An attempt was made to modify the Gambling Functional Assessment (GFA), which was proposed to identify four possible contingencies maintaining the respondent's gambling behavior. However, previous research found that it only identified two contingencies (i.e., positive vs. negative reinforcement), with some items cross-loading on both contingencies and one not loading at all. A total of 1,060 undergraduate students completed a revised version of the GFA containing 22 items. Exploratory factor analyses conducted on a random selection of half of the participants led to a two-factor solution (positive and negative reinforcement) for 16 of the items that strongly loaded on the two factors. Confirmatory factor analyses conducted using structural equation modeling on the data from the other half of the sample confirmed the two-factor model. The GFA-Revised consists of 16 items, 8 each measuring positive and negative reinforcement contingencies. Although this revised measure cleanly parses the two contingencies, the data indicate that gambling maintained by positive reinforcement is more frequent than gambling maintained by negative reinforcement. This outcome will make directly comparing the two contingencies difficult, especially given that evidence suggests that gambling maintained by negative reinforcement is more strongly associated with pathology than gambling maintained by positive reinforcement.