Using log‐linear models in higher education research

Abstract

The data collected in higher education research are not always quantitative or continuous. Until recently, the statistical methods available for analyzing associations among more than two categorical variables were extremely limited.

Assumptions underlying quantitative and qualitative research: Implications for institutional research

“…Institutional researchers have engaged in debates similar to those of educational researchers, but not to the same extent. Institutional researchers have primarily discussed quantitative and qualitative differences at the method level (Bohannon, 1988;Fetterman, 1991;Fincher, 1985;Hinkle, McLaughlin, andAustin, 1988, Jennings andYoung, 1988;Marshall, Lincoln, and Austin, 1991;Tierney, 1991) and how different methodologies yield different information (Peterson and Spencer, 1993). At this level, institutional researchers have been attending, correctly so, to assumptions supporting specific statistical procedures (Bohannon, 1988;Yancey, 1988aYancey, , 1988b, such as having random selection when performing multiple regression, having a sample size larger than five in each cell of an ANOVA, or assuming a normal distribution.…”