ABSTRACTThree experiments explored the usefulness of a visual preference technique for assessing word comprehension in infants. Pairs of slides were presented, and visual fixation to each recorded. The parent (in Experiments 1 and 2) or the experimenter (in Experiment 3) then prompted the child with the name of one of the slides (“Find the ——,” “Do you see the ——?”), and visual fixation to each was recorded again. Percentage fixation to the picture matching the word referent following the prompt question was compared to initial preference for that picture, providing a measure of the acceptability of the picture as a word referent, independent of its initial salience.The first experiment demonstrated increases in comprehension from 8–14 and 14–20 months. The second experiment established longitudinal stability of comprehension from 14 to 20 months. Word comprehension scores were related to parent report of vocabulary knowledge and difficulty of words. The final experiment measured 1-week test-retest reliability of the word comprehension score at 14 and 20 months and replicated the effect of word difficulty. Methodological and theoretical implications of these results are discussed, in particular, the profound effect of stimulus salience and the lack of sex differences in word comprehension.