We study the informativeness of trades via discount and full-service retail brokers. We find that trades via full-service retail brokers are statistically and economically more informative than are trades via discount retail brokers. This finding holds in every year over the 12-year sample period and in various subsamples. We also find that past returns, volatility, and news announcements positively relate to the net volume of discount retail brokers, but these variables are unrelated to the net volume of full-service retail brokers. Our results suggest that broker type selection bias is an important consideration in studying individual investors' trades.