A Test for Program Effect in the Abscence of a Proper Control Group.
Abstract
A statistical method is proposed to analyze certain designs for program effect when no proper control group is present. It is shown that under certain partially verifiable assumptions the data can be analyzed by standard analysis of variance techniques. If these assumptions are doubtful, then a statistic is proposed which is shown to have an asymptotic t distribution. The analysis is then carried out on some research data. Keywords: Analysis of Variance, Control Group, Convergence in distribution, t-distribution.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1985
- Accession Number
- ADA163783
Entities
People
- Eric S. Tollar
Organizations
- Florida State University