On the Determination of Causal Ordering between Vocabulary and Comprehension

Abstract

In a recent issue of this journal, Yap (1979] applied a two-wave, two-variable, cross-lagged panel analysis to an investigation designed to probe the causal ordering between vocabulary and comprehension skills of primary grade pupils. However, Yap has failed explicitly to consider the restrictive circumstances in which unambiguous directionality of effects may be determined in two-wave, two-Variable (2W2V) models. The purpose of this comment is to illustrate-these rather restrictive circumstances in order to caution researchers of reading behavior not to apply indiscriminately the procedures used by Yap. Yap (1979) used the Gates-McGinite reading tests to measure the reading ability of second and third graders in a Hawaiian school. The Gates-McGinite tests yielded one subscale measuring vocabulary, and a second measuring comprehension. The subjects were tested twice, in October 1971, and April 1972.. Applying the 2W2V design, vocabulary and comprehension were measured on two different occasions. Let the measurements of vocabulary be denoted Vt and V2, respectively, and the measurements of comprehension, C, and C2. A general model for the causal linkages that may be present in this system may be expressed in the following equations: