[Paper Review] How Gamification Affects Physical Activity: Large-scale Analysis of Walking Challenges in a Mobile Application

Ali Shameli, Tim Althoff+2 more|arXiv (Cornell University)|Feb 24, 2017

Educational Games and GamificationPsychology35 references19 citations

TL;DR

This study analyzes 2,500 walking challenges in a mobile health app across 800,000 person-days, finding that gamified competitions increase average physical activity by 23%. The research identifies key design principles—such as matching users with similar baseline activity levels and ensuring balanced gender composition—to maximize engagement and effectiveness, and develops a predictive model with 75.1% AUC for identifying highly engaging competitions.

ABSTRACT

Gamification represents an effective way to incentivize user behavior across a number of computing applications. However, despite the fact that physical activity is essential for a healthy lifestyle, surprisingly little is known about how gamification and in particular competitions shape human physical activity. Here we study how competitions affect physical activity. We focus on walking challenges in a mobile activity tracking application where multiple users compete over who takes the most steps over a predefined number of days. We synthesize our findings in a series of game and app design implications. In particular, we analyze nearly 2,500 physical activity competitions over a period of one year capturing more than 800,000 person days of activity tracking. We observe that during walking competitions, the average user increases physical activity by 23%. Furthermore, there are large increases in activity for both men and women across all ages, and weight status, and even for users that were previously fairly inactive. We also find that the composition of participants greatly affects the dynamics of the game. In particular, if highly unequal participants get matched to each other, then competition suffers and the overall effect on the physical activity drops significantly. Furthermore, competitions with an equal mix of both men and women are more effective in increasing the level of activities. We leverage these insights to develop a statistical model to predict whether or not a competition will be particularly engaging with significant accuracy. Our models can serve as a guideline to help design more engaging competitions that lead to most beneficial behavioral changes.

Motivation & Objective

• To understand how gamified walking competitions in mobile health apps affect physical activity levels across diverse user demographics.
• To identify game design elements that maximize user engagement and physical activity increases during competitions.
• To develop a predictive model capable of forecasting the engagement level of future competitions based on user and competition features.
• To examine the impact of participant composition—such as gender balance and activity level similarity—on competition dynamics and outcomes.
• To provide actionable design guidelines for mobile health applications aiming to leverage competition to promote long-term physical activity.

Proposed method

• Collected and analyzed anonymized step-count data from 2,500 walking challenges in the Azumio Argus app over one year, covering 800,000 person-days.
• Used baseline physical activity levels (outside competitions) and in-competition activity patterns to compute daily step increases and rank changes.
• Constructed statistical models using demographic data (age, gender, BMI), prior competition behavior, and activity level similarity to predict competition engagement.
• Evaluated model performance using ROC AUC for key outcomes: FIRST-LAST rank difference, RANK SWAPS, and ΔACTIVITY (change in activity level).
• Compared models with varying feature sets (e.g., only demographics vs. full feature sets) to identify minimal yet effective input for real-world deployment.
• Validated model robustness by showing high predictive accuracy even with limited features, such as only outside competition activity levels or basic demographics.

Experimental results

Research questions

• RQ1 To what extent do walking competitions in mobile health apps increase physical activity across diverse user groups?
• RQ2 How do participant composition factors—such as gender balance and pre-competition activity level similarity—affect competition dynamics and user engagement?
• RQ3 Can we predict the level of engagement or success in a competition using only basic user and competition features?
• RQ4 What is the relationship between a user’s prior competition behavior and their performance and activity increase during a new competition?
• RQ5 How does the magnitude of physical activity increase vary across different user subgroups, including inactive users and those with varying BMI or age?

Key findings

• On average, users increased their daily physical activity by 23% during walking competitions compared to their baseline levels.
• The largest increases in physical activity—over 2,500 steps per day—were observed among users who were previously inactive, indicating strong potential for behavior change in low-activity populations.
• Winners of competitions increased their activity by 40–60% compared to baseline, while last-ranked users were on average less active than before the competition.
• Competitions with highly unequal pre-competition activity levels among participants showed significantly reduced overall physical activity increases, highlighting the need for balanced matching.
• Competitions with a balanced mix of men and women were more effective in increasing physical activity than same-gender groups.
• A statistical model using only prior outside competition activity levels achieved 70.7% ROC AUC for predicting the final rank difference (FIRST-LAST), and a model incorporating both outside and in-competition history reached 75.1% AUC, demonstrating strong predictive power for engagement outcomes.