California Community College Student Outcomes Abroad Research Project

An interaction effect is typically identified as when two independent variables interact such that the effect of one of the variables differs depending on the level of the other variable. In a regression-based context it is possible to identify the existence of statistically-significant interactions. So if two variables have a positive effect on a student outcome, perhaps the interaction effect between the variables, in which one enhances the effect of the other, may lend further credence to the notion of study abroad participation significantly affecting student outcomes.

While study abroad participants are becoming increasingly diverse, most participants continue to be white females. Thus, analyses that aggregates participants may produce inaccurate expectations for the effects of study abroad for students from certain backgrounds who less well represented among study abroad students. Running analyses on the effect of study abroad for such groups as Latinos, Blacks, Asians, students with lower-SES backgrounds, and academically lower-performing students may be significant to the program completion agenda that exists at the community college level.

The duration of study abroad programs range from short-term programs to semester-length programs. In addition, different types of courses are offered abroad, which may be language and non-language experiences. These particulars may be important to identify, as such information could be useful in designing an optimal length course that balances increased effects with costs in time and resources to the student and the institutions.

With the use of regression techniques to control for differences between study abroad and non-study abroad groups lies the potential for over-extrapolation of data and overestimation of study abroad effects. Some suggest that regression is not an ideal approach for estimating treatment effects when the complete overlap of covariates across treatment and control groups is lacking, requiring extrapolation that may lead to inaccurate treatment effect estimates. A matching method, particularly propensity score matching, may produce more accurate estimated effects of study abroad.