Reconciling Intent with Action: Factors Associated with the Alignment between Transfer Intent and Coursework Completion Patterns among Two-Year College Students in STEM

ABSTRACT

In this study, we explored the alignment between initial transfer intent and subsequent course-completion patterns among 1,668 first-time STEM-aspiring students at two-year institutions. Using survey and transcript data, we conducted latent profile analysis and subsequent path analysis that revealed five distinct course-completion patterns: Trailing/Time-Out, Non-transfer-Focused, Transfer: STEM Concentrated, Transfer: Humanity/Social Science Con-centrated, and Transfer: Developmental Education Concentrated. Students’ math and science self-efficacy, academic engagement, and transfer-oriented interactions moderated the alignment between initial transfer intent and later course completion patterns. Based on the findings, we discuss the importance of cultivating strong academic self-efficacy and building effective advising practices to help students actualize their transfer intent.

KEYWORDS:

• Two-year college
• community college
• STEM
• transfer intent
• course completion
• self-efficacy

Acknowledgments

The authors would like to thank the institutional researchers at the study sites, as well as Seo Young Lee and Yen Lee for research assistance.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. The analytical weight was the product term between the sampling weight and the reciprocal of response rate. The sampling weight was the reciprocal of the sample selection probabilities. Participants with a smaller chance to be sampled would receive a larger sampling weight than those with a higher chance, and vice versa. The response rate was the number of participants divided by the initial sample size within each specific stratum.

2. In statistical analyses routinely adopted in higher education research, transfer intent would be considered and modeled as an observed, instead of latent, multinomial variable. In our research context and analytical framework, however, treating transfer intent as a latent construct is a preferred approach not only due to technical settings in Mplus, but more importantly because of a substantive reason: Because our research questions focused on the evolving process of how each given level of transfer intent, including no intent to transfer, transpires into coursework, using transfer intent as an observed multinomial variable would not allow us to capture this fluid, unique process among students holding no initial intent to transfer, since this category would be used as the reference category with which other levels of transfer intent would be compared. With this conventional approach, we would have ended up with the relative profile probabilities for only those students reporting intent to transfer and no unique information for students reporting no intent to transfer (Masyn, 2013). While we might be able to force such an approach in Mplus, we would have to change the reference category twice, estimate two more path models, and more problematically, risk the inflation of type I error rate. In contrast, within the LTA analytical framework and given the appropriate technical features for LTA in Mplus, modeling transfer intent as a latent variable enabled us to chart three distinct ways describing the intent-to-coursework transition corresponding to each of the three specific levels of transfer intent. Accordingly, as part of the procedures associated with the LTA techniques, transfer intent was transformed into a latent variable to be analytically viable in Mplus. Essentially, for students expressing intent to transfer into STEM fields, the logits representing the posterior profile probabilities were fixed at 15.00, −15.00, and −15.00 for the three types of transfer intent, respectively. For students expressing intent to transfer into non-STEM fields, the logits were fixed at −15.00, 15.00, and −15.00, respectively. The logits for students reporting no intent to transfer were specified as −15.00, −15.00, and 15.00, respectively. The resulting latent variable still distinguished transfer intent as three levels, yielding nuanced findings aligned with our research questions and the measure in a substantive sense.

Additional information

Funding

This material is based upon work supported by the National Science Foundation under Grant No. DUE-1430642 awarded to the second author.