THE GENERATION EFFECT IN READING AND PROOFREADING - IS IT EASIER OR HARDER TO DETECT ERRORS IN ONES OWN WRITING

Using a naturalistic text generation and proofreading task, we investi gated two questions concerning the effect of text familiarity on proof reading performance. Can experimental evidence be provided for the int uition that it is harder to proofread one's own writing than someone e lse's? Will the effect of text familiarity on proofreading differ as a function of whether the familiarity is self-generated or experimental ly-induced? Subjects spent 30 minutes composing an essay on student li fe; after a 20 minute interval (Experiment 1) or a two week interval ( Experiment 2) they proofread their own essay, another subject's essay after being familiarized on an error-free version of it, or another su bject's essay without the benefit of a preview. Experiment 1 showed th at subjects were less able to detect errors (e.g., The best part of st udent like is socializing.) in self-generated essays than in unfamilia r other-generated essays; on the other hand, they were better able to detect errors in familiar other-generated essays that in unfamiliar on es. Experiment 2 showed that the disadvantage for proofreading self-ge nerated text is likely a by-product of extreme familiarity rather than any special quality of self-generated knowledge per se. The results h ave implications for models of skilled reading.