Consistency of teacher communication: A sampling problem in interaction analysis

✍ Scribed by Jenny R. Armstrong; M. Vere De Vault; Eleanore Larson

• Publisher: John Wiley and Sons
• Year: 1967
• Tongue: English
• Weight: 271 KB
• Volume: 4
• Category: Article
• ISSN: 0033-3085
• DOI: 10.1002/1520-6807(196701)4:1<74::aid-pits2310040119>3.0.co;2-p

✦ Synopsis

A major problem confronting researchers and curriculum workers interested in interaction analysis techniques in education is that of determining the time required to adequately sample a teacher's communication behavior. This study was undertaken to provide some tentative answers to this persistent problem. Three different suggestions have been reported in the literature regarding how long one should sample teacher behavior in order to get a representative sample of that behavior. Flanders (1965) recommended 20 minutes, whereas in an earlier investigation done by Jersild and Meigs (1939) a series of five 15-minute time samples was suggested.

In still another study (Moffitt, 1961)" . . . it was found that a satisfactory specimen record of teaching covering twenty to thirty minutes of teaching could be made . . ."

In the reported studies, however, there was no evidence that a systematic analysis of this problem had been undertaken.

The purpose of this study then was to determine whether or not twenty minutes, or even an hour, is a long enough time period to adequately sample a teacher's verbal communication behavior when using either the Wisconsin Teacher Education Research Project (TERP) system of categorization or the Flanders system of categorization.