Analysing Poor Reading Skills: A Comparison Between PIAAC Literacy Levels and Reading components

Abstract

The aim of this paper is to compare two measures of literacy skills used in the Programme for the International Assessment of Adult Competencies (PIAAC), i.e. literacy levels and reading components. Different from assessing literacy levels, which cover the entire range of reading skills from very high to very low, the reading components assessment is specifically designed to test the reading skills of the low literate. As literacy skills are an important determinant for labour market success, we examine whether the reading components assessment is a better predictor of employment of the test takers than the general literacy level assessment used in PIAAC. We compare different logistic regression models that include literacy levels and the reading components and use the Wald-test to compare the model fit of each examined model. In a further step we test whether the predictive validity differs between low and high literate test takers. Our results indicate that the PIAAC reading components are more useful in predicting employment of the low literate than the PIAAC literacy levels.