Abstract

This study describes how the career and family choices of female graduates of the Claremont Colleges within 15 years of undergraduate graduation (unless otherwise specified) have changed across the graduation years of 1960 to 1994. Specifically, I show that female graduates of the Claremont Colleges have clearly shifted away from having their family first (i.e., having at least one biological child) and a job second (i.e., having a job after 15 years of receiving their undergraduate degree but having very weak labor force attachment prior to that) toward simultaneously having both a career (i.e., very strong labor force attachment) and a family for those that graduated after 1979. Finally, I find that the primary mechanisms that allowed for the observed shift toward “career and family” for those that graduated post-1979 appear to be increased access to paid parental leave and childcare.

Introduction

Are women who have graduated from elite institutions of higher education wasting their human capital investments: that is, education? This is what recent media accounts suggest. Specifically, Belkin (2003) posited that highly educated women (particularly those from elite institutions), relative to their less-educated counterparts, are opting-out of the labor force to care for their family at higher rates today than in earlier time periods. Many other media reports followed (see, e.g., Story 2005; Trunk 2005; Wallis 2004).

Understanding the career and family choices of women who have graduated from elite institutions of higher education is particularly pertinent given the increased importance of highly educated women in the skilled workforce. Although a number of studies have examined the career and family choices (including the opt-out phenomenon) among college graduates from elite colleges and universities (see, e.g., Bertrand et al. 2010; Goldin 2006; Goldin and Katz 2008; Herr and Wolfram 2012; Hersch 2013),1 the career and family choices among college graduates exclusively from elite liberal arts colleges have received little attention.2

Several reasons explain why a focus on the career and family choices of female graduates of elite liberal art colleges is valuable. In particular, although graduates of liberal arts colleges account for only a small share of college graduates, they represent a disproportionate share of leaders in all professions, including business, medicine, public service, science, education, and the arts (Janeksela 2012). This disparity might be due to a difference in preferences, which could arise from selection (i.e., the type of people who choose to attend a liberal arts school have distinct preferences) or from the experience of a liberal arts education (i.e., a liberal arts education generates specific tastes).3 Alternatively, this could be driven by treatment (i.e., liberal arts education might prepare students differently than other types of institutions of higher education, such as highly selective research universities).4 I argue that irrespective of why graduates of liberal arts colleges end up doing relatively better in the world, the fact that they do makes it important to study their career and family outcomes.

Perhaps of greater importance, by tending to focus on the broad experiences of a limited number of cohorts, the existing literature provides an incomplete picture of the timing of employment around the timing of children. Moreover, the existing literature has limited information on the potential work–family management mechanisms (such as parental leave, work responsibilities pre- and post-birth, and childcare) that may help one ascertain why the work–family choices differ across graduation cohorts. Using a unique data set from an elite set of liberal arts colleges located in Southern California—namely, the Claremont Colleges which includes Claremont McKenna College, Harvey Mudd College, Pitzer College, Pomona College, and Scripps College5—this article seeks to fill these gaps in the literature to provide a more nuanced picture of the career and family choices of female graduates for graduation cohorts spanning several decades.

Specifically, I track the timing of employment around the timing of children within 15 years of receiving their undergraduate (BA) degree (henceforth referred to as BA graduation)—unless otherwise specified—over a large number of BA graduation cohorts (henceforth referred to as graduation cohorts) dating back to 1960 through 1994 for female graduates from the Claremont Colleges. The first sets of career and family choices I consider are simultaneously having a career (i.e., a very strong attachment to the labor market) and a family (i.e., having at least one biological child); having a career and then having a family (i.e., having at least one biological child after 15 years of BA graduation); and having a career and not having a family (i.e., not having a biological child at any time within my sample frame). I also consider three analogous categories for female graduates who pursued a “job” (i.e., had a weaker attachment to the labor market) rather than a “career.” The final sets of career and family choices I consider are having a family first and then having a job (i.e., having a job after 15 years of BA graduation but very weak labor force attachment prior to that); and having a family but not having a job. Finally, I examine a number of potential work–family management mechanisms employed by female graduates from the Claremont Colleges to help shed light on possible differences in the career and family choices of these graduates across graduation cohorts.

Looking across graduation cohorts since 1960 through 1994, I find a clear shift away from “family then job” across graduation cohorts; in fact, virtually no female graduates of the Claremont Colleges pursued this path post-1984. In contrast, there has been a clear shift toward “career and family” post-1979, particularly for female graduates of the Claremont Colleges who obtained at least one first professional (i.e., MBA, law, and medicine) and/or doctorate degree. Interestingly, increased provision of paid parental leave and access to childcare appear to be the primary mechanisms that have allowed for the observed shift toward “career and family” for those that graduated post-1979, although there is also suggestive evidence that female graduates who graduated in the most recent graduation cohort have greater access to workplace flexibility, which does not appear to come at the expense of career advancement (i.e., fewer workplace responsibilities).

Are the Career and Family Choices of Female Graduates of the Claremont Colleges Different?

Before proceeding to my primary analysis, I begin by comparing the career and family choices of female graduates from the Claremont Colleges with those of graduates from all U.S. colleges, as well as graduates from other elite institutions of higher education. This exercise serves the purpose of placing my results within the context of the broader literature. Specifically, I use data from the 2003 National Survey of College Graduates (NSCG), and following Hersch (2013), I group college graduates into four categories: all U.S. college graduates, graduates of private Research 1 universities, graduates of private Liberal Arts colleges, and graduates of public Research 1 universities.6 To ensure that female graduates have sufficient time to make their career and family choices, I restrict the NSCG sample to female graduates who graduated between 1985 and 1989. I also further restrict the sample to traditional students: that is, female graduates who obtained their undergraduate degree within five years.

For comparability, I focus on the same cohort of graduates from the Claremont Colleges sample: that is, 1985–1989. I also replicate the results from Goldin and Katz (2008) for the Harvard sample for the 1979–1982 cohort, which is the closest graduation cohort available in their analysis.7 The family variables that are available across the data sources are the percentage married (for Harvard, it is the percentage ever married) and the percentage with a first birth (for NSCG, it is the percentage living with children). The career variables include the level of education (for Harvard, the data show only if graduates pursued Law/MBA/Medicine+Doctorate degree combined; for the other two samples, there are more detailed education categories) and percentage working. All variables are measured within 15 years of BA graduation unless otherwise indicated.8

Table 1 reveals that female graduates from the Claremont Colleges are equally likely to be married relative to their counterparts from private liberal arts colleges but are more likely to be married than female graduates from private Research 1 universities. Looking at the percentage ever married, however, there appears to be no difference between female graduates from Harvard and the Claremont Colleges. Moreover, female graduates from the Claremont Colleges and Harvard appear to be less likely to have children relative to their counterparts at other elite institutions. Note, however, that the definition for percentage with children differs between the NSCG sample and the other two samples.9 Finally, female graduates from the Claremont Colleges are more (less) likely to pursue professional and doctorate degrees and be working 15 years post–BA graduation relative to private Research 1 universities and private liberal arts colleges (Harvard).10

Taken together, these patterns provide evidence that the career and family choices of female graduates of the Claremont Colleges who graduated between 1985 and 1989 differ from those at other elite institutions and U.S. college graduates in general, yet more closely align with very elite private Research 1 universities (i.e., Harvard). The remainder of the article examines the career and family choices of female graduates of the Claremont Colleges in more detail by focusing on richer data as well as tracking the patterns across a series of graduation cohorts.

Data

The data I use are based on an Alumni Survey conducted at the Claremont Colleges in the Spring and Summer of 2010. The survey included all five undergraduate institutions at the Claremont Colleges: Claremont McKenna College, Harvey Mudd College, Pitzer College, Pomona College, and Scripps College. All Claremont Colleges alumni, from every graduating year, who had an e-mail address were surveyed.11 The survey was e-mailed to 32,742 graduates, resulting in 7,107 graduates responding to the survey with 6,808 useable responses, for an overall response rate of 20.8 %.12,13 I restrict my analysis to “traditional” students (i.e., obtained their undergraduate degree within five years) who reported information on their graduation year and graduated from the Claremont Colleges between 1960 and 1994. I exclude more-recent graduates (i.e., post-1994 graduation cohorts) to ensure that the respondents have sufficient time to make their career and family choices. Specifically, unless otherwise specified, I focus on outcomes within 15 years of BA graduation, which is also consistent with the previous literature (see, e.g., Goldin and Katz 2008). I exclude the earlier graduates (i.e., pre-1960 graduation cohorts) because of a limited sample of female graduates within these graduation cohorts. Moreover, I exclude respondents with missing information on any of the variables of interest. Table 9 in the appendix provides detailed information on the number of excluded respondents by sample restriction.14 These restrictions result in a final sample size of 2,645 graduates: 1,433 female and 1,212 male.

Family

The data include detailed information on a respondent’s family (i.e., marital and birth history), which allows me to calculate the respondent’s age at first marriage, the relative timing of their marriage(s), and their graduation year from the Claremont Colleges, as well as the respondent’s age at first birth and the relative timing of their first birth and their graduation year from the Claremont Colleges. See Table 10 in the appendix for variable definitions.

Female graduates of the Claremont Colleges from the 1960–1964 cohort married relatively early; 40 % of women from this graduation cohort got married within two years of receiving their undergraduate degree (see Table 2). This marriage rate drops dramatically by the 1975–1979 cohort to 30 % for women, and continues to drop to roughly 20 % for women who graduated after 1984. Although male graduates were less likely to get married within two years of BA graduation relative to their female counterparts in the earlier graduation cohorts, the gender gap in early marriage rates closed post-1979. The average age at first marriage increased across graduation cohorts until the 1980–1984 cohort for women and remained relatively stable thereafter. Specifically, the average age at first marriage for female graduates was roughly 25 in the pre-1975 cohorts, 26 in the 1975–1979 cohort, and 27 in the post-1979 cohorts. Although first marriages occurred at older ages for male graduates relative to their female counterparts in the earlier graduation cohorts, the gender gap in the age at first marriage closed post-1979.

Overall, roughly 79 % of the women were married within 15 years of BA graduation irrespective of graduation cohort. Moreover, there was a drop in the separation/divorce/widowed rates within 15 years of BA graduation of female graduates in the later graduation cohorts relative to the earlier graduation cohorts. In general, there is no gender gap in either marriage rates or separation/divorce/widowed rates, irrespective of graduation cohort.15

In terms of fertility, roughly 62 % of women had at least one biological child within 15 years of BA graduation in my sample irrespective of graduation cohort. Although very few graduates had their first biological child within five years of BA graduation, this number is substantially higher for graduation cohorts in the 1960s. For example, 32 % of women from the 1960–1964 cohort had their first child within five years of receiving their undergraduate degree, while 15 %, 6 %, and 2 % of the women from the 1970–1974, 1980–1984, and 1990–1994 cohorts, respectively, had their first birth within this same time frame. In general, there is no gender gap in having at least one biological child within five or 15 years of BA graduation, irrespective of graduation cohort.16 There remains a substantial fraction of Claremont Colleges female graduates, irrespective of graduation cohort, who have not had a first birth 15 years post–BA graduation (i.e., roughly 38 %).

Turning to the timing of fertility within 15 years of BA graduation, female graduates’ average age at the birth of their first biological child increased across all graduation cohorts from age 27 in the 1960–1964 cohort to age 31, 10 years later; age 32, 20 years later; and age 33, 30 years later. Although female graduates were younger than their male counterparts at the time of the birth of their first biological child pre-1970, the most recent cohort of female graduates were actually older than their male counterparts at the time of the birth of their first biological child. On average, women who graduated from the Claremont Colleges had two biological children (conditional on having a first birth), irrespective of graduation cohort. This average, however, masks some interesting differences in the distribution of children within families across graduation cohorts. Women in more recent graduation cohorts are less likely to have three or more children (conditional on having a first birth) than earlier graduation cohorts. Interestingly, pre-1985 male and female graduates had families of roughly the same size, yet post-1984 male graduates had larger families relative to their female counterparts.

Career

Education

The data also include detailed information on a respondent’s education, which allows me to calculate whether they pursued a graduate degree—and, if so, the type of degree they pursued and the timing of their graduate work relative to their graduation from the Claremont Colleges (see Table 10 in the appendix for variable definitions). Table 3 presents descriptive statistics on educational attainment within 15 years of BA graduation by gender and graduation cohort. The percentage of women who obtained at least one first professional (i.e., MBA, law, and medicine) and/or doctorate degree increased substantially until the 1970–1974 cohort (i.e., 9 % in the 1960–1964 cohort, 19 % in the 1965–1969 cohort, and 32 % in the 1970–1974 cohort) and remained relatively stable thereafter. Although female graduates of the Claremont Colleges were less likely to pursue at least one first professional and/or doctorate degree relative to their male counterparts, they closed this gender gap post-1984: the difference between male and female attainment rates of these degree types is no longer statistically significant post-1984.

Looking within the broad category, the percentage of women pursuing a first professional degree, with the exception of the earlier two graduation cohorts, was relatively stable across graduation cohorts irrespective of type of degree. Although female graduates of the Claremont Colleges continue to lag behind their male counterparts in terms of first professional degrees in the most recent graduation cohorts (this gender gap, however, is not statistically significant in the 1985–1989 cohort), this discrepancy is largely driven by the gender gap in MBA obtainment. Interestingly, female graduates of the Claremont Colleges have completely closed the education gender gap in doctorate degrees post-1979, although to a large degree because male graduates have become less likely to pursue a doctorate across graduation cohorts.

The incidence of master’s (excluding MBA) degrees is relatively stable across graduation cohorts, although there is some evidence that women who pursued a master’s degree in education (conditional on pursuing a master’s degree) declined from 32 % in the 1960–1964 cohort to roughly 20 % to 25 % going forward. Finally, the gender gap in terminal BA degrees (i.e., not pursuing a degree beyond a BA within 15 years of acquiring the BA) is completely erased post-1979 both because female graduates are more likely to pursue graduate degrees across graduation cohorts and because male graduates are less likely to pursue graduate degrees post-1979.17

Labor Market Attachment

The survey also has extremely rich data pertaining to a respondent’s employment history, including current employment, most recent employment if not employed, and employment in all other jobs (up to 10 jobs) back to the first job held for a period of at least six months post–BA graduation.18 This detailed employment history allows me to calculate the respondents’ employment status, their actual labor market experience, their years out of the labor force, and whether they incurred any spells out of the labor market in two periods—currently and within 15 years of BA graduation—conditional on completion of their formal education. See Table 10 in the appendix for variable definitions.

For each job, respondents were asked the year, their full-time work status, their occupation, their job title, and the type of employer both at the start and end of the job.19 Although I have incomplete information on the timing of full-time and part-time employment spells throughout a respondent’s career, I have detailed information on this dimension surrounding the birth of a respondent’s first child, which arguably is the period when many women make this type of labor supply adjustment. I discuss this in detail in the upcoming section, “Potential Work–Family Management Mechanisms.”

According to Table 4, female labor force participation (15 years post–BA graduation) has increased in the most recent graduation cohort relative to earlier graduation cohorts (pre-1975). Specifically, roughly 77 % to 80 % of women who graduated before 1975 were employed compared with 87 % for those women who graduated in the 1990–1994 cohort.20 Although the gender gap in labor force participation 15 years post–BA graduation has narrowed across graduation cohorts, women who graduated in the 1990–1994 cohort are still 9 percentage points less likely to participate in the labor force relative to their male counterparts.

The percentage of female graduates without any spells out of the labor market within 15 years of BA graduation generally increases across the graduation cohorts. In particular, 33 % of women who graduated from the Claremont Colleges between 1960 and 1964 had no spells out of the labor market compared with 51 %, 62 %, and 65 % who graduated 10, 20, and 30 years later, respectively. This increase observed for female graduates across graduation cohorts has led to a substantial decline in the gender gap in uninterrupted employment from a gap of 36 percentage points in the 1960–1964 cohort to roughly 7 percentage points in the 1990–1994 cohort. Finally, female graduates experienced a substantial drop in the length of out-of-the-labor-force spells within 15 years of BA graduation across graduation cohorts (i.e., a drop of almost three years between the earliest and latest graduation cohort for women), and have closed the gender gap in the length of spells out of the labor force in the most recent graduation cohort.

Career and Family Choices

How do women who graduated from the Claremont Colleges combine career and family? Do the career and family decisions of female graduates of the Claremont Colleges differ across graduation cohorts? Do they follow a similar pattern as that described in Goldin (2004) for female U.S. college graduates in general? Specifically, Goldin (2004) described the career and family choices/aspirations of five cohorts of female U.S. college graduates as follows: cohort 1 (1900–1919) had “family or career”; cohort 2 (1920–1945) had “job then family”; cohort 3 (1946–1965) had “family then job”; cohort 4 (1966–1979) wanted to achieve “career then family,” and 13 % to 18 % did so by age 40; and finally, cohort 5 (1980–1990) desired “career and family,” and 21 % to 28 % did so by age 40.21

I first attempt to answer these questions by examining the level of labor market attachment of female graduates of the Claremont Colleges within 15 years of BA graduation across graduation cohorts and across education levels by the presence of family (i.e., the birth of at least one biological child within 15 years of BA graduation). Table 5 reveals that women with at least one biological child appear to have longer spells out of the labor force than their childless counterparts irrespective of graduation cohort or level of education, although the difference is less stark for the post-1984 graduation cohorts and for women who obtained first professional degrees (i.e., MBA, law, and medicine).22 Although there is very little variation in the length of spells out of the labor force for childless women across graduation cohorts, there is a substantial decline in these spells for women with at least one biological child across graduation cohorts. Specifically, the two biggest drops occur between the 1960–1964 cohort (5.17 years) and the 1965–1969 cohort (3.33 years), and between the 1980–1984 cohort (2.11 years) and the 1985–1989 cohort (1.29 years).23 To ascertain whether the number of biological children influences the length of spells out of the labor market, I also further categorize women with at least one biological child as follows: women with one biological child, women with two biological children, or women with three or more biological children. The patterns for these subgroups are generally very similar to the overall patterns, although they are less precisely estimated because of small sample sizes. In addition, the difference between childless women and women with one biological child within a given graduation cohort is less pronounced. These patterns suggest that women who graduated from the Claremont Colleges do not appear to be opting out of the labor market at higher rates in more recent cohorts as Belkin (2003) suggested.24

There is little evidence that women with children have different spells out of the labor force relative to their childless counterparts if they pursued a law degree, a medical degree, or a MBA degree. Although there is some evidence that women with children who pursued a doctorate degree do have somewhat longer spells out of the labor force, the spells are less than one year for both childless women and women with children. For women with a master’s degree (terminal BA degree), having children results in approximately one (two) additional year(s) out of the labor force relative to not having children within 15 years of graduating from the Claremont Colleges.25

To this point, I have documented that women with families who graduated in the most recent graduation cohorts appear to be more attached to the labor market than their earlier counterparts. I have not, however, provided direct evidence on how female graduates of the Claremont Colleges have combined career and family. Therefore, I create a hierarchy of career and family choices to see how these choices have changed across graduation cohorts. Specifically, I create the following mutually exclusive categories:

  • Career and family

  • Career then family

  • Career no family

  • Job and family

  • Job then family

  • Job no family

where “career” indicates a female graduate had no spells out of the labor force within 15 years of BA graduation; “job” indicates a female graduate had one to five years out of the labor force within 15 years of BA graduation; “and family” indicates that a female graduate had at least one biological child within 15 years of BA graduation; “then family” indicates that a female graduate had at least one biological child after 15 years of BA graduation; and “no family” indicates that a female graduate did not have a biological child at any time within my sample frame.26 For the three “career” categories, I also look at the subsamples of female graduates who acquired at least one first professional and/or doctorate degree within 15 years of BA graduation and those who did not, given that the former subsample might be more career oriented than the latter subsample.

I also consider the following two additional mutually exclusive categories:

  • Family then job

  • Family no job

where “family” indicates that a female graduate has at least one biological child within 15 years of BA graduation; “then job” indicates that a female graduate had six to 15 years out of the labor force within 15 years of BA graduation and worked after 15 years of BA graduation; and “no job” indicates that a female graduate had six to 15 years out of the labor force within 15 years of BA graduation and did not work after 15 years of BA graduation.27

Table 6 reveals several noteworthy patterns. There has been a clear shift away from “family then job” across graduation cohorts; in fact, virtually no female graduates of the Claremont Colleges pursued this path in the 1985–1989 or the 1990–1994 cohort (i.e., roughly 2 %) compared with 28 % in the 1960–1964 cohort; 12 %, 10 years later; and 7 %, 20 years later. There is also some weak evidence that female graduates post-1979 were less likely to choose “job and family” (i.e., roughly 17 %) relative to their counterparts who graduated between 1965 and 1979 (i.e., roughly 26 %).

In contrast, there has been a clear shift toward “career and family,” which reached its steady state level by the 1980–1984 cohort at roughly 39 % compared with roughly 31 %, 10 years earlier; and 15 %, 20 years earlier. This shift toward “career and family” appears to be driven by female graduates of the Claremont Colleges who obtained at least one first professional and/or doctorate degree. Specifically, post-1979, roughly 17 % of female graduates who obtained a first professional and/or doctorate degree achieved “career and family” compared with roughly 0 % and 9 % in the 1960–1964 and 1970–1974 cohorts, respectively. In comparison, female graduates who did not obtain a first professional or doctorate degree reached a steady state level by the 1980–1984 cohort at roughly 20 % compared with 15 % 20 years earlier.28

These patterns suggest that women who graduated from the Claremont Colleges may follow a different path toward “career and family” than their counterparts from U.S. colleges in general. Specifically, unlike Goldin (2004), I do not find evidence of a large group of women graduating from the Claremont Colleges pursuing “career then family” in the graduation years between 1965 and 1979. Moreover, although I find that the largest transition away from “family then job” does occur between the 1960–1964 cohort (27.7 %) and the 1965–1969 cohort (14.8 %), it does not reach its steady state until post-1984, when almost no female graduates from the Claremont Colleges are pursuing this track. This suggests that female graduates of the Claremont Colleges may have moved slower than U.S. college graduates in general who shifted away from the “family then job” track post-1965 (Goldin 2004).

Potential Work–Family Management Mechanisms

What work–family management mechanisms allowed female graduates of the Claremont Colleges who graduated post-1979 to shift toward “career and family”? I begin by examining women’s employment patterns, including full-time status, surrounding their first birth within 15 years of BA graduation because the summary measure of labor force attachment that underlies my “career” definition may obscure some important underlying work–family choices. I then explicitly examine several potential work–family management mechanisms—access to the same job/occupation; work responsibilities (including access to flexible work arrangements); access to parental leave; and childcare responsibilities—for women who were employed both the year before and the year after the birth of their first biological child, unless otherwise specified.

Table 7 reveals that women are more likely to be employed and working full-time the year before and the year after the birth of their first biological child in more recent graduation cohorts (post-1980) relative to earlier graduation cohorts (pre-1980). One can also see a big increase in the number of women who worked full-time in the year before and shifted to part-time work in the year after post-1984. This increase suggests that one of the mechanisms allowing women to stay attached to the labor market is the ability to work part-time in the year after the birth of their first biological child. Moreover, relative to the earliest graduation cohort (1960–1964), there is a big drop in the number of women who worked in the year before and exited the labor market in the year after for all successive graduation cohorts. Finally, the latter two graduation cohorts (post-1984) are much less likely to have not been employed in the year before and year after the birth of their first biological child. Similar patterns are found if one looks at the year before and the year after in isolation (see Table 7). Taken together, these patterns surrounding the birth of the first biological child provide further support that the most recent graduation cohorts are more career-oriented and have a desire to pursue both career and family. Given this, I now explicitly examine the relative importance of the potential work–family management mechanisms described earlier.

According to Table 8, the majority of women had the same job and the same occupation one year after the birth of their first child relative to one year before, irrespective of graduation cohort. There is, however, weak evidence that for women who did change jobs, the probability of staying in the same occupation increased in the 1990–1994 cohort relative to earlier graduation cohorts. There is also evidence to suggest that women in the most recent graduation cohort (1990–1994) enjoyed more flexible work hours as well as more hours for pay at home, and traveled less for work in the year after the birth of their first biological child relative to the year before; interestingly, this increased access to workplace flexibility was coupled with more work responsibility. These patterns are encouraging—especially in light of the aforementioned evidence on the shift from full-time status to part-time status following childbirth—because they suggest that employers may be adopting more work- and family-friendly policies that afford women the opportunity to not only stay attached to the labor market but also to pursue a “career” within their occupation of choice.

The biggest changes, however, appear to be due to access to paid parental leave and childcare (conditional on being employed pre- and post-childbirth). Specifically, less than 25 % of women who graduated from the Claremont Colleges prior to 1970 had access to paid parental leave. Access to paid parental leave increased to 40 % in the 1970–1974 cohort and further increased to 68 % (79 %) 10 (20) years later. Although the number of weeks of paid parental leave also increased across the graduation cohorts, if one conditions on taking a paid parental leave, the number of weeks is relatively stable across graduation cohorts. Moreover, there is little evidence to suggest that access to unpaid parental leave differs across graduation cohorts. There is, however, some evidence that the number of weeks of unpaid parental leave (conditional on taking an unpaid leave) declined post-1974.

There has also been a clear shift in the childcare responsibilities for the first biological child across graduation cohorts. Specifically, the share of childcare responsibilities undertaken by the woman herself declined substantially across graduation cohorts. For example, the mother was responsible for 70 % of childcare responsibilities in the 1960–1964 cohort, and this percentage fell to 54 %, 50 %, and 41 % in the 1970–1974, 1980–1984, and 1990–1994 cohorts, respectively. This large decline in women’s role in childcare shifted to either their spouse or toward daycare/preschool, although the shift to their spouse seems to have occurred earlier (post-1969) than the shift to daycare/preschool (post-1979). Despite this increase in spousal participation in childcare responsibilities, it does not appear that the male graduates in my sample (who may serve as proxies for spouses) have experienced any large shocks to their labor supply (i.e., increases in their years out of the labor force or a shift from full-time to part-time employment) that their female counterparts continue to experience (see Table 4).

Conclusions

Using a unique data set from an elite set of liberal arts colleges (i.e., the Claremont Colleges), I examine the career and family choices of female graduates across graduation cohorts since 1960 through 1994. My results show that more recent female graduates of the Claremont Colleges—compared with earlier graduation cohorts—shifted toward delaying their families (i.e., the birth of their first biological child). However, within 15 years of graduating from their undergraduate (BA) degree, there is virtually no difference in the timing of families across graduation cohorts (i.e., roughly 62 % of female graduates had their first birth within 15 years of BA graduation). At the same time, female graduates of the Claremont Colleges from more recent graduation cohorts are more likely to pursue their careers within 15 year of BA graduation (i.e., obtain advanced degrees and stay attached to the labor market) relative to their counterparts from earlier graduation cohorts.

Upon further examination of the timing of employment around the timing of children within 15 years of BA graduation, I find that female graduates from the Claremont Colleges have almost entirely shifted away from “family then job,” with only roughly 1 % of female graduates pursuing this track post-1984; and have shifted toward “career and family,” with roughly 40 % of female graduates pursuing this track post-1979. Interestingly this shift toward “career and family” appears to be driven by female graduates of the Claremont Colleges who obtained at least one first professional and/or doctorate degree. Specifically, no female graduates who obtained at least one first professional and/or doctorate degree achieved “career and family” in the 1960–1964 cohort; while post-1979, 17 % of these women did achieve “career and family.” In contrast, 15 % of female graduates who did not obtain a first professional or doctorate degree achieved “career and family” in the 1960–1964 cohort, and this metric reached its steady state of roughly 20 % post-1979.

Although Goldin (2004) also found a shift toward “career and family” for female graduates from U.S. college in general in more recent graduation cohorts (post-1980), she described the path that they took to get there as follows: “family then job” for female graduates who graduated in 1946–1965 to wanting to achieve “career then family” for female graduates who graduated in 1966–1979 with 13 % to 18 % doing so by age 40, and finally to desiring “career and family” for female graduates who graduated in 1980–1990 with 21 % to 28 % doing so by age 40. Although I find that the largest transition away from “family then job” does occur between the 1960–1964 cohort (27.7 %) and the 1965–1969 cohort (14.8 %), I find that very few female graduates of the Claremont Colleges pursue “career then family” irrespective of graduation cohorts (roughly 4 %). Thus, there is suggestive evidence that female graduates of the Claremont Colleges followed a different path to “career and family.” Finally, I find that the primary mechanisms that allowed for the observed shift toward “career and family” in more recent graduation cohorts appear to be increased access to paid parental leave and childcare. There is also, however, suggestive evidence that female graduates who graduated in the most recent cohort have greater access to workplace flexibility, and this does not appear to come at the expense of career advancement (i.e., fewer workplace responsibilities). Further understanding the potential work–family management mechanisms that will allow more women to successfully achieve “career and family” is an extremely important avenue for future research.29

Acknowledgments

I thank Mary Evans, Serkan Ozbeklik, Oana Tocoian, Tomoe Kanaya, Sherylle Tan, Kirsti Zitar, Hiram Chodosh, and participants at both the Gender and Family in the New Millennium, Broom Workshop in Santa Barbara, CA, and the Claremont McKenna Workshop in Claremont, CA, for comments and suggestions on earlier versions of this article. I also thank the Alumni offices at each of the five Claremont Colleges for administering the Alumni Survey on my behalf. In addition, I thank Claudia Goldin, Marianne Bertrand, and Jeffrey Greenhaus for allowing me to fashion some of the questions in the Claremont Colleges Alumni Survey after the surveys they conducted at their respective institutions (i.e., Harvard University, University of Chicago Booth School of Business, and Drexel University LeBow College of Business, respectively). I gratefully acknowledge research support provided by both the Berger Institute for Work, Family and Children and the Lowe Institute of Political Economy, Claremont McKenna College.

Appendix

Notes

1

A number of studies have examined the opt-out revolution among college graduates in general. See, for example, Antecol (2011), Boushey (2005, 2008), Cohany and Sok (2007), Fortin (2013), Goldin (2006), Percheski (2008), and Vere (2007). Several studies have also sought to understand why highly educated women are exiting the labor market and what the associated costs of exiting may be (see, e.g., Hewlett 2007; Hewlett et al. 2005; Stone 2008; Stone and Lovejoy 2004; Williams et al. 2006); these studies tend to be more qualitative in nature. Finally, several studies have examined the fertility side of opting-out (see, e.g., Percheski 2008; Shang and Weinberg 2013; Vere 2007).

2

Although Goldin (2006) did include several liberal arts colleges in her analysis of the career and family choices of female graduates from one graduation cohort (i.e., the entering class of 1976), these liberal arts colleges are looked at in conjunction with other highly selective universities (i.e., the sample is pooled across all 34 highly selective colleges and universities). Moe and Shandy (2010) examined the career and family choices of 857 alumnae who graduated between the years of 1970 and 2006 from a liberal arts college; the authors argued, however, that their sample of female college graduates is not “elite.”

3

The learning environment and pedagogical philosophies at liberal arts colleges (e.g., small size, residential nature, student orientation of its faculty, and student–student and student–faculty interactions) are different than those at highly selective research universities (see, e.g., Astin 1999 and Pascarella et al. 2004, 2005).

4

Although the empirical literature on the impact of liberal arts education on student preparedness is limited and tends to suffer from methodological problems (Pascarella et al. 2005), it provides suggestive evidence that liberal arts colleges, particularly elite institutions, prepare students differently relative to other institutions (see, e.g., Astin 1999; Pascarella and Terenzini 1991; Pascarella et al. 2005; Umbach and Kuh 2006). There is also some evidence to suggest that liberal arts experiences positively influence both short-term (Seifert et al. 2008) and long-term (Pascarella et al. 2005) student learning outcomes.

5

On average, the Claremont Colleges (as a whole) ranked roughly in the top 20 of the U.S. News & World Report Best Colleges Liberal Arts Colleges Rankings in 2010 (of 100 plus colleges). There is, however, variation across the Claremont Colleges: Claremont McKenna College ranked 11, Harvey Mudd College ranked 14, Pitzer College ranked 49, Pomona College ranked 6, and Scripps College ranked 25.

6

Research 1 universities (whether private or public) offer a full range of baccalaureate programs, offer graduate education through a doctorate degree, place a high priority on research, award at least 50 doctoral degrees per year, and receive at least $40 million in federal support annually (Hersch 2013:478).

7

The other graduation cohorts they examined are 1969–1973 and 1989–1992.

8

For the NSCG sample, the variables are measured within 14–18 years of BA graduation. For expositional ease, however, this is delineated as “within 15 years of BA graduation.” This allows for large enough sample sizes across the college and university categories in the NSCG sample.

9

The mean age at first birth also tends to be lower among college-educated women in the United States generally compared with graduates from the Claremont Colleges or Harvard. Specifically, the mean age at first birth for college-educated women in the United States was 24.9 in 1970–1974 and 29.7 in 1990–1994 (Martin 2004), while it was roughly 31 to 33 since the 1970 graduation cohort for female graduates from the Claremont Colleges (see Table 2) and Harvard (Goldin and Katz 2008).

10

Women who graduated from the Claremont Colleges are substantially more likely to obtain degrees in law, medicine, MBA, and/or doctorates than the national average, even among the 1970–1974 cohort, but less likely than their Harvard counterparts. For example, 1 % (0.4 %) and 3.3 % (1.2 %) of all female college graduates entered law (medical) school in 1970 and 1990, respectively (Goldin and Katz 2008), while 10.0 % (6.0 %) and 8.0 % (5.0 %) of female graduates from the Claremont Colleges completed law (medical school) in the 1970–1974 and 1990–1994 cohorts, respectively (see Table 3) and 17.3 % (18.2 %), and 20.4 % (16.6 %) of female Harvard graduates completed law (medical school) in the 1970 and 1990 cohorts, respectively (Goldin and Katz 2008).

11

The survey was administered by the Alumni offices at each of the five colleges on my behalf. Not all the e-mail addresses that the Alumni offices had were viable: roughly 3 % to 7 % failed upon delivery, depending on the college.

12

The overall response rates varied somewhat by college: Claremont McKenna College, 17.4 %; Harvey Mudd College, 20.3 %; Pitzer College, 19.2 %; Pomona College, 23.0 %; and Scripps College, 21.7 %. Discussions that I had with the Alumni offices revealed that these response rates are in line with these colleges’ typical response rates for Alumni surveys sent via e-mail to all graduates from all graduating years. Pitzer College did not conduct any other e-mail surveys within the last four years; thus, there is no comparison group for this college.

13

Preliminary investigations of the characteristics of the graduates at four of the five colleges in general (i.e., age and graduation cohort) appear to be in line with the respondents of this survey (one college has not supplied information on their graduates in general to date). However, the gender composition of the survey respondents appears to be somewhat skewed toward women relative to the overall gender composition of the graduates at three of the four colleges for which I have information (one college is female only), particularly in more recent graduation cohorts. For the colleges for which I have information, I weighted the survey responses by one of two weights: (1) the overall population by graduation cohort, gender, and college; and (2) the survey population divided by the overall population by graduation cohort, gender, and college. The weighted results are extremely similar to the unweighted results irrespective of the weight employed.

14

Although I find some evidence that the estimation sample is selected based on education and college, using the actual gender composition of each of the Claremont Colleges by graduation cohort (with the exception of Harvey Mudd College, where I use the sample gender composition), I find that a large fraction of the respondents with missing demographic information appear to be men. Given my main interest is examining the career and family choices of female graduates, my main results should not suffer from severe bias due to sample selection. Having said this, I do try to assess how biased my main results are for female graduates based on sample selection using imputed data for respondents with missing gender based on the actual gender composition of each college by graduation cohort. Specifically, I assign these additional observations the educational distribution of the sample with missing gender, as well as the family status distribution and the time out of the labor force distribution (by family status) for the female sample with demographic information by college and graduation cohort. I find that the overall patterns continue to hold, although the sample that includes both observations with and without missing demographic information appear to have somewhat longer spells out of the labor force irrespective of graduation cohort, presence of children, or level of education. Results are available upon request.

15

The main exceptions are that in the earliest graduation cohort, more (less) men were married (separated/divorced/widowed) relative to their female counterparts; while in the most recent graduation cohort, fewer men were married relative to their female counterparts.

16

The main exceptions are that in the earliest graduation cohort, more men had at least one biological child within 15 years of BA graduation relative to their female counterparts; while in the most recent graduation cohort, less (more) men had at least one biological child within 15 (five) years of BA graduation relative to their female counterparts.

17

Given the substantial growth in the U.S. financial sector over the past 60 years (see Philippon and Reshef 2012), this shift for men may be due to a higher pursuit of jobs in finance, which may not require graduate work to obtain them.

18

If they had more than 10 jobs, they were asked to give detailed information on their first job post–BA graduation as well; for all but a handful of respondents, 10 jobs were more than adequate. Of those that reported an 11th job (roughly 30 respondents), it generally was the job that came right after their 10th job.

19

This does not allow me to accurately calculate the respondent’s full-time attachment to the labor market throughout their career as one may have chosen to ramp up or ramp down at different points in their job, and this status change would not be reflected in their start- and end- year full-time status. For example, if a respondent indicated in Job 1 that she worked full-time at the start of the job in 1980 and full-time at the end of the job in 2000, it is unclear whether she consistently worked full-time in the years in between 1980 and 2000.

20

One may be concerned that the level of labor market participation of female graduates from the 1990–1994 cohort may have been negatively impacted by the financial crisis. I argue that this is less of a concern for my sample: roughly 79 % of female graduates from the Claremont Colleges in the 1990–1994 cohort are white, and as Hoynes et al. (2012) pointed out, minority groups were impacted more heavily in terms of employment by the financial crisis relative to their white counterparts.

21

In the latter two cohorts of female U.S. college graduates, Goldin (2004) defined “family” as having a first birth irrespective of when the birth occurred or if the woman was married (currently or ever) at the time of the first birth and “career” as having earnings (annual income or hourly wage) greater than those earned by their male counterparts at the 25th percentile of the male earnings distribution for two or three years (consecutive if available) by roughly age 40.

22

This is similar to women who graduated from 34 selective colleges and universities between 1978 and 1985. Women with children had 2.08 years out of the labor force (excluding education and spells of less than six months) within 15 years of graduation compared with 0.41 year for their childless counterparts (Goldin 2006). Female graduates from Harvard without children also have fewer years out of the labor market than their counterparts with children (Goldin and Katz 2008).

23

Goldin and Katz (2008) similarly found that if anything, women with children who graduated from Harvard in more recent cohorts take less time rather than more time off relative to their counterparts from earlier cohorts. For example, the mean of all out-of-the-labor-market spells within 15 years of graduation for women with children was 24 months for the 1970 cohort, 20 months for the 1980 cohort, and 19 months for the 1990 cohort.

24

These patterns are also consistent with those found for U.S. college graduates generally. Specifically, recent studies have generally concluded that there is little or no evidence to support the popular press’ notion of an opt-out revolution, (see, e.g. Antecol 2011; Boushey 2005, 2008; Cohany and Sok 2007; Fortin 2013; Goldin 2006; Percheski 2008; Vere 2007).

25

Women with advanced degrees who graduated from 34 selective colleges and universities generally had shorter spells out of the labor market than other female graduates, even those female graduates with children (Goldin 2006).

26

Some female graduates in the most recent graduation cohort (1990–1994) may be wrongly classified in the “no family” category given they may in fact have their first child after 2010. Therefore, I may be underestimating (overestimating) the percentage of female graduates who are in the “then family” (“no family”) category.

27

I am unable to consider the “no family then job” (N = 14), “then family then job” (N = 4), and “no family no job” (N = 2) categories because of small sample sizes.

28

Although there appears to be an increase in the percentage of female graduates achieving “career no family” in the most recent graduation cohort, this is likely due to the fact that this is an “overestimate” for this group because they may have a biological child post-2010, but I cannot observe this.

29

Although sociologists have examined the effect of work–family policies on women’s labor supply choices, economists have largely overlooked this. The main exception to the best of my knowledge is Herr and Wolfram (2012).

References

Antecol, H. (
2011
).
The opt-out revolution: Recent trends in female labor supply
.
Research in Labor Economics
,
33
,
45
83
. 10.1108/S0147-9121(2011)0000033005
Astin, A. W. (
1999
).
How the liberal arts college affects students
.
Daedalus
,
128
(
1
),
77
100
.
Belkin, L. (
2003
,
October
26
).
The opt-out revolution
.
New York Times Magazine,
42
47
, 58,
85
86
.
Bertrand, M., Goldin, C., & Katz, L. F. (
2010
).
Dynamics of the gender gap for young professionals in the financial and corporate sectors
.
American Economic Journal: Applied Economics
,
2
,
228
255
.
Boushey, H. (
2005
).
Are women opting out? Debunking the myth (CEPR Briefing paper)
.
Washington, DC
:
Center for Economic Policy Research
.
Boushey, H. (
2008
).
Opting out? The effect of children on women’s employment in the United States
.
Feminist Economics
,
14
(
1
),
1
36
. 10.1080/13545700701716672
Cohany, S. R., & Sok, E. (
2007
).
Trends in labor force participation of married mothers of infants
.
Monthly Labor Review, February,
9
16
.
Fortin, N. M. (
2013
).
Gender role attitudes and women’s labor market participation: Opting-out, AIDS, and the persistent appeal of housewifery
. Unpublished manuscript,
Department of Economics, University of British Columbia
,
Vancouver, Canada
. Retrieved from http://faculty.arts.ubc.ca/nfortin/Fortin_Gender.pdf
Goldin, C. (
2004
).
The long road to the fast track: Career and family
.
Annals of the American Academy of Political and Social Science
,
596
,
20
35
. 10.1177/0002716204267959
Goldin, C. (
2006
).
The quiet revolution that transformed women’s employment, education, and family
.
American Economic Review Papers and Proceedings
,
96
,
1
21
. 10.1257/000282806777212350
Goldin, C., & Katz, L. F. (
2008
).
Transitions: Career and family life cycles of the educational elite
.
American Economic Review Papers and Proceedings
,
98
,
363
369
. 10.1257/aer.98.2.363
Herr, J. L., & Wolfram, C. (
2012
).
Work environment and opt-out rates at motherhood across high-education career paths
.
Industrial and Labor Relations Review
,
65
,
928
950
. 10.1177/001979391206500407
Hersch, J. (
2013
).
Opting out among women with elite education
.
Review of the Economics of the Household
,
11
,
469
506
. 10.1007/s11150-013-9199-4
Hewlett, S. A. (
2007
).
Off-ramps and on-ramps: Keeping talented women on the road to success
.
Boston, MA
:
Harvard Business School Press
.
Hewlett, S. A., Luce, C. B., Shiller, P., & Southwell, S. (
2005
).
The hidden brain drain: Off-ramps and on-ramps in women’s careers (Harvard Business Review Research Report No. 9491-PDF-ENG)
.
Boston, MA
:
Harvard Business Publishing
.
Hoynes, H. W., Miller, D. L., & Schaller, J. (
2012
).
Who suffers during recessions?
.
Journal of Economic Perspectives
,
26
(
3
),
27
48
. 10.1257/jep.26.3.27
Janeksela, G. M. (
2012
).
The value of a liberal arts education
.
Academic Exchange Quarterly
,
16
(
4
),
37
42
.
Martin, S. (
2004
).
Delayed marriage and childbearing: Implications and measurement of diverging trends in family timing
. In Neckerman, K. (Ed.),
Social inequality
(pp.
79
119
).
New York, NY
:
Russell Sage
.
Moe, K., & Shandy, D. (
2010
).
Glass ceilings & 100-hour couples: What the opt-out phenomenon can teach us about work and family
.
Athens, GA
:
The University of Georgia Press
.
Pascarella, E. T., & Terenzini, P. T. (
1991
).
How college affects students
.
San Francisco, CA
:
Jossey-Bass
.
Pascarella, E. T., Wolniak, G. C., Cruce, T. M., & Blaich, C. F. (
2004
).
Do liberal arts colleges really foster good practices in undergraduate education?
.
Journal of College Student Development
,
45
,
57
74
. 10.1353/csd.2004.0013
Pascarella, E. T., Wolniak, G. C., Seifert, T. A., Cruce, T. M., & Blaich, C. F. (
2005
).
Liberal arts colleges and liberal arts education: New evidence on impact (ASHE Higher Education Report, Vol. 31, No. 3)
.
Las Vegas, NV
:
Jossey-Bass
.
Percheski, C. (
2008
).
Opting out? Cohort differences in professional women’s employment rates from 1960 to 2005
.
American Sociological Review
,
73
,
497
517
. 10.1177/000312240807300307
Philippon, T., & Reshef, A. (
2012
).
Wages and human capital in the U.S. financial industry 1909–2006
.
Quarterly Journal of Economics
,
127
,
1551
1609
. 10.1093/qje/qjs030
Seifert, T. A., Jorgensen, J. D., Woniak, G. C., Pascarella, E. T., & Blaich, C. F. (
2008
).
The effects of liberal arts experiences on liberal arts outcomes
.
Research in Higher Education
,
49
,
107
125
. 10.1007/s11162-007-9070-7
Shang, Q., & Weinberg, B. (
2013
).
Opting for families: Recent trends in the fertility of highly educated women
.
Journal of Population Economics
,
26
,
5
32
. 10.1007/s00148-012-0411-2
Stone, P. (
2008
).
Opting out? Why women really quit careers and head home
.
Berkeley
:
University of California Press
.
Stone, P., & Lovejoy, M. (
2004
).
Fast-track women and the “choice” to stay home
.
Annals of the American Academy of Political and Social Science
,
596
,
62
83
. 10.1177/0002716204268552
Story, L. (
2005
,
September
20
).
Many women at elite colleges set career path to motherhood
.
New York Times,
p.
A1
.
Trunk, P. (
2005
,
April
17
).
New generation puts the focus on family
.
The Boston Globe,
p.
G1
.
Umbach, P. D., & Kuh, G. D. (
2006
).
Student experiences with diversity at liberal arts colleges: Another claim for distinctiveness
.
Journal of Higher Education
,
77
,
169
192
. 10.1353/jhe.2006.0008
U.S. News & World Report
. (
2010
).
Best colleges liberal arts colleges rankings
.
Washington, DC
:
U.S. News & World Report, LLC.
.
Vere, J. P. (
2007
).
“Having it all” no longer: Fertility, female labor supply, and the new life choices of generation X
.
Demography
,
44
,
821
828
. 10.1353/dem.2007.0035
Wallis, C. (
2004
).
The case for staying home
.
Time
,
163
,
50
56
.
59
.
Williams, J., Manvell, J., & Bornstein, S. (
2006
).
“Opt-out” or pushed out?: How the press covers work/family conflict. The untold story of why women leave the workforce
(Center for Work Life Law Working paper).
San Francisco
:
University of California, Hastings College of the Law
. Retrieved from http://www.worklifelaw.org/pubs/OptOutPushedOut.pdf