Abigail Knowles, Jon McCann, Rachel Newell, and Kyle Hurley
After examining the trends of postsecondary education, such as major choice and student debt, it is important to understand what factors may affect how likely someone is to attend college. This chapter will focus on factors that may influence a student’s decision to attend college. These factors include gender, race, age, and generation. In addition to this, a student’s decision to attend a public versus a private school will also be analyzed for both secondary and tertiary levels of education. The last section of this chapter includes an empirical analysis of how likely someone is to enroll in college. In this section, an ordinary least squares regression analysis is generated that includes variables such as gender, ethnicity, Armed Services Vocational Aptitude Battery (ASVAB) score, as well as the highest degree of education completed by both biological mother and father. Ultimately, these variables are then used to determine how likely someone is to enroll in postsecondary education.
Public vs. Private School Attendance
Historically, demand for private schools greatly correlates with income, as found in the years 1970-1980 according to Long (1988). The higher the familial income, on average the greater the probability that a child will attend private school over public, this effect similarly occurs with Caucasians. For grades K-12, White students were found to have a 5.69% greater probability of attending religious private school than non-Whites, and a 1.68% greater probability of attending secular private school (Long, 1998). Demand for private school is more impacted by income at the high school level than at the elementary level. Parents may be more prone to investing in private elementary schools because their children are younger. This is indicated by the greater effect income has on private high school demand, although Long (1988) does not offer any explanation for this occurrence. The effect of race and income on school choice also appears to decline from 1970 to 1980. Long (1988) cites a degradation in public school quality during this period, leading to both White and Non-White families, high and low income, opting for private schools over public at a greater rate by the end of the decade. Given this explanation, the decrease in segregation for both race and income in private schooling is not necessarily a sign of progress, but rather a sign of lower-income and non-White families choosing private schools at higher rates by necessity.
Di Carlo (2021) analyzes both public and private schools in New York City in the 2017-18 school year using data from the National Center for Education Statistics. To analyze segregation in primary schools, compare the proportion of students of different races that go to school with a specific race. For example, if a student is White in public school then it was found that a proportion of 0.415 of the students in that same school are also White, 0.185 are Asian, 0.065 are Black and 0.10 are Hispanic. For Black students in public schools, the proportion of Black students among their classmates was 0.531, compared to just 0.065 when looking at White students (Di Carlo, 2021). When looking at the proportion of White classmates for Black students, the proportion increases in comparison to public schools. However, this is most likely because a larger number of White students by proportion attend private schools than other races (Di Carlo, 2021). A much larger proportion of White students attend private schools in comparison to non-white students, and their classmates are also largely White in the schools they go to.
Familial income positively correlates with private school attendance from 1968 to 2013, with a decrease in middle-class private school attendance over this same period (Murnane, 2018). For high-income families, 18% of children enrolled in private schools in 1968, compared to 12% for middle income, and just 5% of low-income families. However, 16% of high-income families attended private school in 2013, and middle-income families’ attendance rate fell to just 7% (Murnane, 2018). This shift away from private schools by the middle class has also widened the 90-50 gap for private elementary schools, from 5.5% to 9.3% in 1968 and 2013 respectively (Murnane, 2018). Catholic and non-sectarian schools also accept differing student demographics at varying rates. For example, high-income families have moved towards non-sectarian private schools in the last 40 years whilst non-catholic religious schools have enrolled a higher proportion of middle-income students, growing from 3% to 4%, and fewer high-income students, falling from 11% to 10% over the same period (Murname, 2018). The proportion of White students decreased by 5% from 1959 to 2013, the proportion of attendance from Hispanic students also decreased by 10%, and the attendance of Black students increased by 2% during this same period (Murname, 2018). The author attributes these shifts to possible changes in income over this period, and the closure of many Catholic private schools which historically have had low tuitions, with many of the remaining open Catholic schools charging higher tuition fees (Murname, 2018). With the increased closure of schools that historically admit low to middle-income students at a higher rate, the inequality of private school attendance will most likely continue to grow and be reflected by the 90-50 gap increasing.
To find out the extent of racial segregation in a school district, comparing the school district’s racial makeup to the surrounding population may yield significant results. Saporito (2009) found that around 75% of the largest 22 school districts are more segregated by race than the average of zones those schools admit students from (Saporito, 2009). Why are large primary schools more segregated than corresponding school attendance zones? To study school enrollment, Saporito (2009) gathered U.S. Census data and analyzed students ages 14-18 enrolled in high school between the years 1990 and 2000. When comparing these two years, and finding remarkably similar results, this is a testament to how strong the quality of the data is and that a strong correlation between race and private school exists with significant certainty. In 2000, when less than 10% of children in a district were Black, around 8% of the White students in this community attended private school. However, when 71% to 80% of the children in a community were Black, over 40% of the White students residing in that community attended private school (Saporito, 2009). The opposite effect occurs among Black students, with the data reflecting that Black students have a lower chance of attending private school when the proportion of Black students in their community grows (Saporito, 2009). This data suggests that in 2000, White students living in majority Black communities had a remarkably higher chance of attending private school than White students who live in a majority White community. The shift in attendance according to the racial makeup of a community further reveals the racial aspects of private school attendance. Overall when considering the relevant research regarding the topic of public and private school attendance, it can be said with strong confidence that race, ethnicity, and familial income are all significant factors in private school enrollment.
College Attendance by Gender
One aspect of life that is impacted by one’s gender is the likelihood of embarking on and completing a degree in higher education. In many industrialized countries, the gender education gap recently changed direction such that the educational attainment of females now often exceeds that of males (Riphahn & Schwientek, 2015). Previous economic literature, with Jacob (2002), Reijnders (2018), Welsch & Winden (2019), and Davis & Otto (2016) being a few examples, has explored the college attendance gender gap and the subsequent reversal of the college attendance gap within the past few decades and the possible reasons for that change.
Non-cognitive skills, examples of which are the ability to follow directions, work in groups, pay attention in class, organize materials, an ability to successfully navigate the school environment, and social maturity and behavior, seem to have a significant influence on one’s ability to attend college and were found to be higher amongst women (Jacob, 2002). Low non-cognitive skills may reduce college attendance rates indirectly by decreasing the likelihood that a student graduates high school or leading to poor performance in high school, which decreases the chance of being accepted to college or receiving financial aid, and may increase the non-pecuniary costs of college, which results in decreasing the likelihood of enrollment for males (Jacob, 2002). In that same vein, it was found the contribution of women’s superior performance in high school to women’s increased college attendance combines both the effects of exogenous changes in how high schools prepare women for college and changes in high school performance induced by women’s responses to increased labor market opportunities (Cho, 2007). Overall, women in high school are now considerably better prepared for college entrance than they were two decades ago, and improvement in college-readiness could include both cognitive and non-cognitive skills (Cho, 2007).
Another crucial factor explaining women’s higher college attendance rates is their higher college premium, so women gain much more than men would gain in completing a college degree resulting in higher education being a worthy investment that many women are willing to make to ensure higher future wages (Jacob, 2002). This is a commonly cited reason for the increase in women’s college attendance, but it has been argued that women’s higher college premium has dwindled over time and has been on-par with the males’ college premium for at least a decade (Hubbard, 2011). The trend of increasing college attendance rates for females can now seem to be harder to explain, but an alternative factor that could now be outweighing the premium is that women tend to have lower nonpecuniary costs when attending college (Hubbard, 2011). A potentially higher college premium coupled with higher non-cognitive skills that allow for an individual to complete a degree with a lesser amount of stress and struggle results in many more women following through in higher education.
Changes in society’s norms have also influenced college attendance by gender, as over the years the expectation put upon women to marry has continually decreased, and this societal shift affects college attendance through three exogenous factors: a decrease in the probability of marriage, a decrease in the gender wage gap, and an increase in the college wage premium (Reijnders, 2018). As previously stated, single women have more to gain from a college degree and this is due to the fact that women earn lower wages, especially when considering the diminishing marginal utility of wealth, women have incentives to increase their lifetime earnings through higher education (Reijnders, 2018). The presence of the marriage market tends to minimize the overall benefit of education to women since they have been taught to expect to marry a wealthier spouse and focus on home production, and the decline in marriage rates remove this minimization from some women (Reijnders, 2018). Throughout the latest century, women have continually become less dependent on marriage for economic stability and the suitable alternative that affords them financial security has become higher education. Whilst a factor in women’s higher college attendance rates in recent years is their increased returns to education, there is also a ‘marriage benefit’ to attending as a college-educated individual is more likely to have a college-educated spouse and enjoy benefits from the spouse’s earnings power as well as their own (Ge, 2011). Even though both genders benefit from educational balance within the household, especially when both partners have a degree in higher education, the presence of potential benefits from finding a spouse plays a more significant role in women’s decision to attend college (Ge, 2011). Marriage plays a significant role in the decision-making process regarding whether or not to attend college, and the ideas that on one hand lower marriage rates have been increasing college attendance for women and on the other hand women are more likely to attend college in order to capture the potential ‘marriage benefit’ that results from finding a college-educated spouse seem to be able to coexist when explaining the reversal of the college gender gap.
Continuing to evaluate the gender gap reversal in colleges, a novel reasoning is that the opportunity cost of attending is rising for males due to the increase in rewards to becoming a superstar in occupations typically dominated by men, like professional sports (Rossi & Ruzzier, 2018). An increase in male earnings in those fields has a significant positive effect on the female to male ratio in colleges, it can increase the ratio by as much as 3.62 percent in European countries where individuals do face an actual choice between going to college and pursuing a superstar career (Rossi & Ruzzier, 2018). This chance at superstardom entices boys to pursue this path from a young age, but there are consequences of the superstar path gaining popularity because boys will continue to neglect their studies in favor of sports to try to reach that high level of achievement with no success, and are more likely to stay longer in the that path to the detriment of their accumulation of human capital, furthering the gender gap reversal (Rossi & Ruzzier, 2018).
Additionally, high school counselors are more likely to recommend that a female student would be successful at a highly selective college but are less likely to recommend them to pursue math in college, so whilst it is now more likely that women will attend college than it is for men, the gender gaps that reside in majors, occupations, and wages still very much exist and can factor into someone’s willingness to enter higher education (Welsch & Winden, 2019). The role and influence of high school counselors on student college decision-making is significant as students often use their counselors for college advice and act as a source of information when students are in the gathering or search stage of the decision-making process (Welsch & Winden, 2019). Furthermore, when families do not have the information students need, school counselors often become the primary source of information to help students access college and at one point, providing guidance on college was the primary role of counselors, whereas today they also focus on academic achievement (Welsch & Winden, 2019). This is important to consider as counselors are still people, and each individual has their own biases and beliefs that could impact the advice they give to students, whether it be consciously or unconsciously. The recommendation for which major a student should choose is dependent on both the gender of the counselor and the student, as a female counselor are 13.6 percent less likely to recommend math as a major to a female student compared to a male student due to societal stereotypes and female counselors seem to have a bias against male students in terms of success at a selective institution as they tend to rank outstanding male students approximately 0.42 points lower, on a 10 point scale, than an outstanding female student (Welsch & Winden, 2019). The authors offer the potential explanation that female counselors’ historical sociocultural values along with their own perception of their strengths and weaknesses end up being projected onto the next generation of female students, which ultimately leads to a pattern in their recommendations (Welsch & Winden, 2019). These recommendations come at pivotal moments in a students’ college decision-making process and can either encourage or discourage an individual’s willingness to attend college as they have the ability to instill confidence or insecurity into a student.
The likelihood of someone attending college seems to be greatly impacted by their gender, but people do not have just one identity defining them, and this can be seen with the intersection of race and gender. Davis & Otto (2016) explore the impact of academic determinants, which include GPA, the academic intensity of a student’s course load, and math and reading test scores, social determinants, which include parental expectations, teacher expectations, and peer influence, social class and family financial resources, and school poverty, on an idividuals’ likeliness of attending college and how these factors interact with an individual’s race and gender. Black male students fare worse than all other race-gender groups on virtually all predictors of college enrollment, with the strongest determinant of the Black gender gap being the lower academic performance of Black male students (Davis & Otto, 2016). In terms of college attendance, black women do better than black men, but not as well as white men, who do less well than white women, illustrating how the intersectionality of race and gender has a great influence on college attendance (Davis & Otto, 2016). This can be explained by the finding that both black men and women gain less from having good grades, positive peer influences, and having parents who are able to spend more on higher education, so even if they were given the same resources as white men and women, it still wouldn’t result in an equal amount of college enrollment from each group (Davis & Otto, 2016). In recent years, African-American women have increasingly used higher education as a means to increase their socioeconomic status more so than their male counterparts, and that can be seen as each older college-educated sibling increases college attendance by significantly more for African-American women than for men (Loury, 2004). Black female students respond positively to academic success more than black male students, as a one unit increase in test scores leads to a 2 percent increase in the odds of enrolling in a four-year college whereas a one unit increase in test scores composite leads to no significant change for black male students in the odds of enrolling in college, indicating that black women are more encouraged by positive academic performance to embark on higher education likely due to an increase in self-confidence (Davis & Otto, 2016).
The gender gap in higher education has been examined for many years by many researchers in previous economic literature and is found to be impacted by both endogenous and exogenous factors. Factors like an increase in the college premium for women and an increase in male earnings as a superstar in occupations like professional sports have influenced individuals internally when making their decision on whether or not to attend college, whereas the societal shift of a declining marriage market externally impacts individuals and their decision-making process. Overall, gender does play a significant role in an individual’s likelihood of attending college.
College Attendance by Race
How does the race of individuals affect their likeliness to attend college? There have been many factors that play a role in this relationship, such as socioeconomic status (SES), affirmative action, and tuition expenditures. Black, S. E., & Sufi, A. (2002) look into the socioeconomic effects when comparing college attendance between black individuals and white individuals. Previous research has shown that in the 1970’s and 1980’s “blacks were more likely to attend college than equivalent whites” (Black, S. E., & Sufi, A. 2002, pg. 3). However, it would appear these results are largely being driven by blacks that are at the lower end of the family background spectrum. This effect is also flipped when looking at the other end of the spectrum. Blacks who are at the higher end of the SES “are less likely to attend college than equivalent whites” (Black, S. E., & Sufi, A. 2002, pg. 3).
When considering college tuition costs it at first appeared that tuition had little effect on college attendance. However, Black, S. E., & Sufi, A. (2002) suggest that the relationship between the two is more diverse and that it should be looked at as such. Many potential factors were considered when determining the cause of this relationship, such as grants, credit constraints, affirmative action, etc. However, it was found that low-income blacks are extremely sensitive to changes in tuition cost, while middle to high-income blacks are not. Black, S. E., & Sufi, A. (2002) estimated that a 5% increase in tuition would result in a decrease of 35% in college enrollment among low-SES blacks. This relationship slowly disappeared over time until it was completely gone by the 1990’s in which blacks were then no more likely to attend college than whites at any end of the SES distribution (Black, S. E., & Sufi, A. 2002). Black, S. E., & Sufi, A. (2002) also suggest that educational policies should not only take race into consideration but also different responses based on family background as well.
Light, A., & Strayer, W. (2002) used data from the 1979 National Longitudinal Survey of Youth (NLSY79) to examine the effects of affirmative action on college attendance. Using the Federal Interagency Committee on Education (FICE) codes Light, A., & Strayer, W (2002) were able to determine who attends college. SAT scores were then used to determine the quality level of each college. Light and Strayer (2002) built estimates for both college attendance and college completion. They were able to determine that minorities are more likely to attend college than whites when excluding unobserved factors such as labor market returns to college or unmeasured ability. Light and Strayer (2002) believe this to be a result of affirmative action. The authors also found a positive coefficient of 0.176 for minorities when estimating college completion meaning minorities are more likely to complete college than their white counterparts. However, this may not actually be the case as a college’s primary objective is to accept individuals who have the highest likelihood of earning a degree (Light and Strayer 2002, pg. 41). Once the effects of the unobserved factors are considered, minorities become less likely to graduate from college than their white counterparts. This discrepancy is believed to be the result of affirmative action as from an observable standpoint it would appear that minorities have greater success with getting into college but that success is brought back down once the unobservable factors are introduced.
Affirmative action, being an often discussed political topic frequently undergoes scrutiny in the U.S. state level legislative system. Several states have banned utilizing race as a metric in college admissions, offering the opportunity to study the effects of the absence of these policies. Backes (2012) compiled data from the Integrated Postsecondary Education Data System to study both enrollment and graduation statistics. After the ban, the share of enrollment for Black students dropped 1.6% for top universities (p. 447), indicating a significant negative effect for Black enrollment. In medium selectivity institutions an insignificant increase in Black enrollment occurred after the ban, and in the bottom 70% of universities had an insignificant decrease in Black enrollment. Hispanic enrollment decreased 2.9% after the affirmative action ban in top universities, and an 8% decrease in average institutions (Backes, 2012, p. 447). For enrollment, banning affirmative action policies appears to have a negative effect on Hispanic and Black students. For graduation after the affirmative action ban an insignificant decrease in the graduation of Hispanic and Black students occurred for average institutions. Overall, there appears to be an effect of decreasing minority students in top universities for both graduation and enrollment after affirmative action bans (p. 450). Affirmative action appears to be most effective at top universities for increasing minority enrollment and graduation rates, and possibly insignificant at medium and low selectivity institutions.
College Attendance by Age
How likely is someone to attend college based on their age? Are older or younger people more likely to enroll in post-secondary education? Davis and Bauman (2013) find that 71% of students enrolled in undergraduate college were 24 years old or younger. This was true for both two-year, 64%, and four-year college enrollment, 75% (pg. 9). This report provides a guide to data from the U.S. Census Bureau on the 83 million people who were enrolled in school in 2011. The data used in this report relies heavily on the American Community Survey (ACS), which is a nationwide survey that is part of the census program, and the Current Population Survey (CPS). The data was collected from 83 million people who were enrolled in school in 2011. Information on their age, sex, where they live, progress through school, as well as the types of schools they attend was gathered. According to this report, in 2011, 20.4 million people were enrolled in college. In addition to this, “Students enrolled in graduate school were more likely to be older with 74 percent of them aged 25 and over,” (Davis & Bauman, 2013, pg. 9). Therefore, this leads to the conclusion that most students who enroll in an undergraduate program are most likely to be 24 years of age and younger. However, those who enroll in a graduate degree program are more likely to be older than 25 years of age. This is a reasonable conclusion since most students must complete an undergraduate program, in order to be eligible for a graduate program. For example, in order to apply for a master’s degree, a student must first hold a bachelor’s degree. Thus, most students in graduate school are likely to be older than those who are in undergraduate school.
Juszkiewicz (2016) focused primarily on community college enrollment. The paper is the third of a series of American Association of Community Colleges reports on the national trends observed in community college enrollments and uses data published by the U.S. Department of Education. In general, Juszkiewicz (2016) found that part-time students are more likely to be older than full-time students with the average age of full-time students enrolled in an undergraduate program being 21.8 years and the average age of a part-time student being 27.2 years. Furthermore, Juszkiewicz (2016) analyzed the completion rate of different age groups and observed that students over the age of 24 had a lower completion rate (36.6%) than those 20 or younger (40.7%), but a higher completion rate than students between the ages of 20 and 24 (25.1%). One reason as to why part-time students are more likely to be older than full-time students may be due to holding different responsibilities. For example, a young woman in her early twenties is more likely to be a full-time student than an older woman who is in her thirties because the older woman has different responsibilities, such as taking care of children. An older person may also have different financial responsibilities than a younger student. For example, a person in their thirties is more likely to not live with their parents and either rent or own a house. This makes them financially responsible. They may also have to provide for children. Therefore, an older person may only be able to attend school part-time since they must also maintain a full-time job and flow of income in order to care for themselves and their children. However, a younger person is more likely to not have as many financial responsibilities. Thus, they are able to attend school full-time. Different financial responsibilities, as well as having children, may also contribute to why students over the age of 24 have a lower completion rate than those who are 20 years old or younger. In addition to this, Juszkiewicz (2016) found that the average age of a part-time student is 27.2 years old and the average age of a full-time student is 21.8 years. This leads to the average age of a student in general being 24.5 years, which corresponds to Davis and Bauman’s (2013) finding that 71% of students enrolled in undergraduate college are 24 years old and younger.
Hanson (2021) used data from multiple sources, including the U.S. Census Bureau’s Current Population Survey (CPS) as well as the National Center for Education Statistics and reports on college enrollment and student demographic statistics. Hanson (2021) found that as of January 2021, 92.0% of college students were under the age of 24. This is interesting considering that in 2011, Davis and Bauman (2013), found that 71% of students enrolled in undergraduate college were 24 years old or younger (pg. 9). Therefore, one trend noted in the age of college students is that the enrollment of students over the age of 24 is declining, while the enrollment of students under the age of 24 is increasing. According to Hanson (2021), one possible explanation for the decrease in enrollment for older age groups may be partially due to economic improvement. Economic improvement that may deter older age groups from enrolling in college may include an increase in employment opportunities or higher paying jobs, which would improve a person’s quality of life without them having to attend college. Furthermore, this paper also breaks down college attendance data into age groups. For example, in 2021, 0.2% of college students were 55 and older, whereas 2.8% of college students were under the age of 18. This leads us to believe that when analyzing college attendance by age, the data may display a bell-curve shape with the highest enrollment being those between the ages of 20-24 years. The lowest enrollments are those under the age of 18 and over the age of 55.
This analysis also corresponds with the findings of Monaghan (2021). Monaghan (2021) examines predictors of enrollment as an individual progresses through the life-course. The data used comes from nationally representative data from the United States, National Longitudinal Survey of Youth 1979, and follows a cohort from ages 18-45. One major finding in this study is that social background and academic preparation are only weakly predictive past the age of 24. This means that for those ages 18-24, family background and high school performance are large predictors of college enrollment. However, as an individual ages, these predictors’ influence declines rapidly and in some cases are even reversed. One reason for this may be because “the influence of one’s past fades in importance as one ages,” (Monaghan, 2021, pg. 15). In addition to this, a better economy is also associated with lower odds of postsecondary enrollment for younger ages. For example, if a student who just graduated from high school is able to establish a career and earn a high-wage, then they would not attend college. However, when the economy is worse and it is more difficult for a high school graduate to either find a job or earn a high salary, then they may enroll in college in order to establish a career or earn a high salary in the future. Monaghan (2021) also found that marriage is a barrier to college enrollment, but only until the age of 25. Furthermore, after the age of 25, marital dissolution or divorce positively affects enrollment among females, but negatively affects enrollment among males (Monoghan, 2021, pg. 16). In addition to this, involuntary job loss causes those who are age 35 and older to enroll in college. This correlates to Hanson’s (2021) speculation that the state of the economy affects how likely an older person is to attend college. If there is a high unemployment rate, it is likely that unemployed people who are aged 35 and older will enroll in a college degree program. Also, with this older age group of 35-45 years, “responsibility for very young children continues to discourage enrollment, as does a faster-growing national economy,” (Monaghan, 2021, pg. 12). Therefore, the college enrollment predictors, as well as the effects of these predictors, found by Monaghan (2021), further supports the analysis that college enrollment by age group data forms a bell-shaped curve in which the highest enrollment rate is found in those age 20-24, and the lowest enrollment rates are found in those under the age of 18 and over the age of 55.
How Likely is Someone to Enroll in College?
As of January 22, 2021, there are approximately 19.6 million students enrolled in postsecondary education institutions in the United States (Hanson, 2021). This number includes both undergraduate and graduate students. Therefore, with such a high number of students enrolled in college degree programs, it is pertinent to know what causes a particular person to choose to enroll in postsecondary education. Having the knowledge of the factors that may influence a person’s decision to enroll in postsecondary education could help policymakers decide on and pass effective legislation that incentivizes college enrollment. In addition to this, being aware of the factors that influence a person’s decision to attend college could also lead to more programs and acts designed to help those who are less likely to enroll in a degree program. For example, it has been noted that older women are less likely to enroll in postsecondary education due to the responsibility of taking care of their children. Therefore, if policymakers and education institutions are aware of this influence, they can develop programs to assist mothers who wish to enroll. One example of this may be offering a stipend to mothers to use towards daycare services. In this paper, we will determine what factors are considered significant and their impact on how likely someone is to enroll in college through the use of an ordinary least squares regression analysis using data from the National Longitudinal Study of Youth, 1997 cohort. Factors such as gender, ethnicity, Armed Services Vocational Aptitude Battery (ASVAB) score, type of school attended, and highest degree of education completed by both biological mother and father will be analyzed.
Prior Research
Prior research in this area includes a number of longitudinal studies of high school students and their decisions to attend postsecondary educational institutions. These early empirical studies have established that family income, parental education, high school peer relationships, and the proximity of a college to the student’s home are all factors that may influence a student’s decision to enroll in college. However, while these studies have confirmed that these factors are important, they have not been able to describe the magnitude of importance.
Kohn (1976) attempted to quantify these variables by developing a theoretical and empirical model of student behavior that would help to forecast enrollment patterns at the postsecondary level. In addition to this, Kohn’s model would also help to “forecast the effect of proposed federal and state policies on the number and composition of enrollments and on the distribution of students among institutions,” (Kohn, 1976, pg. 391). However, Kohn’s model lacks the improved data necessary to fully estimate enrollment patterns. Kohn (1976) states that the data used in his model was unreliable of true student behavior and estimates, such as financial aid amounts, were not accurate, thus making his model limited. Today, this can be solved by using public data collected from sources such as the U.S. Bureau of Labor Statistics. In addition to this, Kohn’s (1976) model lacked currency in data. This study was performed in 1976 and since then, student’s college-going behaviors have changed. Therefore, the need for a more recent study is pertinent in order to properly forecast current enrollment patterns and policies. Another limit of Kohn’s (1976) model is that the data used was only collected from students living in North Carolina. Thus, his findings are only applicable to the state of North Carolina since students’ behaviors may differ across states. There is a need for further research in either other individual states or as a whole across the country of the United States.
In a recent study, Monaghan (2021) also examined college enrollment patterns, but did so by age. His study focused on how predictors of enrollment changed as individuals progressed through the life-course. While Monaghan (2021) did analyze factors such as race, parental education, and AFQT score, his findings mainly focus on the factors of marriage, marital dissolution, and children. Therefore, there is a demand to analyze how factors like race, parental education, and AFQT score affect a student’s decision to enroll in college.
Data
The data used in this study comes from the National Longitudinal Survey of Youth, 1997 cohort (NLSY-97). This data is funded and directed by the U.S. Bureau of Labor Statistics. The NLSY-97 consists of a sample size of 8,984 men and women born during the years of 1980 through 1984 and were living in the United States at the time of the initial survey in 1997. Between the years of 1997 to 2011, interviews were conducted annually. After 2011, interviews were conducted biennially. Overall, the NLSY-97 collects substantial data on these respondents’ educational experiences, as well as labor market behaviors. In addition to this, information on their youth, as well as family and community backgrounds has been collected.
The dependent variable used in our model is the highest degree received. This variable was created for all respondents and provides the answer of the respondents’ most recent interview. Respondents can indicate no degree, GED, high school diploma, associate’s degree, bachelor’s degree, master’s degree, PhD, or professional degree. Predictor variables include gender, location such as urban or rural, biological father’s highest grade completed, biological mother’s highest grade completed, race/ethnicity, type of school attended, mother’s age at first birth, household income, and the Armed Services Vocational Aptitude Battery (ASVAB) score percent. Indicator variables were employed for respondents’ gender, location, race/ethnicity, and type of school attended. Table 1 includes all of the variables used within the model, as well as their descriptions. The variables gender and race/ethnicity were included in the regression analysis since it has been well documented in previous research that these socioeconomic factors contribute to one’s college-going decision. In the past, males were often more likely to attend college than females. However, this has changed in recent years and now females are more likely to attend college than males. In addition to this, it has been found that caucasians are more likely to attend college than minority groups. The highest grade completed by both biological parents was included since it has also been noted as significant to one’s college decision in previous literature. For example, if both biological parents received undergraduate degrees, then their child is more likely to also receive an undergraduate degree than a child whose biological parents’ highest level of education completed was a high school diploma. The ASVAB score percent were included in our regression analysis to evaluate if level of intelligence influenced one’s college-going decision. In addition to this, location was included in order to assess if living in an urban or rural area influences college enrollment. Due to previous significant findings, the mother’s age at first birth has been included. It has been formerly noted that the younger a mother is at first birth, the less likely her children may be to attend a postsecondary education institution. Therefore, the mother’s age at first birth is an important factor to include in our analysis. Household income is also included in the regression analysis since a student from a household with higher income is more likely to attend college than a student from a lower income household. One reason for this may be because children from higher income households are able to afford more opportunities that make them more apt to attend college, like playing on sports teams or learning an instrument. Finally, the type of school attended has been included in order to assess if the type of school a student attends, like public or private, may make them more apt to enroll in college.
Table 1: Variables and their Descriptions
Variable Description
| HighestEdu | Highest level of education achieved by child |
| Gender | Gender of Individual |
| Male | Dummy for Male Gender |
| Female | Dummy for Female Gender |
| Ethnicity | Ethnicity of Individual |
| Black | Dummy for Black race/ethnicity |
| Hispanic | Dummy for Hispanic race/ethnicity |
| MixedRace | Dummy for Mixed race/ethnicity (Non-Hispanic) |
| NonBlackHispan | Dummy of Non-Black/Non-Hispanic race/ethnicity |
| Urban | Dummy for Urban |
| FatherGrade | Highest level of education completed by father |
| MotherGrade | Highest level of education completed by mother |
| AgeAt1st | Mother’s age at birth of first child |
| Income | Household income |
| SchoolType | Type of School Individual has Attended |
| PublicSchool | Dummy for Public School |
| TechnicalSchool | Dummy for Technical/Vocational School |
| CatholicSchool | Dummy for Cathoic School |
| PrivateReligion | Dummy for Private School – other religious affiliation |
| PrivateNonReligion | Dummy for Private School – no religious affiliation |
| AlternativeSchool | Dummy for Alternative School |
| OtherSchool | Dummy for Other Schooling option |
| HomeSchool | Dummy for HomeSchool |
| ASVAB | ASVAB Math and Verbal Test Score |
| logIncome | Log of Household Income |
| logASVAB | Log of ASVAB |
Summary Statistics
The sample size of this analysis is 2202 individuals. 48% are male and 52% are female. 18% are black whereas 66% are white (nonblack/hispanic). 69% of the sample were from an urban location. The average household income was $56,417 and the average age of the mothers at the birth of their first child was 23.6 years old. The highest level of education was very similar for the mothers and the fathers with both completing one year of college on average.
Table II: Summary Statistics
| Variable | N | Mean |
| Income |
2202 |
56416.93 |
| AgeAt1st | 2202 | 23.61 |
| ASVAB | 2202 | 54815.30 |
| FatherGrade | 2202 | 13.21 |
| MotherGrade | 2202 | 13.22 |
| Urban | 2202 | .6903 |
| Female | 2202 | .5163 |
| Black | 2202 | .1771 |
| Hispanic | 2202 | .1503 |
| MixedRace | 2202 | .0073 |
| TechnicalSchool | 2202 | .0091 |
| CatholicSchool | 2202 | .0454 |
| PrivateReligion | 2202 | .0145 |
| PrivateNonReligion | 2202 | .0082 |
| AlternativeSchool | 2202 | .0023 |
| OtherSchool | 2202 | .0009 |
| HomeSchool | 2202 | .0009 |
Results
An ordered probit analysis was conducted on the data and the results are as follows:
| Variable | Coefficient | P-Value |
| logIncome | 0.0894467 | 0.0018 *** |
| logASVAB | 0.591757 | 8.50e-064 *** |
| FatherGrade | 0.0618188 | 1.06e-09 *** |
| MotherGrade | 0.0306851 | 0.0042 *** |
| AgeAt1st | 0.0276353 | 1.26e-06 *** |
| Urban | -0.107384 | 0.0374 ** |
| Female | 0.34327 | 1.37e-013 *** |
| Black | 0.324327 | 2.46e-06 *** |
| Hispanic | 0.0584871 | 0,4150 |
| MixedRace | 0.325274 | 0.2301 |
| TechnicalSchool | 0.121130 | 0.6225 |
| CatholicSchool | 0.148233 | 0.1775 |
| PrivateReligion | -0.0388516 | 0.8379 |
| PrivateNonReligion | 0.189055 | 0.4520 |
| AlternativeSchool | -0.150331 | 0.7549 |
| OtherSchool | -0.514741 | 0.5072 |
| HomeSchool | -1.60679 | 0.0519 * |
It can be seen that the variables logIncome, logASVAB, FatherGrade, MotherGrade, AgeAt1st, Urban, Female, Black, and HomeSchool are all statistically significant. HomeSchool has a negative coefficient meaning that, when controlled for PublicSchool, a homeschooled individual is less likely to have a higher level of education than an individual that attended public school. This is to be expected as homeschooling generally results in the least amount of preparation for the students. When controlled for public school, private school was found not to have a statistically significant effect on maximum level of education. This indicates that private school may not have a significant benefit or drawback in maximum educational attainment for students, despite the income and racial inequities in private school enrollment (pp. 2-5). Females are more likely to achieve a higher level of education than males which is also to be expected as previous research has shown that higher educational attainment has been increasing over time for women. The model also shows that an increase in household income results in an increase in the likelihood of a higher level of education for an individual. Black individuals are also estimated to be more likely to obtain a higher education level than non-black/non-hispanic races. This is to be expected as previous research has found increases in educational attainment for blacks when compared to whites. An increase in an individual’s ASVAB test scores is also estimated to result in a larger likelihood of obtaining a higher level of education. This shows how telling the scores that individuals receive are when it comes to future education.
Conclusion
Several factors have a heavy influence on whether or not an individual will embark on a degree in higher education. Being a female has a positive effect on the highest educational achievement level over being a male. In terms of ethnicity, a Non-Black/Hispanic person is less likely to attend college than a Black person. If a student was home-schooled, then they are significantly less likely to attend college than if they were to attend a public high school. The higher the household income of an individual, the more likely that individual is to attend college and complete more years of schooling. All of these elements, and several others contribute to the likelihood that an individual will continue into higher education. Determining how factors like race, age, and gender can impact an individual’s likelihood of attending college is important because research in this area can be utilized when writing policies meant to encourage enrollment for groups who are less likely to continue their education. Education is an effective way to increase one’s human capital, with their wages and socioeconomic status following suit and can be a tool used to lessen income inequality in the United States.
References
Backes, B. (2012). Do Affirmative Action Bans Lower Minority College Enrollment and Attainment? Evidence from Statewide Bans. Journal of Human Resources, 47(2), 435–455.
Black, S. E., & Sufi, A. (2002, November 4). Who goes to college? differential enrollment by race and family background. NBER.
Bureau of Labor Statistics, U.S. Department of Labor. National Longitudinal Survey of Youth 1997 cohort, 1997-2017 (rounds 1-18). Produced and distributed by the Center for Human Resource Research (CHRR), The Ohio State University. Columbus, OH: 2019.
Cho, D. (2007). The role of high school performance in explaining women’s rising college enrollment. Economics of Education Review, 26(4), 450-462.
Davis, J., & Bauman, K. (2013). (rep.). School Enrollment in the United States: 2011 (pp. 1–14). Washington, DC: US Census Bureau.
Davis, T., & Otto, B. (2016). Juxtaposing the Black and White Gender Gap: Race and Gender Differentiation in College Enrollment Predictors. Social Science Quarterly, 97(5), 1245–1266.
Di Carlo, M., Wysienska Di Carlo, K., Fenelon, E., Flood, K., Milborn, E., & Rodriguez, C. (2021). Public and Private School Segregation in New York City. Albert Shanker Institute.
Ge, S. (2011). Women’s college decisions: how much does marriage matter?. Journal of Labor Economics, 29(4), 773-818.
Hanson, M. (2021, January 22). College enrollment statistics [2022]: Total + by demographic. Education Data Initiative.
Hubbard, W. H. (2011). The phantom gender difference in the college wage premium. Journal of Human Resources, 46(3), 568-586.
Jacob, B. A. (2002). Where the Boys Aren’t: Non-cognitive Skills, Returns to School and the Gender Gap in Higher Education. Economics of Education Review, 21(6), 589–598.
Juszkiewicz, J. (2016, March). Trends in Community College Enrollment and Completion Data, 2016, Washington, DC: American Association of Community Colleges.
Kohn, M. G., Mansk, C. F., & Mundel, D. S. (1976). An empirical investigation of factors which influence college-going behavior. In Annals of Economic and Social Measurement, Volume 5, number 4 (pp. 391-419). NBER.
Light, A., & Strayer, W. (2002). From Bakke to Hopwood: Does Race Affect College Attendance and Completion? Review of Economics and Statistics, 84(1), 34–44.
Long, J. E., & Toma, E. F. (1988). The Determinants of Private School Attendance, 1970-1980. The Review of Economics and Statistics, 70(2), 351–357.
Loury, L. D. (2004). Siblings and gender differences in African-American college attendance. Economics of Education Review, 23(3), 213-219.
Monaghan, D. B. (2021). Predictors of College Enrollment across the Life Course: Heterogeneity by Age and Gender. Education Sciences, 11(7), 344. MDPI AG.
Murnane, R. J., Reardon, S. F., Mbekeani, P. P., & Lamb, A. (2018). Who goes to private school? Long-term enrollment trends by family income. The WERA Educational Journal, 40.
Reijnders, L. S. M. (2018). Wealth, Wages, and Wedlock: Explaining the College Gender Gap Reversal. Scandinavian Journal of Economics, 120(2), 537–562.
Riphahn, R. T., & Schwientek, C. (2015). What Drives the Reversal of the Gender Education Gap? Evidence from Germany. Applied Economics, 47(52–54), 5748–5775.
Rossi, M. A., & Ruzzier, C. A. (2018). Career Choices and the Evolution of the College Gender Gap. World Bank Economic Review, 32(2), 307–333.
Saporito, S. (2009). School Choice in Black and White: Private School Enrollment among Racial Groups, 1990-2000. Peabody Journal of Education, 84(2), 172–190.
Welsch, D. M., & Winden, M. (2019). Student Gender, Counselor Gender, and College Advice. Education Economics,27(2), 112–131.