This report presents the results of the largest survey and testing program for students in home schools to date. In Spring 1998, 20,760 K-12 home school students in 11,930 families were administered either the Iowa Tests of Basic Skills (ITBS) or the Tests of Achievement and Proficiency (TAP), depending on their current grade. The parents responded to a questionnaire requesting background and demographic information. Major findings include: the achievement test scores of this group of home school students are exceptionally high--the median scores were typically in the 70th to 80th percentile; 25% of home school students are enrolled one or more grades above their age-level public and private school peers; this group of home school parents has more formal education than parents in the general population; the median income for home school families is significantly higher than that of all families with children in the United States; and almost all home school students are in married couple families. Because this was not a controlled experiment, the study does not demonstrate that home schooling is superior to public or private schools and the results must be interpreted with caution. The report clearly suggests, however, that home school students do quite well in that educational environment.
The fairness or appropriateness of measures of aptitude or achievement have long been of interest and theoretical concern, but more recently have surfaced as variables requiring empirical investigation. One of the first instruments that addressed these concerns was a purported culture fair test devised by Eells, Davis, Havighurst, Herrick and Tyler (1951). From that time through the present, these issues continue to stir controversy for test developers, researchers and the legal profession (Williams, Mosby and Hinson, 1978).In an attempt to provide a framework for reducing or minimizing bias in educational assessment, several methods which address the problems of either test bias or item bias have been proposed. Investigations of test bias have used models ~hich determine whether a test unfairly favors examinees of particular groups, e.g., cultural or linguistic groups. Invesigations of item bias seek to identify specific items within a test which exhibit dissimilar response patterns for examinees having equal ability but different group memberships. The first type of investigation is of primary interest to test users who need to choose intact existing tests for particular applications. The second is of interest to test developers for use in their attempts to identify and eliminate biased items.The present study focused on several models for detecting item bias and the differential performance of these models when applied to a variety of bias structures likely to be encountered in educational assessment data. The data in this investigation were produced with a Monte Carlo procedure in which the amount and type of item bias were specified a priori. TECHNIQUES INVESTIGATEDInvestigations of item bias provide an empirical basis for the identification and elimination of items which appear to measure different traits for different population groups. This study applied two transformed item difficulties approaches, three item characteristic curve (1CC) approaches, and two chi-square approaches to obtain measures of the degree of bias for each item within generated item pools. The procedures used for the application of these approaches are described in the following paragraphs.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.