Appendix B

Appendix B: Response Rates

In general, a response rate reflects the proportion of total sampled eligible cases from which data were obtained. In the TIMSS Videotape Classroom Study, the response rate indicates the percentage of sampled schools for which videotapes were completed. In each country, response rates can be computed both before and after replacement. The response rate before replacement identifies the proportion of originally sampled schools that participated; the response rate after replacement gives the percentage of all schools sampled (original and replacement schools) that participated.

In addition, response rates can be either unweighted or weighted. Unweighted response rates, computed using the actual numbers of schools, reflect the success of the operational aspects of the study (getting schools to participate). Response rates weighted to reflect the probability of being selected into the sample describe the success of the study in terms of the population of schools to be represented. Table B1 provides unweighted and weighted response rates for each of the three countries before and after replacement.

Table B1

Unweighted and weighted school response rates before and after replacement for Videotape Classroom Study,1994-95, by country

Unweighted
Response rate type	Germany	Japan	United States
Before replacement	85.0	96.0	63.3
After replacement	87.0	96.2	66.4

Weighted
Response rate type	Germany	Japan	United States
Before replacement	86.0	95.8	65.3
After replacement	87.7	95.9	68.5

NOTE: In Japan and the United States, the final video sampling weights included nonresponse adjustments. This adjustment was factored out in the calculation of the weighted response rates for Japan, and for the weighted U.S. response rates, base weights (which did not include the nonresponse adjustment) were used.

SOURCE: U.S. Department of Education, National Center for Education Statistics, Third International Mathematics and Science Study, Videotape Classroom Study, 1994-95.

In Germany, 85 of the 100 originally sampled main TIMSS schools for the videotape study agreed to participate. For the other 15 schools, one of the main TIMSS replacement schools for the school which refused the videotaping was contacted; all 15 of the replacement schools cooperated with the study.

In Japan, 50 schools were sampled; 48 of these agreed to the videotaping. One of the schools which refused was replaced with a school matched to it from the same stratum. Another replacement school was identified, but it could not be matched to the other refusal school and was weighted to be self-representing.

Sixty-nine of the originally sampled 109 schools in the United States agreed to participate in the videotaping. As described in chapter 2, the paired schools for 13 of the refusals were contacted, and 12 of these replacements agreed to participate.

Table B1 shows that response rates in Japan were very high, and response rates in Germany were also good. The response rates in the United States were somewhat lower. Many of the refusals were due to a high refusal rate for participation in the main TIMSS study: The Population 2 weighted response rate before replacement for the main TIMSS pencil-and-paper assessment was 77 percent (Martin and Mullis, 1996, Table 2.7), and roughly three-quarters of the sampled video schools which did not participate had refused to participate in TIMSS altogether. Refusal to participate in TIMSS overall also meant that the paired schools for many of the refusal schools were not contacted as potential replacements. Due to a somewhat lower response rate for the United States, caution should be used in generalizing the sample results to the population from which it was drawn.

Questions, problems or comments with this website? Contact timss@ed.gov