ERIC Number: ED164609
Record Type: RIE
Publication Date: 1978-Mar-28
Reference Count: 0
Estimation of the Regression Effect Using a Latent Trait Model.
Quinn, Jimmy L.
A logistic model was used to generate data to serve as a proxy for an immediate retest from item responses to a fourth grade standardized reading comprehension test of 45 items. Assuming that the actual test may be considered a pretest and the proxy data may be considered a retest, the effect of regression was investigated using a percentage of the group from the lowest tail of the pretest distribution. The mean of the retest group is assumed unbiased by errors of measurement. Thus the means of the pretest and the proxy test should differ, on the average, by an amount equal to the regression effect. Regression effects for various sizes of samples tested, from 25 to 500, and size cuts, from the fourth percentile to the 89th percentile, were reported. Item parameters had been estimated by Birnbaum's three-parameter logistic model separately for four groups of 1,000 individuals each, selected by sex and by race. The regression effect was most powerful with samples from the extreme lower tail, and decreased to a low, stable level at about the 30th or 40th percentile. (Author/CTM)
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: N/A
Identifiers: Birnbaum Models
Note: Paper presented at the Annual Meeting of the American Educational Research Association (62nd, Toronto, Ontario, Canada, March 27-31, 1978)