The dimensionality of speaking and writing: a multilevel factor analysis of situational, task and school effects.

Abstract

This article addresses the question whether communication apprehension (CA) should be regarded as a unidimensional or alternatively as a multidimensional construct. The answer is not only interesting from a theoretical point of view. There might also be practical implications for the treatment of CA. If CA were to appear to be unidimensional and a student's level of CA were to be the same across situations and tasks, there would be no need to tailor the treatment to particular speaking situations or specific writing problems. If, however, CA appeared to be multi-dimensional and a student might have a variety of different levels and types of CA, one type of treatment might be more effective for one student than for another one. To examine the effects of situational, task and school effects on speaking and writing apprehension. Use was made of the dataset of the 1987-1988 National Assessment of Language Performance in the Netherlands. The nationally representative sample consisted of 1448 students from 184 secondary schools; 52% of the students were boys and 48% were girls; the mean age of the students was 15 years 6 months. Speaking and writing apprehension were measured by means of self-report measures in grade 9. Multilevel factor analysis (MLFA) was used to determine the dimensionality of the measurement of speaking and writing apprehension. First, all seven speaking situations and three out of four writing problems could be distinguished empirically. Speaking and writing apprehension are clearly multidimensional constructs that depend on the speaking situation and the writing task. Second, correlations between speaking and writing apprehension were rather low. Speaking and writing apprehension seem to represent skill-specific constructs, which cannot be considered as equivalent forms of communication apprehension. Third, differences between schools in the level of speaking and writing apprehension were very small compared to measurements of speaking and writing performance.