PDF pending restoration
ERIC Number: ED189122
Record Type: RIE
Publication Date: 1980-Apr
Reference Count: 0
Estimating a Correlation Coefficient Using a Multiple Matrix Sampling Disign.
Estes, Carole; Estes, Gary D.
Multiple matrix sampling is a sampling design in which both test items and examinees are randomly sampled from their respective populations. This study was designed to develop and assess a method for computing an estimate of a correlation coefficient when a multiple matrix sampling design is used. The examinee populations included 212 third-grade pupils and 232 seventh-grade pupils. The item populations consisted of the reading and mathematics subtests of the Comprehensive Tests of Basic Skills: 40, 45, 48, 25, and 25 items, respectively. The population parameters of interest were the correlation coefficients among the five subtests. Four matrix sample sizes were selected (2, 3, 4 and 5), and items and examinees were sampled. The correlation coefficients were adjusted using four different methods: (1) no adjustment; (2) validity-reliability, i.e. the matrix sample estimates and the subtest reliabilities were adjusted; (3) validity; and (4) reliability. The extent to which significant differences existed among the number of matrix samples, the adjustment method, and the interaction of these two factors were examined. The number of matrix samples affected the precision of the estimates. The adjustment method affected the estimate, the amount of bias in the estimate, and the precision of the estimate. (BW)
Publication Type: Speeches/Meeting Papers; Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Comprehensive Test of Basic Skills; Multiple Matrix Sampling
Note: Paper presented at the Annual Meeting of the American Educational Research Association (64th, Boston, MA, April 7-11, 1980).