Show Less
Restricted access

Examining the Interaction among Components of English for Specific Purposes Ability in Reading

The Triple-Decker Model


Yuyang Cai

The book presents the Triple-Decker Model that describes the interactive mechanism among key constituents of English for Specific Purposes Ability (ESPA): language knowledge, strategic competence and background knowledge (nursing knowledge).

First, the model reveals that ESPA constituents can be assigned to three groups according to their roles in determining ESPA in reading: automators (language knowledge and background knowledge) that respond most directly, assistants (the cognitive aspect of strategic competence) that come to assist when automators are insufficient or boggle down, and regulators (the metacognitive aspect of strategic competence) that supervise all cognitive activities. Second, the model demonstrates the effect of strategic competence and background knowledge on ESPA fluctuated with the continuous increase of language knowledge.

The book also demonstrates the use of two innovative analytical techniques: composite scores based on bifactor multidimensional item response theory for scoring ESP reading tests and the multi-layered-moderation analysis (MLMA) for detecting linear and nonlinear moderation relations.

Show Summary Details
Restricted access

Chapter 4 Preliminary Results


The purpose of this chapter was to calibrate and score responses to the four scales used in the current study (i.e., he MNKT, GKT, MNRCT and SCQ) and compute composites (composite scores) representing the domains tapped by each scale. It proceeds in four sections. The first section provides summarized and item-level descriptive statistics and reliability estimates. The second section presents results of a series of analyses aimed to evaluate the appropriateness of applying bifactor-MIRT/MGRM models (i.e., dimensionality assessment, local dependence detection, and model specification). The third section elaborates on the process of composites computation based on the bifactor-MIRT/MGRM discrimination estimates. The presentation flows scale by scale following the order of the MNKT, GKT, MNRCT and SCQ.

This section provides summarized and item-level descriptive statistics (i.e., the means, standardized deviations, kurtosis, and skewness) and reliability estimates (i.e., internal consistency reliability estimates) for each scale. These analyses served to examine how well the items performed in measuring different types of ability they were designed to measure (see, Carr, 2008; Kunnan & Carr, 2012).

Table 4.1 provides summary descriptive statistics for the MNKT scale. The overall mean was 13.55 out of 24 total possible points (i.e., 56.46 % correct) and the standard deviation was 4.09. Skewness was -0.03 and kurtosis was -0.13, suggesting reasonable normal distribution.

Table 4.1: Summary Descriptive Statistics for the MNKT (N=1491)

Table 4.1 provides the descriptive and reliability statistics for participants’ responses to the MNKT scale. The means...

You are not authenticated to view the full text of this chapter or article.

This site requires a subscription or purchase to access the full text of books or journals.

Do you have any questions? Contact us.

Or login to access all content.