Modeling ESP ability in reading : a focus on interaction among grammatical knowledge, background knowledge and strategic competence

Abstract

A current dilemma in assessment of English for Specific Purposes (ESP) is the burgeoning of ESP assessment programs and the vaguely understood concept of ESP ability. Douglas (2000) posits the nature of ESP ability lies in the interaction among its three components: language knowledge, subject-matter background knowledge and strategic competence. The purpose of the study was to investigate the validity of Douglas‘ hypothesis through exploring potential interaction among grammatical knowledge (one aspect of language knowledge), subject-matter background knowledge (operationalized as medical and nursing knowledge), and strategic competence in explaining ESP ability in reading (ESPAR). The study investigated how different components of strategic competence could combine with themselves and with background knowledge and grammatical knowledge to enhance ESPAR. A pool of 1491 medical and nursing students from eight medical and healthcare colleges in China participated by responding to four instruments that were designed to measure medical and nursing knowledge, grammatical knowledge, medical and nursing English ability in reading, and strategic competence, respectively. The study was designed as correlational study. It used two different statistical methods in two stages. In the first stage, the bifactor-multidimensional item response theory (bifactor-MIRT) was used to calibrate, score and formulate composite scores representing students‘ ability measured by each scale and to examine the evidence of factorial validity of the underlying constructs. In the second stage, structural equation modeling with the bifactor-MIRT based composites was applied. This was to done to explore the one-moment main effect of, and the two-moment and three-moment interaction effects among grammatical knowledge, background knowledge and strategic competence in explaining ESPAR. The results of the study offered empirical evidence in support of the validity of the underlying constructs operationalized in the grammatical knowledge test, the background knowledge test, the English for Specific Purposes (ESP) reading test, and the strategic competence survey. The findings suggested that grammatical knowledge produced largest explained variance of the ESPAR, followed by background knowledge and strategic competence. The results also suggested linear moderation between grammatical knowledge and strategic competence, that is, they enhanced or constrained each other‘s capacity in explaining ESPAR. More importantly, the results also showed quadratic moderation of grammatical knowledge on background knowledge. That is, the explanatory capacity of background knowledge increased with the increase of grammatical knowledge until the latter reached a high threshold (theta=0); when grammatical knowledge continued to increase, background knowledge effect started to decrease. Furthermore, the results showed that different components displayed a Triple-Decker pattern that best enhanced ESPAR, apart from the linear moderation between evaluating and grammatical knowledge and the quadratic moderation of grammatical knowledge on background knowledge. On the lowest deck were automatized grammatical knowledge and background knowledge (Automators); on the middle deck were cognitive strategies (Assistants) that assisted the deficiency of the Automators; and on the top deck were metacognitive strategies (Regulators) that regulated the Assistants.