A comparative study of univariate and multivariate methodological approaches to educational research

Thumbnail Image
Date
1989
Authors
Crawford, Robert
Major Professor
Advisor
Anton J. Netusil
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Abstract

The purpose of this research was to examine a meaningful set of data from a longitudinal educational research project using univariate and multivariate methodological approaches. It was hypothesized that the multivariate analysis would provide a more in-depth and complete understanding of the underlying factors. A second purpose was designed to provide a practical, clear, and concise guide for other educational researchers to draw upon;The researcher first explored the possible relationship between teacher education graduates' teaching certification level and their adequacy ratings on 33 areas of preparation in their teacher education program. A comparison was made between the results of multiple t-tests and the multivariate Hotelling T[superscript]2 followed by discriminant analysis. A second relationship explored the time at which the students graduated and their adequacy ratings in the 33 areas. The researcher also examined the canonical correlation between the 33 preparation areas with five measures of the graduates' overall satisfaction with their education program;It was concluded that the dependent variables identified to be significant using the univariate approach were also judged to be significant using the multivariate approach. In addition, other variables previously found not to be significant were judged to be significant when using the multivariate approach. Variables that centered around the concept of classroom instruction were found to have significant differences among the graduation years. This same cluster of variables differentiated between the two teaching levels, with those certified at the elementary level having higher measures of adequacy;There was general agreement between the results of the univariate approach compared to the multivariate analysis. The advantage of the multivariate approach was in the researcher's confidence in the conclusions. The type I error rate was maintained and most importantly, the multivariate technique was found to be the most appropriate for the design of the research.

Series Number
Journal Issue
Is Version Of
Versions
Series
Academic or Administrative Unit
Type
dissertation
Comments
Rights Statement
Copyright
Sun Jan 01 00:00:00 UTC 1989
Funding
Supplemental Resources
Source