Stanovich K E. Toward an interactlve-compensatory model of individual differences inthe development of reading fluency. Read. Res. Quart. 16:32-71, 1980.

My colleague Richard F. West and I began studies of context effects on the word recognition of children when we were both graduate students at the University of Michigan in the mid-i9705.At the time we were heavily influenced by the popular top-tiown models of the reading process.’ We fully expected to make an experimental mark by providing more rigorous confirmationsof the individual-differences predictions of these models than had previously existed. To our surprise our initial studies 2 failed to confirm one ofthe most basic predictions ofthe topdown modelnt that the word recognition of children who were better readers would be more influenced by contextual information. Instead, we observed piecindy the opposite the word recognition processes of better readers were less influenced by contextual manipulations. This finding—which contradicted assumptions about the effects of context that were common in the reading literature and that had been widely disseminated in textbooks on reading education—was confirmed in numerous subsequent investigations.~ 5