Do transactive memory systems predict team performance better than shared mental models?

Two different kinds of team knowledge

A shared mental model is what it sounds like: team members hold similar understandings of the task and of how the team works. A transactive memory system is different—it is a division of labor for knowledge, where members specialize, know who knows what, and can retrieve information from each other [3]. One is about thinking alike; the other is about knowing who to ask.

These are not competing descriptions of the same thing, which is why asking which one predicts performance better is partly a question about which team outcome you care about. The studies here tend to examine them separately rather than pitting them head-to-head.

What each contributes to performance

For transactive memory, a seven-wave longitudinal study of project teams found that shared leadership improved team performance, and that the transactive memory system mediated this relationship [3]. The effect of shared leadership on performance was stronger early in the team's life cycle, again with transactive memory mediating that conditional effect [3]. This suggests that a well-functioning system of distributed expertise is one route through which leadership translates into results, particularly when teams are new.

For shared mental models, a study of 266 individuals across 48 teams in Italian manufacturing companies found that the degree to which individuals aligned their task- and team-related mental models affected both their tendency to proactively solve problems and their ability to coordinate team problem-solving [4]. Those proactive behaviors and coordination, in turn, were linked to enhanced team performance [4]. The authors also note that shared mental models can lead to groupthink, potentially diminishing proactive problem-solving—so alignment is not uniformly beneficial [4].

When both are in play at once

One study of temporary project teams found that both transactive memory and shared mental models mediated the relationship between acute stress and team performance [2]. That is consistent with the idea that the two constructs can operate together rather than one simply outperforming the other as a predictor.

In human–AI teams, an experiment with 128 participants found that high knowledge complementarity enhanced team effectiveness and member satisfaction, with shared mental models acting as a mediating mechanism and a positive team climate amplifying the effects [5]. This study is framed around shared mental models rather than a direct comparison with transactive memory systems, so it speaks to the value of shared understanding in a newer kind of team rather than to which predictor is stronger.

Why the evidence does not crown a winner

The comparison is hard to settle because most studies examine one construct or the other, or examine both as parallel mediators rather than as competing predictors. One paper investigates both shared mental models and transactive memory systems in relation to team learning and effectiveness, including team performance [1], but its abstract does not report a direct comparison of their predictive strength.

What the evidence does support is that both are plausible levers. Transactive memory appears to carry the effect of shared leadership onto performance over time [3], and shared mental models appear to support the proactive problem-solving and coordination that accompany better performance [4]. Which one predicts performance better may depend on the team, the task, and the stage of the team's life—a question the current studies do not resolve.

Sources used in this answer

Related research questions