Home Articles Reader Opinion Editorial Book Reviews Discussion Writers Guide About TCRecord
transparent 13
Topics
Discussion
Announcements
 

Postnatal College Trajectories: When and Under What Circumstances Do Mothers Enroll?


by Melissa Radey - 2017

Background/Context: In light of increasingly common, non-traditional pathways to college enrollment and potential importance of post-secondary education for family wellbeing, this article examines maternal college enrollment. I employ a sociological application of rational action theory in which costs of reentry, probability of success, and utility of education influence enrollment.

Objective: Recently available longitudinal data provide the opportunity to (a) describe maternal college enrollment during the first 9 years after giving birth and (b) consider the influence of demographic (race, nativity, age), economic (poverty level, employment status, occupation) and social (informal social support, public assistance receipt, household composition, maternal health status) covariates on maternal enrollment.

Research Design: The analysis uses secondary data, the Fragile Families and Child Wellbeing Study (FFCWS), to examine how college enrollment evolves over the child’s first 9 years (n = 2,330 mothers, n = 7,808 observations).

Data Collection and Analysis: The FFCWS is a longitudinal survey of 4,898 children born in 1998–2000 to mostly unmarried, low-income mothers (the sample is comprised of 75% unmarried mothers and a 25% married comparison group). Through a stratified random sampling design, U.S. cities with populations of 200,000 or more were selected, followed by hospitals within the cities, and, finally, beds within hospitals. Using multilevel models of change, the analysis considers social and economic factors that influence college enrollment, and examines how the influences of covariates change over time.

Findings: One-third of mothers enrolled in college at least once during their child’s first 9 years and college enrollment increased as children aged. Mothers’ enrollment levels largely reflected rational action theory such that enrollment costs and probability for success influenced enrollment. Models illuminated differential rates of change. Safety net access and marital status at the child’s birth became more important in differentiating students as children aged.

Conclusions/Recommendations: Mothers’ interest coupled with the short- and long-term benefits of college enrollment suggest that mothers should receive additional support to finance their educations. Findings indicate the need for policies, including welfare policy reform, to support the growing number of student mothers enrolling in college discontinuously and their high levels of disadvantage.



The traditional trajectory of school completion followed in succession by job attainment, marriage, and parenthood, while economically beneficial, is becoming less common (Haskins & Sawhill, 2009). In particular, a growing number of people do not follow “the success sequence” and place parenthood before school completion or marriage (Sawhill, 2014). In 2012, 51% of all undergraduates and 55% of female undergraduates met criteria as independent students primarily by being older than 24 years, mothers, or married (CLASP, 2015). Approximately 62% of independent undergraduate women were mothers—one-third of all female students (U.S. Department of Education, 2014). Single parent college students, a particularly vulnerable population due their economic, child care, and time demands, are a rapidly growing student population nearly doubling from 8% in 1992 to 15% in 2012 (U.S. Department of Education, 2002; CLASP, 2015). Almost 80% of enrolled single parents are single mothers; therefore, among female undergraduates, a full 21% are single mothers (U.S. Department of Education, 2014).


The rise in maternal college enrollment, particularly among single women, is important given college’s benefits on child and family wellbeing. Post-secondary education provides one mechanism for mothers and their children to escape poverty and its consequences. In an examination of the influence of college education for single mothers using the Panel Study of Income Dynamics (PSID), Zhan and Pandey (2004) found that single mothers with at least some college were more likely to live above the federal poverty level, earned $5,063 more in annual labor income, received $157 more annual child support, and received $742 less in annual welfare income than single mothers with only a high school degree. In addition, maternal postsecondary school enrollment is associated with child gains including increased books in the home, engagement in family activities, maternal involvement in the child’s education, and educational aspirations and attainment for children (Domina & Roska, 2012; Kalmijn, 1994). Using the Early Longitudinal Study of Kindergarteners (ELS-K), Domina and Roska (2012) found that mothers who enroll in college during their children’s elementary years increased parental involvement, books in the home, and maternal involvement by .13, .28, and .37 standard deviations, respectively.


In light of increasingly common, non-traditional pathways to college enrollment and the potential importance of post-secondary education for family wellbeing, this article uses the Fragile Families and Child Wellbeing Study (FFCWS) to examine how maternal postsecondary enrollment evolves over the child’s first 9 years among a cohort of urban, disproportionately single, mothers who gave birth in 1998. Using multilevel models of change (Singer & Willett, 2003), I also identify social and economic factors that influence enrollment and examine how the influences of covariates change over time.


LITERATURE REVIEW


Rising student diversity is somewhat incongruent with conventional college persistence models. Education scholars often emphasize the importance of education without interruption (e.g., Tinto, 1987) and college persistence models focus on attributes and factors that facilitate continuous enrollment through graduation or program completion (Caspar, 2015; Tinto, 1987). Most models emphasize the importance of early recruitment such as educating middle school students on college’s financial benefits or addressing the “summer melt” where 10% to 20% of recent high school graduates with college intentions forego fall enrollment (Castleman & Page, in press).


A limited, yet growing, body of scholarship recognizes the common part-time and non-continuous patterns of college enrollment and examines the trajectories of students who juggle adult and student responsibilities (Astone, Schoen, Ensminger, & Rothert, 2000). In one of the first studies examining mothers’ return to school, Bradburn, Moen, and Dempster-McClain (1995) used a two-wave panel study of married mothers in upstate New York born in 1905–1933 (n = 294) and found that 22% of women returned to school within 30 years. In the Pathways to Adulthood Study of children born in the 1980s in Baltimore, Astone et al. (2000) found that 38% of African American women reentered school between ages 16 and 27. Although neither of these studies utilized national samples or recent data, both studies indicate the potential importance of postnatal school enrollment in families’ lives.


Applied to the current study, mothers may utilize college education after childbirth because of earlier obstacles to education and the additional employment opportunities that enrollment may provide. Increased financial need as well as modeling the value of education for their children may increase mothers’ needs and desires for college education (Sealey-Ruiz, 2013). Indeed, qualitative evidence indicates that becoming a young, single mother increases educational aspirations despite limiting enrollment short-term (Zachry, 2005).


Mothers, on average, have greater financial constraints, fewer academic role models, less academic preparation, greater employment commitments, and greater family commitments than traditional, dependent women (Goldrick-Rab & Sorensen, 2010; Institute for Women’s Policy Research, 2014). Perhaps, not surprisingly then, only 5% of 1995–1996 single mother college students obtained a bachelor’s degree in 6 years compared to 29% of nonparent undergraduates (Goldrick-Rab & Sorenson, 2010). While mothers’ lower graduation rates are well-documented, much less in known about their patterns of college enrollment postnatal (Goldrick-Rab, Carter, & Wagner, 2007). Recently-available longitudinal data provide the opportunity to (1) describe maternal college enrollment during the first 9 years after giving birth and (2) consider the influence of demographic (race, nativity, age), economic (poverty level, employment status, occupation), and social (informal social support, public assistance receipt, household composition, maternal health status) covariates on maternal enrollment.


CONCEPTUAL FRAMEWORK


The study’s conceptual framework relies upon Goldthorpe’s (1998) sociological application of rational action theory as related to educational enrollment. Under the theory, enrollment reflects both the objective and subjective realities of the costs of reentry, the probability of college success, and the utility of college education. While previous work examines rational action in decisions of continuing school (Goldthorpe, 1998) or reentering school at any level (Astone et al., 2000), the current study examines entering or, continuing, college postnatal.


First, costs may influence enrollment. In terms of the first objective, mothers’ enrollment over time, although earlier studies indicate that enrollment decreases with age (e.g., Astone et al., 2000; Bradburn et al., 1995), this pattern may not apply to new mothers. Mothers’ competing demands (e.g., child care needs, time demands of caregiving) often dissipate as children age (Kimmel & Connelly, 2007) and mothers may find college enrollment less costly in terms of time and child care costs as their children age.


In terms of covariates, individuals living in poverty achieve fewer years of schooling, on average, than their more advantaged counterparts (Duncan & Brooks-Gunn, 1997). School enrollment translates to direct costs and indirect costs in forgone earnings. Also, the number and ages of children likely influence enrollment levels. Additional children dilute available resources by requiring additional child care, time, and money. Likewise, as mentioned above, younger children are particularly demanding (e.g., Kimmel & Connelly, 2007).


College enrollment introduces vulnerability because mothers can only project its costs. Difficult course requirements or unforeseen child needs can increase the difficulty of school navigation. Student mothers often experience little schedule flexibility because they add school commitments to already hectic schedules rather than exchanging an obligation (e.g., work) for school (Center for Women Policy Studies, 2004). As such, mothers may call upon personal networks to facilitate school enrollment. Among individuals emerging into adulthood, students received more support from their parents than non-students (Swartz, Kim, Uno, Mortimer, & O’Brien, 2011) and, thus, perceived access to financial, housing, and child care assistance, or a safety net, may increase enrollment levels.


Public assistance receipt, including cash welfare and food stamps, also may be important to college enrollment, particularly among low-income mothers. Public assistance receipt may be inversely related to college enrollment rates insofar as receipt comes with constraints (e.g., work requirements) and discourages long-term planning. Some evidence indicates that recipients’ college enrollment levels were lower than non-recipients due to post-welfare reform, strict work requirements (Jacobs & Winslow, 2003). However, cash and food stamp assistance may provide necessary resources to facilitate enrollment. Likewise, public assistance receipt demonstrates mothers’ ability to navigate the government benefit system. This navigation ability would likely serve them well in securing financial aid in college.


The presence of a spouse, partner, or parent may increase enrollment through providing additional financial support to facilitate college entry. Perhaps partner availability provides a financial and child care cushion for mothers, lowering education costs and thereby increasing enrollment levels among married or cohabiting women. Alternatively, spouse or partner absence may increase enrollment because single mothers often have limited sources of income and college enrollment provides an important mechanism to achieve self-sufficiency (MacGregor, 2009). The lack of alternate partner income may increase enrollment because additional education becomes more necessary or appealing with limited household contributions and resources.


Job loss or hour reduction is another potential enrollment factor. While job loss was a cost to school enrollment in earlier birth cohorts (Taniguchi & Kaufman, 2006), employment promoted enrollment in more recent cohorts (Elman & O’Rand, 2007). The availability of online, evening, and weekend classes allows many student mothers to negotiate college enrollment while maintaining employment (Ross-Gordon, 2011). Being employed also demonstrates mothers’ willingness and ability to secure child care, a need for student mothers with young children. Likewise, stay-at-home mothers may perceive less value in additional education because they do not seek employment.


Based on limited available, sometimes inconsistent, literature, I hypothesize the following:

Hypothesis 1a: Maternal college enrollment increases as children age.

Hypothesis 1b: Mothers with more resources including less poverty, fewer children, safety nets, and public assistance enroll in college postnatal at higher rates than mothers with fewer resources.

Hypothesis 1c: Partner or parent absence increases college enrollment such that mothers living alone are more likely to enroll than those living with other adults.  

Hypothesis 1d: Employed mothers enroll in college at higher rates than those not currently employed.


Second, under Goldthorpe’s (1998) rational action theory, mothers’ probability for success influences enrollment. Mothers likely contemplate several factors when deciding whether or not they have the necessary skills and supports to succeed in school. Cognitive ability is one such skill because individuals with at least average ability have more academic success than those with lower than average ability (Jencks & Phillips, 1998). Prior college exposure also facilitates success. The college environment can be intimidating and mothers may question their ability to succeed. Prior enrollment eliminates many of the collegiate atmosphere unknowns. Similarly, family role models who obtained high levels of education may increase success perceptions (Jacob & Weiss, 2011). The enrollment process, including financial aid applications, advisement, course registration, scheduling, and campus navigation, takes initiative and skill. The process can be daunting without role models or mentors. No college connections may lower mothers’ success perceptions, also, because they do not know people like themselves in college.


Alternatively, youth may facilitate entry. Student success relies upon imposed structure and requires unique skills to navigate a schedule without clearly delineated hours to dedicate to school. Mothers who have not been enrolled in any type of schooling for a period of time may find enrollment more intimidating than their younger counterparts (e.g., Astone et al., 2000).


Therefore, I hypothesize the following:

Hypothesis 2a. Higher cognitive scores increase college enrollment rates.

Hypothesis 2b. Prior college enrollment and parents’ higher levels of education increase enrollment rates.

Hypothesis 2c. Younger mothers enroll in college at higher rates than older mothers.


Third, the utility of college education may influence enrollment rates. A college education’s utility may depend on job history. In focus groups of adult college women, Deutsch and Schmertz (2011) found that women viewed their current jobs as promoting “unsustainable pattern[s]” (p. 489) and viewed education as a rare gateway to better economic and family opportunity. Likewise, job schedules of service sector jobs are often incompatible with motherhood (Blank, Schulman, & Frohlich, 2014). With few job alternatives, mothers in service sector jobs may return to college at higher rates than mothers in other industries. Alternatively, Bradburn et al. (1995) found that job sector was not statistically significant in predicting school return.  


Therefore, based on the limited, available literature, I hypothesize the following:

Hypothesis 3: Mothers in service sector jobs will return to college at higher rates than mothers in other sectors.


METHODOLOGY


This study uses the Fragile Families and Child Well-Being Study (FFCWS), a longitudinal survey of 4,898 children born in 1998–2000 to mostly unmarried, low-income mothers (the sample is comprised of 75% unmarried mothers and a 25% married comparison group). Through a stratified random sampling design, U.S. cities with populations of 200,000 or more were selected, followed by hospitals within the cities, and, finally, beds within hospitals (see Reichman, Teitler, Garfinkel, & McLanahan, 2001, for details about the FFCWS study and sampling procedures). Although income was not a stratifying criterion, the marital criterion along with the over-selection of hospitals with close to a 3:1 unmarried to married birth ratio, resulted in a largely low-income sample. Mothers and fathers completed separate face-to-face interviews within 72 hours of giving birth, in either English or Spanish. They answered questions about their relationships with each other, their families, living arrangements, neighborhoods, employment and earnings, and psychological and physical health. Respondents completed follow-up telephone questionnaires 1, 3, 5, and 9 years later. Mothers’ response rates were 86% for the baseline survey and 90%, 88%, 87%, and 76% for the follow-ups, respectively (see CRCW, 2008). This study uses the baseline data to measure background characteristics and data from the 1-, 3-, 5-and 9-year follow-ups to model school enrollment.


SAMPLE


I restricted the sample to mothers who did not have a bachelor’s degree at the time of their child’s birth (n = 4,374) and those who participated in the Year 1 follow-up or a later wave (n = 4,201). Because covariates were measured at Year 1 (i.e., age at first birth) and Year 3 (i.e., cognitive score), I also restricted the sample to mothers who completed those waves (n = 3,558). I excluded mothers who did not report at least a high school diploma or GED by Year 9 (n = 1,029). Due to the interest in college enrollment among parenting mothers, I excluded mothers who did not report living with the focal child at least once in the follow-up period (n = 3) and those who reported an illogical sequence of educational attainment (e.g., reported having only a high school diploma after reporting some college at an earlier wave) (n = 26). After these restrictions, I employed listwise deletion for each wave when respondents did not contribute any data due to non-participation in a wave or missing values (n = 170).


The sampling restrictions necessary to examine college enrollment mean that the analytic sample is not representative of all mothers in the sample. Table 1 displays Year 1 descriptive statistics for the analytic sample as well as for the groups excluded from analysis. As expected, mothers without high school diplomas significantly differed from those with college degrees. Mothers missing data at Year 1, however, were remarkably similar to the analytic sample except that those excluded had significantly higher levels of education, lower cognitive scores, and lower levels of employment. Findings and implications must be considered in the context that the analytic subsample is not nationally representative of all mothers.


[39_21801.htm_g/00001.jpg]


+ Highest level of enrollment

Letters indicate difference in predictor and educational level or response outcome at p < .05 or better: a Sample-No High School diploma difference; b Sample-College graduate difference; c Sample-Missing difference


MEASURES


Dependent Variable


The dependent variable measures college enrollment and is based on two survey items. First, mothers indicated whether they were currently attending either a community college or a 4-year college. If not, mothers indicated if they had attended or graduated from such a college since the previous interview. Mothers who were attending or had attended college since the last interview were coded 1, 0 otherwise. Although community college and 4-year colleges are distinct, over two-thirds of female students entering community colleges aspire to baccalaureate degrees (Wang, 2013) and both community college and 4-year college enrollment increase students’ earning potential (Jepsen, Troske, & Coomes, 2014). Likewise, similar academic environments with writing, reading, and math requirements distinguish them from other vocational or technical schools.


Time Constant Covariates


Mothers who were married at the time of their child’s birth were coded 1, 0 otherwise. To consider parity, mothers specified if the focal child was their first child and a variable distinguishes firstborn births from higher order births. To measure socioeconomic background, maternal educational attainment measures whether or not the mother’s mother (the child’s grandmother) had at least a high school diploma.


To consider prior college exposure, mothers who had any college education at the time of her child’s birth are distinguished from those without such exposure. The analysis also measures job type. At Year 1, employed mothers provided their job type and those without outside employment provided the sector of their last job. Job categories included: professional/technical/management, administrative support, sales, service, other, or none. As a measure of intelligence, mothers completed an eight-item similarities subtest of the Weschler Adult Intelligence Scale–Revised (WAIS-R), which measures verbal and reasoning ability. Although mothers completed this scale at Year 3, the scale’s targeted concept of intelligence is relatively stable over time (Kaufman & Lichtenberger, 2006).

 

Time Varying and Cumulative Covariates


Age was measured in years at each time point as the spacing between interviews varied. Economic and household covariates were initially dummy-coded. However, rather than measuring them at each time point irrespective of previous values, I measured them cumulatively such that the 0–1 value range at Year 1 expanded to a 0–4 range at Year 9 where 0 indicates a continued absence of a predictor and 4 indicates its continued presence.


To consider economic characteristics, I measured employment status, safety net presence, poverty status, and public assistance receipt at each wave. Mothers who reported that they worked for pay during the two weeks prior to the survey are coded 1, 0 otherwise. To measure safety net presence, mothers specified whether they had access to each of the following in an emergency: $200, child care, and a place to live. Mothers who reported access to each of these items were coded as having a safety net, 0 otherwise. Poverty status was based on the federal definition of poverty computed from income and household size. Mothers who fell below 100% of the poverty level were coded as living in poverty. Mothers who reported receiving food stamps or Temporary Assistance to Needy Families (TANF) were coded as receiving public assistance, 0 otherwise.


To measure household composition, the mother completed a household roster at each wave including names, ages, and the mother’s relationships to all people living in the household. I consider whether or not the mother lived with a spouse or a cohabiting partner at each survey period, whether or not she lived with a parent, and whether or not she had given birth or adopted additional children younger than the focal child since the prior interview.


Control Variables


Following Taniguchi and Kaufman (2007) and Elman and O’Rand (2007), the analysis controls for race and ethnicity and nativity. Bennett and Xie (2003) found “Black advantage” such that net of socioeconomic position and academic background Black high school graduates enrolled in college at higher rates than their White counterparts. While less is known about Hispanic-White differences or immigrant-native differences, I include race and ethnicity and nativity in the analysis. To measure race and ethnicity, mothers self-identified as Hispanic/Latino or not and specified their race. From these items, I constructed four categories: Hispanic, non-Hispanic Black, non-Hispanic White, and non-Hispanic of another race. For nativity, mothers specified their county of birth and those born outside the United States were coded 1, 0 otherwise. Following Bradburn et al. (1995) and Astone et al. (2000), I also consider maternal health status. I measure it cumulatively distinguishing between mothers reporting fair or poor health from their counterparts reporting excellent, very good or good health at each wave.


ANALYTIC TECHNIQUES


First, I provide a statistical description of mothers at Year 1 (the first wave of data used in the analyses) and provide percentage distributions across the covariates. For time-varying covariates, I provide percentages of mothers who experienced change subsequent to Year 1.


Second, I used multivariate models of change in college enrollment, or the multilevel model of change (Singer & Willett, 2003), a technique allowing us to examine both structural (i.e., fixed) and stochastic (i.e., random) effects simultaneously, using melogit in Stata 13. I measured (1) initial enrollment differences among mothers measured at each data point and (2) differential rates of change. To conduct the analyses I converted each record to a person-year format (i.e., wide to long) in order to measure the time-varying outcome (i.e., college enrollment) and the fixed and time-varying covariates. The model estimates change trajectories among all eligible sample members and considers the influence of right-censoring (i.e., early exits from sample). Therefore, the sample size differs at each wave because mothers obtain their high school diplomas and gain sample eligibility, mothers obtain their 4-year college diplomas and lose sample eligibility, or eligible mothers complete or fail to complete a study wave.


To measure the passage of time between waves, century months were utilized, a demographic convention based on months elapsed since January 1900. Through the use of century months, the analysis considers the actual number of months elapsed since the child’s birth at each point of data collection accommodating varying interview schedules among mothers in the FFCWS study and avoiding the use of hypothetical timepoints (e.g., the average time distance between waves).


The multilevel model of change includes equations at two levels. The Level-1 model measures within mother change (e.g., how does college enrollment change for mothers over their child’s first 9 years?) and provides the opportunity to describe how mothers’ enrollment patterns change as their children age. The Level-2 model measures between-mother differences in modeling change (e.g., how do measured covariates influence which mothers enroll in college?) and provides the opportunity to examine the shape of enrollment trajectories given the individual’s covariate values and the interrelationships among the covariate values. The Level-2 equations consider (a) how covariates influence mothers’ college enrollment in the year after giving birth, and (b) how these covariates influence the rate at which their enrollment patterns change during the subsequent 9 years. Descriptive statistics, primarily relying on Year 1 variables, are weighted using Year 1 weights (FFCWS, 2008). Multivariate analyses are unweighted because longitudinal weights are currently unavailable (FFCWS, 2008). The analyses include whether or not mothers were married at baseline, the main sample stratification criterion. For ease of interpretation, results are presented as odds ratio form. Odds ratios are the antilogs (e.g., exponentiated values) of the model coefficients. Odds ratios greater than one indicate the percentage increase in the odds of attending college associated with a one-unit change in the independent variable. Likewise, odds ratios lower than one indicate the percentage decrease in the relative probability associated with a one-unit increase in the independent variable.

RESULTS


DESCRIPTIVE FINDINGS


The left column in Table 1 provides a descriptive snapshot of the national sample of mothers with at least a high school diploma 1 year after giving birth. Examining the outcome variable, college enrollment first, 8% of mothers were enrolled in community college or a 4-year institution within 1 year of giving birth. Students were evenly divided between the two types of institutions. While not illustrated in table form, additional descriptive analyses revealed enrollment increased gradually over time (11% at Y3, 16% at Y5, 18% at Y9). Students remained equally divided among 2- and 4-year institutions at Years 3 and 5 while 4-year institution enrollment was more prevalent at Year 9 (14% vs. 4%).


The next variables are time-fixed covariates and illuminate mothers’ background. Due to the oversampling of unmarried births, 46% of mothers were married at Baseline and 35% were giving birth for the first time. In terms of race and ethnicity, 42% were non-Hispanic, Black, 26% were Hispanic, 27% were non-Hispanic, White, and 6% were of another race. Largely among the Hispanic population, 22% of mothers were born outside of the United States.


Our selection of mothers with at least a high school diploma translated to a sample of mothers with higher levels of human capital than the total sample. Approximately 67% of mothers had mothers who had earned at least their high school diploma and almost 40% had some prior college exposure themselves. Most mothers were employed in sales or service sector jobs; only 17% reported being in professional, technical, or managerial positions at Year 1 or in their most recent positions.


The Year 1 measurement of the time-varying covariates shows that mothers of 1 year olds had an average age of 28. Almost 55% of mothers were employed, 73% were living with a spouse or partner, and 72% had emergency access to $200, housing, and childcare. However, almost one-third of mothers lived in poverty and 25% received public assistance. In terms of health, 10% reported fair or poor health.


Table 2 shows mothers’ strengths and vulnerabilities as their children age. First, 30% of mothers entered college at least once in their child’s first 9 years. Of mothers who returned and remained in the study, 54% returned at only one survey point, 32% returned at two or more consecutive time points, and 14% returned inconsistently, more than once at non-consecutive time points. With these high enrollment levels, although graduation rate is not the current analytic focus, 234 mothers (approximately 10% of the sample) graduated with a 4-year degree (results available upon request).


The volatility of the time-varying covariates on Table 2 demonstrates how mothers’ lives may interfere with college graduation. Approximately 58% of mothers entered or exited employment, 48% entered or exited poverty, 36% entered or exited public assistance, and 42% gained or lost a cohabiting spouse or partner. When changes occurred, net gain was less common than instability or loss. On average, one half of mothers who experienced a status change in an area experienced more than one change in that area over their child’s first 9 years as the “Percent Inconsistent” column indicates. However, in terms of employment and partners, mothers fared better as their children aged as 36% of mothers with an employment change gained and maintained employment while only 13% lost one without replacement. Almost 30% gained and maintained a partner, compared to only 18% who lost a partner without replacement.


[39_21801.htm_g/00003.jpg]


MIXED-EFFECT LOGISTIC REGRESSION ESTIMATES


Table 3 provides the results of the mixed-effect logistic regression models of maternal college enrollment. Model A is the “unconditional means” model (Singer & Willett, 2003) used to determine whether there is sufficient variation in mothers’ college enrollment (e.g., do enough mothers enroll in college?) to warrant further analysis. The between-person constant in the fixed effects model estimates the average log odds of enrolling in college across mothers and over time. The constant indicates that the average probability of returning to college over the 8 years of data collection is .072 calculated from the coefficient of the between-mother constant -3.330 (P = exp[-3.330]/1+exp[-3.330]). The constant’s statistical significance indicates that the probability of enrollment varies across mothers and over time.


[39_21801.htm_g/00005.jpg]


The residual intraclass correlation coefficient of latent responses, rho, is the average correlation between any two residuals (e.g., between Y1 and Y3) and it measures between-subject heterogeneity through approximating the within- and between-person variance components (Singer & Willett, 2003). Rho in Model A is 0.59, indicating that over one-half of the total variance of college enrollment exists between mothers, rather than changes over time.


Model B, an unconditional growth model (Singer & Willett, 2003) introduces time, measured in annualized century months, to obtain the rate of change in college enrollment, and a random time component. The statistical significance of Model B’s between-mother constant indicates that the log odds of college enrollment are not equal across individuals at Year 1. Likewise, the within-mother constant in Model B is positive and statistically significant indicating that the average probability of change increases over time. In other words, mothers are more likely to enter (and exit) college as their children age. In Model B, rho decreases to 0.028 from 0.594 in Model A. This reduction reflects that the within-mother variance dominates the between-mother variance. In other words, differences between mothers influence college enrollment greater than the passage of time.


Findings in Models A and B confirm that mothers’ college enrollment rates change as their children age and indicate that between-mother differences largely account for this change. Following, Model C builds upon Model B by including demographic, economic, and household covariates. The statistically significant decreases in the deviance and AIC statistics indicate that the inclusion of the covariates provides a better fit to the data. The positive within-mother constant now reflects the average monthly rate of change in college enrollment net of all covariates and indicates that mothers are more likely to enroll in college as their children age.


Model C’s significant coefficients indicate that demographic and economic characteristics influence which mothers enrolled in college. As expected, aging and giving birth after the focal child decreased the odds of college enrollment. With each 1-year increase of age, mothers’ odds of college enrollment decreased by 13% (expb = 0.87) while having an additional child decreased odds by 23% (expb = 0.77). Educational experience, through beginning college before the focal child’s birth or through having a mother who has at least a high school diploma, increased the odds of college enrollment. Mothers with some college before the focal child’s birth had twice the odds of enrollment compared to those with only a high school diploma. Time-varying characteristics also influenced enrollment. Being employed increased the odds of enrollment by 22% (expb = 1.22) while living in poverty decreased the odds by 26% (expb = 0.74).


Model D introduces covariate-time interactions to provide insight into mothers’ college enrollment. While Model C covariates address between-mother change, the interactions address within-mother change. We interacted each covariate with time, or months elapsed since the focal child’s birth, to examine which mothers enter or exit college over time. Model D includes the interactions that present the best model fit for the data as indicated by deviance and AIC statistics.


Figure 1 illustrates the nature of the significant interaction effects of being married at the time of giving birth and safety net access. Each panel shows the change over time in the predicted probability of college enrollment associated with the specified covariate keeping the remaining covariates at their mean values. Panel A compares the college trajectories of mothers who were married at the time of the focal child’s birth with mothers who were not married at the time of the child’s birth. The graph depicts that while baseline marital status does little to distinguish between students and non-students at Year 1, the disparity grows over time. At the end of Year 9, unmarried mothers at baseline were approximately 10 percentage points more likely to enroll in college compared to their counterparts who were married at baseline. Although the likelihood of college enrollment rose for both groups over time, it rose more steeply for unmarried mothers. Being married at the time of giving birth suppresses college enrollment more so among mothers of elementary-school children than for mothers of 1 year olds.


Figure 1. Predicted Probability of College Enrollment, by Specified Covariates: Fragile Families and Child Well-Being Study, Year 1 to Year 9

[39_21801.htm_g/00006.jpg]



Panel B compares the college trajectories of mothers who had continuous access to safety nets (e.g., $200, place to live, and childcare in an emergency) over the 9-year period to those who did not report access at any time period. Similar to the trajectories of those married and unmarried at the time of the child’s birth, safety net status does little to influence college enrollment at Year 1 and college enrollment rates rise as the focal children age among mothers both with and without safety nets. The influence of safety net access grows dramatically over time, however. While mothers with safety net access have a 5-point college enrollment advantage at Year 1, this advantage increases to over a 20-point advantage by the end of the survey period. Safety net presence is more important in college enrollment for mothers of elementary-school children than for mothers of younger children. Stated differently, safety net absence becomes a steeper barrier to college enrollment as children age.

 

DISCUSSION


Placing motherhood before college completion is becoming commonplace with one third of female undergraduates combining parent and student roles (U.S. Department of Education, 2014). The increase in student mothers is encouraging given the extensive evidence that extolls the benefits of college enrollment for maternal and child well-being (Domina & Roska, 2012; Zhan & Pandey, 2004). While prior work documents increasing college enrollment rates among mothers, this study used the FFCWS to examine the less-understood college enrollment trajectories among mothers during their child’s first 9 years. I also considered how demographic, social, and economic factors influence enrollment and how the influences of these covariates change over time.


College enrollment was common among this sample of urban, college-ready, predominantly single mothers; approximately 30% of mothers enrolled in college at least once during their child’s first 9 years. Mothers’ average age of 28 indicates that neither age nor childbirth impede college enrollment. Although this study’s focus on maternal enrollment during the first 9 years postpartum is unique, the return rate was somewhat higher than the 24% return rate Elman and O’Rand (2007) found among 24 to 55 year-old women in the National Families and Households Survey from 1987–1988 to 1992–1994. The higher rate of return among FFCWS mothers may reflect cohort differences as well as single mothers’ need to increase their job opportunities for self-sufficiency through education. After considering postnatal enrollment, post-hoc analyses indicate that 59% of FFCWS mothers had some postsecondary education, comparable to the 61% of women in the general population aged 25 years with some postsecondary education (U.S. Census Bureau, 2014).


Following Hypothesis 1a, college enrollment increased as children aged: 8% of mothers enrolled in college within 1 year postnatal and 18% enrolled once focal children reached elementary school age after the Year 5 interview. While 54% of student mothers enrolled at only one time point, 32% enrolled at two or more consecutive time points (e.g., Y1 and Y3) and 14% enrolled at two or more non-consecutive time points. High enrollment levels despite high levels of instability in employment, partner coresidence, poverty, and public assistance receipt indicates that, similar to more advantaged mothers in previous decades (Bradburn et al., 1995) and Black mothers (Astone et al., 2000), college enrollment is feasible, albeit difficult, for mothers.


Multivariate results indicate that mothers’ college enrollment largely reflects a sociological application of rational action theory. Net of other covariates, practical resources promoted enrollment (Hypothesis 1b) such that those not living in poverty, those without subsequent births, and those with access to safety nets (i.e., emergency cash, place to live, and childcare) enrolled in college at higher rates than their counterparts. However, the presence of a cohabiting partner—a resource on the surface—decreased enrollment (Hypothesis 1c), perhaps due to single mothers’ greater need for postsecondary education due to fewer potential mechanisms for self-sufficiency (e.g., MacGregor, 2009). Employment status and public assistance receipt did not reach statistical significance in the final model indicating that net of other economic covariates including poverty level, employment status and public assistance receipt did not differentiate between those who enrolled in college and those who did not.


Congruent with a sociological application of rational action theory and Astone et al.’s (2000) findings, mothers’ school enrollment reflected their probability of success (Hypotheses 2a–c). Younger mothers, those with higher cognitive scores, those whose mothers had graduated from high school, and, particularly, those with prior college exposure enrolled in college at higher rates than their peers. The odds ratio of prior college enrollment (7.81) is particularly noteworthy given college persistence models that emphasize the importance of continuous enrollment for college completion. The traditional launching trajectory where college graduation precedes motherhood is not realistic; however, motherhood does not eliminate college opportunities. Given the responsibilities and complexities of motherhood demonstrated in part by the great instability of this study’s analytic time-varying covariates, 6-year graduation rates often used to measure college success may not be a strong measure of success to students with competing family and employment demands.


In terms of college utility, Hypothesis 3 (service sector employment increases college enrollment rates) was not supported. In fact, results indicate that white-collar employment (professional, technical, managerial, or administrative support) increased enrollment over service sector jobs. While little work examines the influence of job sector and college enrollment among mothers, mothers in service sector jobs may not have the basic necessities for enrollment available to white-collar workers, such as higher wages, schedule flexibility, and access to tuition reimbursement programs.


Multilevel models of change also illuminated differential rates of change in college enrollment contingent upon marital status at the time of giving birth and safety net presence. Single mothers’ rate of enrollment grew differentially when compared to their counterparts who were married at the time of giving birth. While single and married mothers enrolled in college at similar rates at Year 1, single mothers’ enrollment grew more rapidly. Perhaps, mothers who gave birth as single parents prioritize their long-term desires for college education once children enter school (SmithBattle, 2007).


Access to friends and outside family support also increases in importance over time. At Year 1, the predicted probability of college enrollment was 0.06 for mothers with private safety nets versus .03 for mothers without nets. At the end of the survey period, the probability of enrollment among those with private safety nets grew to 0.38 compared to only 0.12 for those without such nets. Safety net presence is more important as time passes perhaps because as college enrollment becomes more realistic without the high demands of infants and toddlers, safety nets provide the necessary safeguard for mothers to risk college enrollment.


LIMITATIONS


Before considering implications, findings must be considered in the context of their limitations. First, while FFCWS data provided the opportunity to examine the influence of a breadth of time-varying social and economic covariates, the measures of educational attributes, including the measurement of college enrollment, are admittedly crude. I measured current enrollment or enrollment since the prior wave, but I did not have enrollment dates and datapoints are not equally spaced (i.e., opportunity to enroll varies between waves). For model convergence, I collapsed students at 2- and 4-year institutions. While community college and 4-year students have similar graduation aspirations (Wang, 2013) and enrollment at either promotes future earnings potential (Jepsen et al., 2014), 2-year and 4-year students are distinct. Likewise, data did not allow mean to include variables to contextualize their educational experiences such as financial aid receipt, number of hours enrolled, or educational goals. Second, although FFCWS data can be representative of unmarried mothers giving birth in large cities in 1998, the analytic sample of unmarried and married mothers with only high school degrees is not; rather, the sample is a national, urban sample of mothers. Third, while mixed-effects models control for unspecified, stable characteristics, analyses are not exempt from omitted variable bias such that unmeasured, time varying characteristics may account for mothers’ college enrollment. Educational paths of their peers or their children’s special health needs, for example, may influence enrollment decisions.


IMPLICATIONS


The relatively high rates of college enrollment among college-ready mothers, particularly among those with prior college enrollment, indicate that following the success sequence (college graduation, marriage, children) is only one of several paths. Although maternal age was inversely related with enrollment, child age was positively related with enrollment. The analyses indicate that college enrollment increases as children age, particularly once children reach school age. Safety net presence and spousal absence become more influential in differentiating student mothers as their children age.


These findings lead to two central policy implications. First, although the proportion of students with external demands such as employment, children, or spouses, is increasing, college financing is often contingent upon traditional paths. Scholarships often require that students enroll in college immediately following high school graduation. Florida’s Bright Future’s awards, for example, require that students enroll in college within 2 years and funding expires 5 years post high school graduation (Florida Department of Education, 2015). Yet, mothers’ older ages and low graduation rate in the FFCWS (~=10%) indicate that their enrollment needs may not fit those of their non-parent counterparts. Student mothers should not be penalized if they need to withdraw temporarily from student life in order to meet family or employment demands. Instead, policy needs to recognize adult learners’ needs and adapt educational programs to encourage student mother success (Ross-Gordon, 2011).


Second, the influence of poverty status, safety nets, and occupation sector indicate that many student mothers require their basic needs met before they can enroll in college. Currently, public assistance in many states is not designed for college enrollment. Given the merits of enrollment for mothers, children, and families, including financial benefits, we should reconsider investing in vulnerable mothers’ college educations. If college enrollment could fulfill welfare’s work requirements—for more than 1 year as is currently allowed by many states and locales—mothers could utilize the financial safety net to launch into permanent self-sufficiency for their families.


FUTURE DIRECTIONS AND CONCLUSION


Presented analyses suggest that enrollment increases as children age and that safety net presence and spousal absence become more important in differentiating students from non-student. The data, however, do not explain why these relationships exist. It could be, as I suggest, that school age children provide mothers additional time not available with infants and toddlers. An important direction for future research is to deepen our understanding of why mothers have the enrollment patterns they do. Do they seek to fulfill self-sufficiency, positive role modeling, intellectual curiosity? How do rationales vary by social and economic covariates? In addition to contextualizing mothers’ decision-making processes, future work can provide a deeper description of enrollment patterns and outcomes. Do mothers who enroll in school for short, full-time spells graduate at higher rates than continuous part-time students?


A deeper understanding of maternal college enrollment patterns and mechanisms for graduation success can inform future welfare policy. While welfare policy prioritizes immediate employment over educational attainment, the high college enrollment rates despite high levels of instability among this sample of mothers indicate great interest and dedication to education. Mothers’ interest coupled with the short- and long-term benefits of college enrollment (Domina & Roska, 2012; Zhan & Pandey, 2004) suggest that mothers should receive additional support to finance their educations. Rather than chide mothers for their low 6-year graduation rates, policy makers need to recognize that the typical student with employment, children, or spouses should be supported and congratulated on their commitment to higher education.


References


Astone, N. M., Schoen, R., Ensminger, M., & Rothert, K. (2000). School reentry in early adulthood: The case of inner-city african americans. Sociology of Education, 73(3), 133–154. doi:http://dx.doi.org/10.2307/2673213


Bendheim-Thoman Center for Research on Child Wellbeing (CRCW) (2008). Introduction to the Fragile Families Public Use Data Baseline, One-Year, Three-Year, and Five-Year Core Telephone Data. Princeton, NJ: Author.


Bennett, P. R., & Xie, Y. (2003). Revisiting racial differences in college attendance: The role of historically Black colleges and universities. American Sociological Review, 68, 567–580.


Blank, H., Schulman, K., & Frohlich, L. (2014). Nearly one in five working mothers of very young children work in low-wage jobs. Washington, DC: National Women’s Law Center.


Bradburn, E. M., Moen, P., & Dempster-McClain, D. (1995). Women's return to school following the transition to motherhood. Social Forces, 73(4), 1517–1551.


Caspar, E. (2015). A path to college completion for disadvantaged students. Focus, 31(2), 24–29.


Castleman, B. L., & Page, L. C. (in press). Summer nudging: Can personalized text messages and peer mentor outreach increase college going among low-income high school graduates? Journal of Economic Behavior & Organization.


Center for Law and Social Policy (CLASP) (2015). Yesterday’s non-traditional student is today’s traditional student. Washington DC: CLASP Center for Postsecondary and Economic Success.


Center for Women Policy Studies (2004). A profile of low income women students in postsecondary educational institutions. Prepared for the National Conference of State Legislatures Annual Meeting. Washington, D.C.


Deutsch, N. L., & Schmertz, B. (2011). “Starting from ground zero:” Constraints and experiences of adult women returning to college. The Review of Higher Education, 34, 477–504.


Domina, T., & Roksa, J. (2012). Should mom go back to school? Post-natal educational attainment and parenting practices. Social Science Research, 41(3), 695–708. doi:http://dx.doi.org/10.1016/j.ssresearch.2011.12.002


Duncan, G. J., & Brooks-Gunn, J. (1997). Consequences of growing up poor. New York: Russell Sage Foundation.


Elman, C., & O'Rand, A. M. (2007). The effects of social origins, life events, and institutional sorting on adults' school transitions. Social Science Research, 36(3), 1276–1299. doi:http://dx.doi.org/10.1016/j.ssresearch.2006.11.001


Florida Department of Education (2015). Bright futures student handbook: Initial eligibility requirements. Tallahassee, FL: Florida Department of Education, Office of Student Financial Assistance.


Fragile Families and Child Well-Being Study (FFCWS) (2008). A brief guide to using the mother, father, and couple replicate weights for core telephone surveys waves 1-4. Retrieved from http://www.fragilefamilies.princeton.edu

 

Goldrick-Rab, S., Carter, D. F., & Wagner, R. W. (2007). What higher education has to say about the transition to college. Teachers College Record, 109, 2444–2481.


Goldrick-Rab, S., & Sorensen, K. (2010). Unmarried parents in college. The Future of Children, 20(2), 179–203. Retrieved from http://search.proquest.com/docview/862787217?accountid=4840


Goldthorpe, J. H. (1998). Rational action theory for sociology. The British Journal of Sociology, 49, 167-192.


Haskins, R., & Sawhill, I.V. (2009). Creating an opportunity society. Washington, DC: Brookings Institution Press.


Institute for Women’s Policy Research (IWPR) (2014). IWPR analysis of data from the U.S. Department of Education, National Center for Education Statistics. 2003-04 Beginning Postsecondary Students Longitudinal Study, Second Follow-up (BPS:04/09).


Jacob, M., & Weiss, F. (2011). Class origin and young adults’ re-enrollment. Research on Social Stratification and Mobility, 29, 415–426.


Jacobs, J. A., & Winslow, S. (2003). Welfare reform and enrollment in postsecondary education. American Academy of Political and Social Science, 586, 194–217.


Jencks, C., & Phillips, M. (Eds.). (1998). The Black-White test score gap. Brookings Institution Press.


Jepsen, C., Troske, K., & Coomes, P., (2014), The labor-market returns to community college degrees, diplomas, and certificates. Journal of Labor Economics, 32, 95–121.


Kalmijn, M. (1994). Mother's occupational status and children's schooling. American Sociological Review, 59(2), 257–275. Retrieved from http://search.proquest.com/docview/618492581?accountid=4840


Kaufman, A. S., & Lichtenberger, E. O. (2006). Assessing adolescent and adult intelligence. Hoboken, NJ: Wiley.


Kimmel, J., & Connelly, R. (2007). Mothers' time choices: Caregiving, leisure, home production, and paid work. The Journal of Human Resources, 42, 643–681.


MacGregor, C. A. (2009). Education delayed: Family structure and postnatal educational attainment. Working Paper 09-07-FF. Princeton: Center for Research on Child Wellbeing.


Reichman, N. E., Teitler, J. O., Garfinkel, I., & McLanahan, S. S. (2001). Fragile families: Sample and design. Children and Youth Services Review, 23(4), 303–326.


Ross-Gordon, J. M. (2011). Research on adult learners: Supporting the needs of a student population that is no longer nontraditional. Association of American Colleges and Universities, 29.


Sawhill, I. V. (2014). Generation unbound: Drifting into sex and parenthood without marriage. Washington, DC: Brookings Institution Press.


Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York: Oxford University Press.


Sealey-Ruiz, Y. (2013). Learning to resist: Educational counter-narratives of Black college reentry mothers. Teachers College Record, 115, 1–31.


SmithBattle, L. (2007). ‘I wanna have a good future:’ Teen mothers’ rise in educational aspirations, competing demands, and limited school support. Youth & Society, 38, 348–71.


Swartz, T. T., Kim, M., Uno, M., Mortimer, J., & O'Brien, K. B. (2011). Safety nets and scaffolds: Parental support in the transition to adulthood. Journal of Marriage and Family, 73(2), 414–429.


Taniguchi, H., & Kaufman, G. (2006). Belated entry: Gender differences and similarities in the pattern of nontraditional college enrollment. Social Science Research, 36, 550–568.


Tinto, V. (1987). Leaving college: Rethinking the causes and cures of student attrition. Chicago: University of Chicago Press.


U.S. Census Bureau (2014). Educational Attainment (Table 2). Retrieved from http://www.census.gov/hhes/socdemo/education/data/cps/2014/tables.html


U.S. Department of Education (2002). The Condition of education 2002. National Center for Education Statistics (NCES) 2002–025, Washington, DC: U.S. Government Printing Office.


U.S. Department of Education (2014). Profile of undergraduate students. 2011-12 National Postsecondary Student Aid Study (NPSAS: 12) (Tables 3.4; 240). Washington, DC: National Center for Education Statistics.


Wang, X. (2013). Baccalaureate expectations of community college students: Socio-demographic, motivational, and contextual influences. Teachers College Record, 115, 1–39.


Zachry, E. M. (2005). Getting my education: Teen mothers’ experiences in school before and after motherhood. Teachers College Record, 107, 2566–2598.


Zhan, M., & Pandey, S. (2004). Postsecondary education and economic well-being of single mothers and fathers. Journal of Marriage and the Family, 66, 661–673.





Cite This Article as: Teachers College Record Volume 119 Number 5, 2017, p. 1-28
https://www.tcrecord.org ID Number: 21801, Date Accessed: 10/18/2021 8:54:25 AM

Purchase Reprint Rights for this article or review
 
Article Tools
Related Articles

Related Discussion
 
Post a Comment | Read All

About the Author
  • Melissa Radey
    Florida State University
    E-mail Author
    MELISSA RADEY is an Associate Professor in the College of Social Work at Florida State University. Her research interests include poverty, social support, and welfare policy with a particular interest in the inequalities unmarried and low-income mothers face in terms of education, employment, and well-being. Although she is somewhat new to examining educational outcomes, a manuscript (Radey, M., & Cheatham, L. P. (2013). Do single mothers claim their share? FAFSA completion among aid-eligible female students. Journal of Diversity in Higher Education, 6, 261-275.) examines how student status (student mother, other non-traditional student, or traditional student) influences FAFSA application rates.
 
Member Center
In Print
This Month's Issue

Submit
EMAIL

Twitter

RSS