Three self-report measures of the broader autism phenotype (BAP) were evaluated in terms of their internal consistency, distribution of scores, factor structure, and criterion-related validity in a non-clinical sample. All measures showed a continuous distribution. The SRS-A and BAPQ showed expected sex differences and were superior to the AQ in terms of internal consistency. The proposed factor structure of the BAPQ replicated better than the proposed structures of the other measures. All measures showed evidence of criterion validity via correlations with related constructs and each measure incremented the others in predicting related constructs. However, the SRS-A and BAPQ were generally stronger in this domain. Recommendations for the use of these instruments for measuring the BAP in non-clinical populations are discussed.