A COMPARISON OF STRATEGIES USED BY BLACK AND WHITE STUDENTS IN SOLVING SAT VERBAL ANALOGIES USING A THINKING ALOUD METHOD AND A MATCHED PERCENTAGE‐CORRECT DESIGN

Abstract

ABSTRACTTwo studies are reported which investigate how Black and White examinees differ in their ability to solve analogy items. The first helps to identify three factors which affect differential ethnic performance on analogy items; the second study carried out a controlled experiment in order to investigate, via a thinking aloud paradigm, how the same three factors influenced the types of strategies that were differentially evoked for two matched groups of Black and White students.In Study 1 three variables were found to be significantly associated with ethnic examinee differences in solving analogies: the item difficulty level (easy/hard dimension), whether the item had science content or not, and whether the item used a part/whole relationship in the stem. The conclusion reached in Study 1 was that Black examinees perform differentially better than the matched White examinees on the hard analogy items (typically these had a non‐science content), and appear to systematically perform less well vis‐a‐vis the matched White examinees on the easy analogy items (which often have a science content).The results for Study 2 indicated that three strategies help to account for ethnic differences in solving analogy items: use of an indirect or weak induction (used primarily by Black students on the hard science items; use of this strategy, especially for hard science items, helped Blacks to get more of these items correct). use of word association in selecting a response which typically results in a wrong option being chosen (this was used more often by the Black students for all items except hard science items for which Black and White students were approximately equal in use of this strategy). problems pronouncing the words in the item stem (Black students mispronounced some of the stem words significantly more often than White students). Collectively these results helped to clarify some of the significant differences found between the two ethnic student groups in solving analogy items as a function of the item's characteristics. In general both studies provided evidence that item difficulty (the easy/hard dimension) was a significant factor underlying ethnic performance; the evidence showed that Black students perform better than their matched White student counterparts on the hard items and perform less well with respect to the easy items. But for Study 2 no confirming evidence was found that Black students performed less well on items involving science content, contrary to the results found in Study 1. Reasons for such differences across studies are discussed. Over and above ethnic differences, additional analyses in Study 2 uncovered differential strategy use as a function of whether analogy items contained a science or nonscience content.