Some Prerequisites for Interpreting Results from Reading Level Matched Designs

Abstract

The present paper discusses conflicting interpretations of specific patterns of results from contrasts of poor and normal readers matched either for chronological age (CA) or reading level (RL). Of general concern is the ambiguity inherent in interpreting results comparing CA- and RL-matched groups on cognitive tasks that could be affected by reading ability. It is argued that causal inferences that might be prompted by a given pattern of results are permissible only when (a) that pattern is observed with a high degree of consistency on a wide variety of relevant cognitive tasks; at least some of which are not readily influenced by experience in reading, (b) if the data are organized by a reasonably plausible theory of reading disability; and (c) if the hypotheses generated by this theory are supported by independent research.