Test expectancy and performance on factual and higher-level questions

Abstract

The study involved an investigation of the effect of test expectancy on preferred strategy use and test performance on factual and higher-level questions in learning from expository text. A total of 42 undergraduate college students were instructed to expect factual or higher-level questions, or were given no information regarding the test. Students studied a 3145-word passage from an introductory psychology text, reported their use of study strategies and why each was used, and then completed 42 multiple-choice questions. Results indicated that test expectancy did not affect preferred strategy use, nor did it affect test performance. Subsequent analyses were employed to identify differences between those who exhibited superior or inferior performance on each type of question. Superior performance on factual questions over main points was associated with higher reading comprehension scores, longer study time, higher confidence rating on the questions, and more extensive use of embellished text. Superior performance on factual questions over details was associated with higher confidence rating in answering questions and extensive use of a number of author-provided cues to strategic text processing (embellished text, examples, analogy). In sharp contrast, superior performance on higher-level questions was not associated with confidence rating, nor use of author provided cues. In addition, use of preferred strategy and metacognitive awareness was not associated with performance on any type of question. A majority of students retrospectively reported use of underlining and rereading, and indicated the importance of selection and retention of important information as reasons for deploying the strategies, suggesting a rote-learning approach to learning from text.