The Effect of Research Design on Free Written Recall as a Measure of Reading Comprehension in a Second Language

Abstract

0 An investigation was conducted of the effect of research design in studies using a free written recall task to measure second language (L2) reading comprehension. Although studies (Carrell 1982, 1984a, 1984b, Bernhardt 1983, Connor 1984, Lee and Ballman 1985) using a free recall task differ in many ways, variables selected for examination were 1) the language (native versus target) in which recalls are written, 2) the use of prereading instructions orienting subjects to the recall task, and 3) the level of L2 proficiency of subjects. The 320 subjects participating in the study were volunteers enrolled in four different levels of Spanish classes at Michigan State University and the University of Michigan. Each of the 80 subjects at each level received a packet containing an agreement form, a reading passage (with or without instructions), and a third page on which they were directed to write their recalls either in English (native language) or Spanish (target language). The results of a three-factor ANOVA revealed significant main effects for level and language of recall, as well as a significant interaction between orienting instructions and level. No significant main effect was found for orienting instructions, and none of the other interactions (Orienting Instructions x Language, Language x Level, Orienting Instruction x Language x Level) was significant. In view of the significant interaction of Orienting Instructions x Level, the significant main effect found for level must be cautiously interpreted. However, while a number of studies of L2 reading ability (Bernhardt 1983, Carrell 1984a, Connor 1984) have used subjects at only one level of proficiency, the results of this study suggest that findings cannot be generalized to learners at other levels of proficiency. The significant main effect found for language of recall has definite research implications, however. The main effect found for language of recall (with learners recalling more when writing in their native language than in the target language) suggests that research on L2 reading comprehension