The work-to-college transition: postsecondary expectations and enrolment for young men and women in the US labour force

Abstract

This paper investigates early employment influences on postsecondary expectations and enrolment for working men and women who have recently completed high school in the United States. We find that young workers still have very high expectations for postsecondary education, but that women are more likely to enrol. However, this difference is primarily due to high rates of female enrolment in less-than-four-year institutions. Job training from one’s employer predicts higher odds of future postsecondary enrolment, but this relationship appears only for men. Recommendations are given for assisting young workers in realising postsecondary expectations and advancing social mobility, as well as how gender must be considered in these efforts.

Keywords:

• educational expectations
• postsecondary enrolment
• gender
• job training

Notes

1. We used the ‘ice’ and ‘mim’ commands in Stata for multiple imputation and related analyses. We used default settings, which include imputing five data-sets, which are then pooled for analysis. We also used the ‘svy’ commands in Stata to account for weighting and the complex survey design utilised for NELS.

2. During 1993, the year from which we draw employment and benefits data, a significant piece of federal legislation granting employees time off for family care while protecting job security went into effect. The Family and Medical Leave Act of 1993 (Public Law 103-3), which was signed on 5 February 1993 and was enacted six months later, entitled an employee with a year of tenure with a given employer to up to 12 weeks of leave for the birth or adoption of a child, for the care of a family member with a serious health condition, or for a personal serious health condition. Full text of Public Law 103-3 can be found on the Government Printing Office website: http://www.gpo.gov/fdsys/pkg/BILLS-103hr1enr/pdf/BILLS-103hr1enr.pdf.

3. Specifically, each additional child predicted about 35% lower odds of expecting a bachelor’s degree. Being married (relative to being single) predicted about 40% lower odds of expecting a bachelor’s and being in a marriage-like relationship (relative to being single) predicted about 55% lower odds of expecting a bachelor’s. Being divorced (relative to being single, never married) did not predict a statistically significant difference in odds of expecting a bachelor’s degree.

4. We also examined our results using a Bonferroni adjustment. When we applied the adjustment, none of the interaction terms was statistically significant. However, scholars caution against using the adjustment due to trade-offs and unintended consequences and instead recommend non-adjusted results with appropriate caveats (Perneger 1998). We include our less-conservative, non-adjusted results but want to remind the reader that our findings, especially for interactions, are primarily suggestive and warrant further investigation.

5. Results from Table 4 also shows that the main effects for job training are sensitive to the level of enrolment. In other words, while job training predicts significant differences when examining enrolment vs. no enrolment, as well as four-year enrolment vs. non-four-year enrolment, it doesn’t predict any differences for the comparisons in the multinomial model. Additionally, however, the interaction of female × job training remained significant when added to the multinomial model (results not shown), adding more credence to that result.