only for more recent birth cohorts does being raised with no religion actually imply an upbringing relatively devoid of religion.One indication of this trend may be growing rates of atheism and agnosticism among individuals raised with no religion (see Figure 5).Alternatively, response bias may play a role -nonbelievers may be more likely to self-identify on a survey if they expect it to be less of a normatively deviant behavior.More recent cohorts may be less likely to view non-belief as deviant.
Studies have cautioned that higher rates of non-affiliation among later birth cohorts may recede as individuals have more time to acquire a religion and undergo key life course events linked to increased religiosity such as marriage and childrearing (Sherkat 2001(Sherkat , 2008;;Hout and Fischer 2002).However, Figure 2 demonstrates that the growing tendency for those raised with no religion to have no religion as adults goes back as far as the 1944-1955 cohorts.This trend has only intensified for more recent birth cohorts.Furthermore, while marriage still appears to be associated with acquiring a religious preference after being raised with no religion, it has become less predictive of such a switch.It is likely that as their numbers grow, unaffiliated individuals will be able to marry similarly unaffiliated partners, lessening the rate of switching into religion among those raised with none.It is already increasingly common for unaffiliated individuals to marry one another (Hout & Fischer, 2002), and marriage to an unaffiliated partner is a strong predictor of having no religion (Baker & Smith, 2009b).The current study suggests that much of the cohort differences in the likelihood of staying out of religion are explained by these changes in marriage patterns.
The current study also highlights the political and social dimensions of religious non-affiliation.Table 1 demonstrates that political liberalism and wariness toward religion are associated with decreased odds of a switch to religion among individuals raised with none.Furthermore, cohort differences in the likelihood of staying out of religion diminish when political orientation and views of organized religion are included in the regression analysis.Hout and Fischer (2002) argue that the religion-infused political climate of the 1980s and 1990s turned off many moderate and liberal Americans to religion.In addition, scholars have more recently argued that the social and moral liberalism of recent birth cohorts has contributed to the growth of religious "nones" (Putnam & Campbell, 2010;Vargas, 2011) The findings of the current study likewise suggest that a growing association between liberalism and a nonreligious background could explain why more people raised with no religion prefer no religion as adults.
There is a notable dearth of research on nonreligious Americans, especially those raised outside of a religious tradition.As these individuals appear in greater numbers on major surveys, this research will undoubtedly become easier.Three areas of inquiry in particular seem important for future research.First, to what extent does being raised with no religion amount to a particular form of religious socialization, or rather irreligious socialization?In other words, are individuals raised outside of a faith religious "blank slates," so to speak, or do they reach adulthood with a set of beliefs and preferences about religion?Second, we know little about the growing numbers of households with no religious affiliation.Do these families find different sources of social and emotional support than do religious families?What are the social consequences of religious non-affiliation?Third, future research should be attentive to social influences on religiosity.A notable lack of survey items on perceptions of normative constraints on religious preferences and behavior make studying social influences on religiosity a challenge.We need to better understand the social influences and normative constraints people with nonreligious upbringings encounter.Since both attendance at religious services and levels of religious belief have remained relatively stable over the past few decades (Presser and Chaves 2007;Hout and Fischer 2002), the trends observed in the current study, along with the growth of a preference for no religion in general, suggest that what has occurred is more a shift in religious identity than in religious participation in the United States.Normative constraints on nonreligious and weakly religious individuals may be declining, making these individuals more comfortable expressing no religious preference.Future research should address this possibility.
Finally, if having no religion is becoming more socially acceptable, being a nonbeliever may not be.In a recent study, atheists topped a list of groups that Americans often find troublesome in both private and public life (Edgell, et al., 2006).Most of those with no religion are in fact believers of some kind (Baker & Smith, 2009a;Hout & Fischer, 2002).Normative constraints on the nonreligious or weakly religious may have declined, but it is likely that being an atheist or agnostic is still not socially desirable.Further research is needed to examine the difference in public perception of unaffiliated believers and unaffiliated non-believers, and the impact on religious preferences and behavior.

Table 1 .
demonstrates, part of the cohort difference in the likelihood of staying out of religion, particularly between the earliest and latest cohorts, is explained by these socialization factors.Those of more recent birth cohorts raised with no religion are less likely to have exposure to religion 12 Logistic regression predicting no religious preference as an adult among those raised with no religion,GSS, 1991GSS,  , 1998GSS,  , 2008.   .
a Reference category is pre-1956 birth cohorts b Reference category is 1991 c Reference category is South