Abstract
Consistent with the emerging trend in the educational literature to discount, or in some cases, overlook the unanticipated consequences of model selection on study findings, the intent of this inquiry is to compare and contrast three alternative approaches to utilizing Multiple Regression Analysis (MRA) methodology with longitudinal data in vocational education evaluation studies. These approaches are (1) single equation models, (2) commonality analysis, and (3) path analysis. Following the leads suggested in prior investigations, this study is explicitly designed to answer the following specific research question: Do the actual inferences drawn from the three methods of applying MRA lead to different conclusions regarding some of the effects of vocational education programs? Using each of the three alternative MRA approaches, answers for two explicit evaluation questions are sought. This consideration is to provide an empirical basis for the actual inferences and conclusions to be drawn regarding the effects of vocational education programs. The first evaluation question is: Does participation in vocational education make a difference in weekly income in dollars one year after graduation? The second evaluation question is: What are the influences of the different intervening variables on the external program outcome which is the weekly income in dollars one year after graduation? For the research question, the findings of this study indicate that all of these methods result in the same inferences. However, of the three methods used, path analysis offers the most insight about the relationships among the different variables considered in this study. In response to the first evaluation question, all three methods suggest that vocational education does not make a difference in weekly income in dollars in the year immediately following graduation from high school. The responses for the second evaluation question are based primarily on the path analysis model which suggests the following...
Mohagheghzadeh, Mohammad S. (1980). A comparison of various multiple regression methods applied to longitudinal evaluation studies in vocational education. Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -685144.