In three experiments on the effects of word imagery, length, and frequency on reading difficulty, high-frequency words were found to be easier to read for both good and poor readers. High-imagery words were easier to read for poor readers only. Word length had little effect on reading difficulty for either good or poor readers. The differential effect of word imagery on reading difficulty for good and poor readers was interpreted in terms of the types of reading strategy used -phonics for good readers and whole word reading for poor readers. When children are forced to learn to read words by a whole word method, word imagery predicts ease of learning for both good and poor readers.Although there has been a considerable amount of research on the effects of such word attributes as concreteness, imagery, meaningfulness, and frequency on memory and tachistoscopic word perception, little work has been done on the effects of these variables on word decoding. However, some recent research has studied the effect of word attributes on the reading performance of patients with dyslexia acquired through brain damage (Marshall & Newcombe, 1973; Richardson, 1975a Richardson, , 1975b Shallice & Warrington, 1975). Many of these patients seem to have a specific impairment of the ability to use grapheme-phoneme correspondence rules. Research with these patients has shown that imagery, concreteness, frequency, and part of speech are related to reading difficulty, but word length seems to have little or no effect. Richardson (1975aRichardson ( , 1975b has shown that a word's imagery value is the best This research reported in this article was supported by a grant from the Australian Research Grants Committee. The author wishes to thank Don McNicol for his advice and encouragement throughout the research and Julie Preston for her help with the preliminary testing of subjects.Requests for reprints should be sent to A.