Presented at the 19th Annual Conference of the Society for the Scientific Study of Reading; Montreal, Canada, July 11–14 2012.

Purpose: Matthew effects in reading development refer to a longitudinally widening gap between high and low achievers, whereby the latter fall increasingly behind as a consequence of diminished practice. Although decreased short-term benefit from experience has been documented for poor readers, longitudinal divergence has been difficult to establish. Here it is argued that there are significant theoretical and psychometric impediments to the empirical elucidation of Matthew effects.
Method: Longitudinal analyses of reading accuracy and fluency, spelling, vocabulary, and reading comprehension were conducted on data from 587 children in Grades 2–4 followed over 2 years in 5 measurement waves approximately 6 months apart. Highly correlated pairs of measures were used to cross-select 1st-wave high- and low-performing children. Growth curves were modeled with linear mixed-effects models contrasting high-low groups.
Results: Weak patterns of divergence were seen in spelling and fluency, whereas convergence was observed in vocabulary and comprehension. However, the interpretability of these interactions is compromised by lack of metric equivalence between the “low” and “high” performance levels. Moreover, the decelerating slopes indicate lack of growth equivalence across age groups, rendering comparisons of different grade and ability levels difficult to interpret.
Conclusions: Converging and diverging patterns of performance cannot be taken at face value as either supporting or contradicting the Matthew effects framework, in the absence of properly calibrated psychometric scales. Better understanding of growth patterns is a prerequisite to the definition of age-appropriate performance and expected rates of development. The longitudinal Matthew effects may not be possible to establish conclusively.