Essays on Gamified Online Platforms

Abstract

Gamified online platforms use game design elements in non-game contexts to increase users' engagement and improve their performance outcomes. In this dissertation, using data from two gamified online platforms, I study the effect of disclosing individual's performance ranking and popularity rating on different outcomes. First, I focus on an online weight-loss community which uses gamified challenges to enable users to set short-term weight-loss goals, and incentivize them to pursue their goals by sharing individuals' progress and rankings with other challenge members via leaderboards. I study the effect of participation in gamified challenges on weight-loss progress. I utilize the system GMM and Inverse Probability Weighting (IPW) approach to address endogeneity issues. The results indicate that participation in gamified challenges has a positive and significant but short-term effect on weight-loss. Moreover, participation in gamified challenges are less effective when users focus only on dietary or physical activity instructions, and more effective when users add a numeric weight-loss target to their dietary or physical activity instructions. Second, I focus on a gamified dating platform where users play a game and rank-order members of the opposite sex and are then matched based on a Stable Matching Algorithm. A key piece of information shown to users during the game is a popularity rating, ranging from one to three stars. I examine the effect of a user's popularity on her demand i.e. I quantify the causal effect of a user's star-rating on the rankings that s/he receives during the game and the likelihood of receiving messages after the game. Popularity can increase one's appeal. However, popular people are less likely to reciprocate, and hence users may strategically shade their revealed preferences for them to avoid rejection. To overcome the endogeneity between a user's star-rating and her unobserved attractiveness, I employ non-linear fixed-effects models. The results indicate that three-star users receive worse rankings during the game but receive more messages after. I link the heterogeneity across outcomes and user-level observables to the perceived severity of rejection concerns and establish strategic shading as the mechanism. Further, I show that users' rejection concerns are consistent with Step-1 bounded rationality.