ERIC Number: ED189155
Record Type: RIE
Publication Date: 1980-Apr
Reference Count: 0
Path Analysis with Categorical Data: Applications to Education.
Wolfle, Lee M.
An extension of the methods of path analysis to include studies of categorical data was described and exemplified in a causal study of college dropouts. The usual models and methods of causal (path) analysis were designed for the study of quantitative variables and are not appropriate when the variables under investigation are categorical. Extensions of the loglinear analysis of contingency tables to include cases with a specified order of priority for variables in a causal model is used to expand the range of theoretical questions that educational researchers can address. The example presented in this document concerns the problem of withdrawal from institutions of higher education using data from the National Longitudinal Study. The model related a respondent's race and ability, two exogenous variables, to postsecondary grades and both grades and ability to the dropout variable. The results support the proposed causal model and indicate that the effects of ability and grades are independent of race. (Author/CTM)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Sponsor: National Center for Education Statistics (DHEW), Washington, DC.
Authoring Institution: Virginia Polytechnic Inst. and State Univ., Blacksburg.
Identifiers: Categorical Data; Causal Analysis; Chi Square; National Longitudinal Study High School Class 1972
Note: Paper presented at the Annual Meeting of the American Educational Research Association (64th, Boston, MA, April 7-11, 1980).