| This research examined a revised version of the Merrill-Palmer, a test of childhood development, in order to assess the cultural validity of items (i.e., differential item functioning; DIF). Archival data, obtained from the Stoelting Co., was used to determine the DIF of items for 245 African American, Caucasian, and Hispanic children, ages 3--6. Potential bias was determined through the use of partial correlation. A significant correlation between item and ethnicity indicated that, in general, one ethnic group performed better on a particular item than another ethnic group. For the purpose of this study, those items that showed a significant correlation between item and ethnicity were considered culturally biased. Of the 320 items analyzed, only 38 were found to be biased toward at least one ethnic group. Eight items were found to be biased toward Hispanics, 14 toward African Americans, and 17 toward Caucasians. A binomial test indicated that the number of items which showed DIF did not exceed the number expected by chance. There was no specific pattern in regards to the type of items that showed DIF, or in what direction specific items showed bias. These findings suggest that the Tryout Edition of the Merrill-Palmer, Revised (MP-R), was not systematically biased toward any of the three racial groups examined. Results from the Tryout Edition suggest that the final edition of the MP-R will also show little, if any, racial bias at the item level. These results are promising for clinicians interested in testing children using culturally fair tests. |
