TY - CONF AB - Optimal decision making requires appropriate evaluation of advice. Recent literature reports that algorithm aversion reduces the effectiveness of predictive algorithms. However, it remains unclear how people recover from bad advice given by an otherwise good advisor. Previous work has focused on algorithm aversion at a single time point. We extend this work by examining successive decisions in a time series forecasting task using an online between-subjects experiment (N = 87). Our empirical results do not confirm algorithm aversion immediately after bad advice. The estimated effect suggests an increasing algorithm appreciation over time. Our work extends the current knowledge on algorithm aversion with insights into how weight on advice is adjusted over consecutive tasks. Since most forecasting tasks are not one-off decisions, this also has implications for practitioners. AU - Leffrang, Dirk AU - Bösch, Kevin AU - Müller, Oliver ID - 37312 KW - Algorithm aversion KW - Time series KW - Decision making KW - Advice taking KW - Forecasting T2 - Hawaii International Conference on System Sciences TI - Do People Recover from Algorithm Aversion? An Experimental Study of Algorithm Aversion over Time ER - TY - CONF AB - Despite the widespread use of machine learning algorithms, their effectiveness is limited by a phenomenon known as algorithm aversion. Recent research concluded that unobserved variables can cause algorithm aversion. However, the impact of an unobserved variable on algorithm aversion remains unclear. Previous studies focused on situations where humans had more variables available than algorithms. We extend this research by conducting an online experiment with 94 participants, systematically varying the number of observable variables to the advisor and the advisor type. Surprisingly, our results did not confirm that an unobserved variable had a negative effect on advice-taking. Instead, we found a positive impact in an algorithm appreciation scenario. This study provides new insights into the paradoxical behavior in which people weigh advice more despite having fewer variables, as they correct for the advisor's errors. Practitioners should consider this behavior when designing algorithms and account for user correction behavior. AU - Leffrang, Dirk ID - 50118 IS - 19 KW - Algorithm aversion KW - Data KW - Decision-making KW - Advice-taking KW - Human-Computer Interaction T2 - Wirtschaftsinformatik Conference TI - The Broken Leg of Algorithm Appreciation: An Experimental Study on the Effect of Unobserved Variables on Advice Utilization ER -