An Experimental Appraisal of the Generalizability of Empirical Revision Procedures for Curriculum Materials [electronic resource] / Eva L. Baker.
An experiment was conducted to determine if the use of rules to guide revisions has generalizable effects across programs and programers. The hypothesis was that pupil achievement would be greater for instructional programs revised according to specified rules than for unrevised programs. Each of 10...
Saved in:
| Online Access: |
Full Text (via ERIC) |
|---|---|
| Main Author: | |
| Format: | Electronic eBook |
| Language: | English |
| Published: |
[S.l.] :
Distributed by ERIC Clearinghouse,
1970.
|
| Subjects: |
| Summary: | An experiment was conducted to determine if the use of rules to guide revisions has generalizable effects across programs and programers. The hypothesis was that pupil achievement would be greater for instructional programs revised according to specified rules than for unrevised programs. Each of 10 programers produced a self-instructional sequence for fifth graders. The 50-frame programs including pretest and 10-item posttests were to take approximately one hour to complete. After each was administered to two fifth graders the first draft programs with technical performance data were randomly distributed to the programers with a set of rules for use in revision. Then 64 fifth graders were randomly assigned to receive either a first draft or revised program. Scores were averaged for each condition, first draft or revision, for each program and the mean for each program per condition provided the unit of analysis. A t test based on the paired observation model was employed using the S's mean performance per condition, matched by program, as the entry. An omega square value to test the strength of association was computed. It was inferred from the results that empirical revision based on rules pays off modestly, i.e., instructions can be provided so that the efforts expanded in revision will have some effect on the obtained levels of pupil performance. Confidence limits for the difference were calculated so that cost of revision could be assessed. [Not available in hardcopy due to marginal legibility of original document.] (JS) |
|---|---|
| Item Description: | ERIC Document Number: ED040963. ERIC Note: Paper presented at annual meeting, AERA, Minneapolis, 1970. |
| Physical Description: | 10 p. |