In recent years, there has been increasing debate and research regarding which modality of vocabulary knowledge has the strongest correlation to reading, with particular focus on distinctions between testing L2 form and L2 meaning, and between recall of answers from memory and recognition of answers from fixed options. However, relatively little attention has been paid to find out which modality has the strongest correlation to listening ability. A recent meta-analysis by Zhang and Zhang (2020) indicated that meaning recall was the superior predictor of reading proficiency. Although their results showed that form recall had the highest correlation to listening, the difference between form recall and meaning recall was statistically insignificant. The present study uses data from McLean et al. (2020) of learner responses to 1000-item vocabulary tests employing written tests of meaning recall, form recall, meaning recognition and Yes/No modalities, sampling them with replacement to create thousands of 100-item tests using a bootstrapping approach. The test scores were then correlated to measures of listening and reading proficiency for comparison. The results indicated that for written tests, meaning recall, form recall, meaning recognition and form recognition had the strongest correlations to both reading and listening, in descending order. All comparisons were statistically significant.