Most word learning studies conducted over the past decade have shown that methods of word instruction that emphasize the graphic or phonic features of words presented alone (word alone method) affect learning rate more positively than do methods that emphasize meaning through oral, written, and pictorial cues (context methods). To compare the efficiency of the two methods of word instruction, a 1981 study examined the performance of 158 kindergarten children on two separate modes of assessment--a sentence mode test (paralleling aspects of the context treatment) and an isolation mode test (paralleling aspects of the word alone treatment). The findings of that study suggested that isolation mode tests yielded performance scores biased in favor of children taught by word alone methods. To extend the findings, a second study selected the number of words recognized on the sentence mode test protocols of 11 children from the context method group and matched them with those of 11 children from the word alone group. The oral reading errors made by children on the sentence mode test were analyzed to determine the differential influence of the two methods on their use of graphophonic, syntactic, and semantic constraints. The results showed no significant differences between groups in their use of graphophonic cues. However, context method children used syntactic and semantic constraints more effectively than did word alone children. (FL)