We examined the robustness of the unconscious arithmetic effect with respect to the exclusion criteria used to prepare the raw data. Under the SPSS Files fork, you will find four SPSS data files. Each file contains our participants' mean RT scores following the application of a different exclusion criteria. In each set, we excluded participants who were identified as having awareness of the primes either through an objective awareness check or subjetcive debriefing questions. We also excluded any trials on which participants did not correctly state the target number. The first set of analyses that we present in our paper are based on the data set titled "UnconsciousArithmeticSPSS". These scores were calculated by excluding trials which differed from the participant's mean RT for the respective block by more than 3 SD. Analyses were conducted on log transformed data. The next analyses that we present are based on the data set titled "UnconsciousArithmetic_Revised". Through the review process, it was brought to our attention that our initial criteria did not match that of Sklar et al. (2012). This data set includes scores which were calculated using their exact exclusion criteria: (1) trials with RT > 2000 ms or < 200 ms; and (2) trials which differed from the participant's mean RT overall by more than 3 SD. Analyses were conducted on untransformed data. We also present analyses based on the data set titled "Unconscious Arithmetic_200RT2000" which excludes only the trials with RT > 2000 ms or < 200 ms. Finally, we present analyses based on the data set titled "UnconsciousArithmetic_3SDblock" which uses both the trials with RT > 2000 ms or < 200 ms and the differed by mean RT by more than 3 SD criteria, but by calculating mean RT by block rather than overall.