Countering Hindsight Bias and Memory Decay During Debriefing Through Dual-Track Simulation Game Design

This paper investigates whether Concurrent Decision Logging (CDL), a practice in which learners record their reasoning and confidence before seeing each round's outcome in a simulation game, alters how they reason and later explain their decisions during debriefing. The authors built Lemonade Lane, a five-round browser-based business simulation with delayed word-of-mouth effects and mid-game shock events, and randomly assigned 41 participants to a research group (n=18), which logged rationale before outcomes, or a control group (n=23), which did not. The study is motivated by hindsight bias and memory decay, which may distort post-game accounts of why decisions were made.

Several findings reached statistical significance. Research players adjusted prices far more across rounds, with nearly three times the intra-player pricing variance of controls, and they responded to a mid-game competitor shock differently, with control players more likely to hold prices unchanged. In post-game attributions, loggers wrote about 53% more text, named more distinct causes, and used more causal connectors and self-evaluative language. Natural language analysis also showed that players who had logged operational reasoning later shifted toward counterfactual language such as "should have," a signature of hindsight bias that appeared in none of the pre-outcome logs. Log completion was complete and log length held steady until the final round.

The study found no difference in final cash balances or scores between groups, so CDL did not improve in-game performance. The authors conclude that CDL is a low-cost, scalable design intervention that makes reasoning visible and contestable during debriefing. The evidence is limited by a single game, a single-player browser format, and a small sample, and the authors acknowledge that the effect of showing logs to learners on actual learning outcomes remains untested.