NotesFAQContact Us
Search Tips
ERIC Number: ED202873
Record Type: Non-Journal
Publication Date: 1981
Pages: 21
Abstractor: N/A
Reference Count: N/A
A Common Metric for Integrating Research Findings.
Haladyna, Tom
The choice of a common metric for the meta-analysis (quantitative synthesis) of correlational and experimental research studies is presented and justified. First, a background for the problem of identifying a common metric is presented. Second, the percentage of accounted variance (PAV) is described as the metric of choice, and reasons are given for this choice. Third, two problems relating to the PAV are presented, and their solutions discussed. Finally, an optional procedure for meta-analysis is outlined which incorporates the PAV and the solutions to these problems. The characteristics of PAV are: (1) experimental or correlational results involving two variables, one an independent variable and the other a dependent variable, can be transformed into a PAV metric; (2) the metric ranges on a scale from zero to one; (3) PAV is statistically and computationally simple, and there are several statistical frameworks in which it can be computed; (4) the metric is the percentage of variance accounted in the dependent variable by the independent variable; (5) PAV can be recovered from most statistical results found in research reports; and (6) PAV is a meaningful statistic which can be generally applied to any bivariate relationship. (Author/RL)
Publication Type: Reports - Research; Reports - Descriptive
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A