Skip Navigation
Access to Postsecondary Education for the 1992 High School Graduates

A Multivariate Analysis of Two-Year Public and Four-Year College Enrollment Rates

In examining access to postsecondary education, this report has focused primarily on the factors associated with the relatively low four-year college enrollment rates of Hispanic, black, and low-income graduating high school seniors. Two of the most important factors were found to be that these groups were less likely to be well prepared academically, and among those who were qualified to attend a four-year college, low-income and Hispanic (but not black) students were less likely to take the college entrance examinations and apply for admission, which are the two steps required for acceptance to and enrollment in a four-year college or university. The analysis has therefore been focused on exploring the relationship of family income, parental education, and raceethnicity to four-year college qualification and taking the steps toward four-year college admission, and the relationship of these two intervening variables to enrollment patterns.

To estimate the effect of a single one of these variables, the effects of the other related variables must be controlled, which requires a multivariate analysis. Two weighted least squares regression models were used to estimate how much each of the student background characteristics variables affected postsecondary enrollment, while controlling for college qualification and taking both steps toward four-year college admission, as well as the other background characteristics.

One model examines community college enrollments; the dependent variable is the proportion of 1992 high school graduates who first enrolled in a two-year public institution. A second model focuses on four-year college and university enrollments; the dependent variable is the proportion of 1992 high school graduates who first enrolled in a either a private or a public four-year institution.\43\  In each model the dependent variable is the dichotomous enrollment variable for each subject. The independent variables in both models include whether the student was college qualified (yes/no) and whether the student took both steps toward attending a four-year institution (yes/no) as well as the background characteristics displayed throughout this report: raceethnicity, family income, and parental education.\44\  The regression coefficients in the model were then used to adjust the original estimates of proportions of students enrolled in public two-year or any four-year institution (see appendix B for the details of the methodology). These adjusted estimates, which take into account the joint effects of all the independent variables in the model, may then be compared to the original unadjusted proportion of all seniors who enrolled in a two-year public (26 percent) or any four-year institution (45 percent) as displayed earlier in table 2.

The results in the first column of table 34 show the original (unadjusted) estimates of the proportion of students who enrolled in a two-year public institution by background characteristics (the original estimates are in table 2 and table 16), college qualification, and steps toward four-year college admission. Column two contains the adjusted percentages after controlling for all other independent variables. Asterisks indicate that the estimates are significantly different from the reference group, which is italicized.

For example, the unadjusted percentages indicate that only Hispanics had two-year public college enrollment rates that were significantly higher than whites, who are the reference group (34 percent compared to 25 percent). Controlling for all other characteristics, however, both Asians and Hispanics had higher community college enrollment rates than whites (32 percent of both Asians and Hispanics compared with 25 percent of whites, adjusted). This change may be due to the fact that Asians who were unqualified for four-year institutions had very high two-year public enrollment rates, as was shown in table 33.

Students from low-income families had unadjusted community college enrollment rates, similar to those of students from middle-income families (25 to 28 percent, unadjusted). Controlling for all other characteristics, however, the adjusted rates indicate that low-income students were less likely to enroll in community colleges than middle-income students (22 compared with 29 percent, adjusted). The unadjusted estimates also suggest that the community college enrollment rates of students whose parents were college graduates were lower than those whose parents had less education. Controlling for all other variables, however, indicates that students whose parents are college graduates are only less likely to enroll in community colleges than those whose parents have some college education (25 compared to 29 percent, adjusted).

A major change from the unadjusted to adjusted estimates concerns the effect of college qualification. The unadjusted estimates indicate that students who are qualified to attend four-year colleges are less likely to attend public two-year colleges than those who are marginally or not qualified (23 percent compared to 30 percent, unadjusted). The regression model, however, results in adjusted estimates that reverse the relationship. Controlling for all other variables, it is the college-qualified students who are more likely to attend community colleges than those who are not qualified (29 percent compared to 21 percent, adjusted).

In summary, the regression model indicates that, after controlling for the effect of all other variables, the community college enrollment rates of Asians and Hispanics were higher than the community college enrollment rates of white seniors. It also suggests that it is the students from families with middle incomes who were the most likely to enroll in community colleges (29 percent, adjusted) when other factors are controlled. Moreover, seniors who were qualified to attend four-year colleges were more likely to enroll in public two-year colleges than those who were not academically qualified after controlling for other factors. However, about 1 out of 10 of those who had applied to four-year colleges enrolled in a community college, while over 40 percent (unadjusted and adjusted) of those who did not take both steps toward 4-year college admission enrolled in community college.

The second regression model defines four-year institution enrollment rates as the dependent variable (private and public colleges and universities combined), and the same independent variables as in the two-year public model discussed above. The unadjusted estimates (also shown in table 2 and table 16) are compared to the adjusted regression model estimates in the right panel of table 34. Since much of the analysis in this report focused on four-year college enrollment rates, which included tables showing the effect of controlling for college qualification and taking the two steps toward admission, the results of the regression model confirm most of the previous findings.\45\ 

The adjusted estimates confirm that there was no significant difference in the four-year college enrollment rates of whites compared to Asians, blacks, and Hispanics when controlling for the variation of all the other variables. A statistically significant difference (of 3 percentage points) remains between low-income and middle-income four-year college enrollment rates (42 percent compared to 45 percent, adjusted). The differences between the highest level of parental education (college graduate) and the lower levels also remained. Finally, the strong effect of taking the two steps toward four-year college admission remains in the regression model: (80 percent adjusted, 76 percent adjusted). Similarly, the effect of being college qualified also remained.

 


FOOTNOTES:

43/ See Appendix B for a discussion of regression models appropriate for categorical dependent variables.

44/ Financial aid was not included as a variable because the amounts are known only for those who enrolled.

45/ No other table, however, reports "Steps taken toward a four-year" as a row variable.


Prev Contents Next