We obtain stock market data from two sources. First, we use daily data from CRSP covering the period 1968 to 2014 (551 distraction events). Second, we again employ TAQ data for the period 1993 to 2014 (225 distraction events). We apply the filters and adjustments described by Holden and Jacobsen (2014) to address withdrawn or canceled quotes, and we use their interpolated time technique to improve the accuracy of liquidity measures.18 All TAQ trades are signed using the Lee and Ready (1991) algorithm. Throughout our analyses, we focus on common stocks (share codes 10 or 11) and exclude penny stocks (closing price < $1). We describe our stock market variables in Section IV, where we also present results for the marketwide analysis