The Five Types of Team Training

A common error among practitioners new to team training is to treat “team training” as a single product. It is not. The team training literature distinguishes at least five well-developed approaches, each built for a different kind of team in a different kind of situation. The five are not interchangeable. Selecting the right type for the team in front of the practitioner is part of the design work, and choosing the wrong type produces predictably weak results regardless of how well the chosen type is executed.

This post describes the five types of team training that have the strongest evidence base, identifies the kinds of teams each type fits best, and offers a short selection heuristic.

Type 1: Cross-Training

Cross-training develops members’ understanding of one another’s roles, tasks, and informational needs. Members do not necessarily learn to perform each other’s jobs at expert level, but they develop enough working knowledge to anticipate what teammates need, recognize when a teammate is overloaded, and provide effective backup. The mechanism is improved coordination through better shared mental models of teammate roles (Volpe et al., 1996; Marks et al., 2002).

Cross-training is most useful for interdependent teams with specialized roles in which coordination depends on members understanding what others are doing and why. Surgical teams, special-operations units, rotating field-service teams, and product launch teams whose members come from different functions are typical fits. Cross-training is less useful for teams whose roles are nearly identical or whose work is loosely coupled, because the mental-model gain is smaller in those contexts.

Implementation typically involves three formats of increasing intensity: positional clarification (members read or hear about other roles), positional modeling (members observe other roles in real or simulated work), and positional rotation (members briefly perform other roles). The intensity required scales with the interdependence of the team.

Type 2: Team-Interaction Training

Team-interaction training (sometimes called team coordination training) provides direct instruction and practice on the interpersonal and procedural behaviors that drive coordination: closed-loop communication, mutual performance monitoring, backup behavior, conflict management, and adaptive role-switching (Salas et al., 2008). The mechanism is direct skill acquisition; the curriculum names specific behaviors and practices them.

Team-interaction training is most useful for newly formed teams, teams entering novel environments, and teams whose members come from different professional cultures with different default coordination norms. It is the most commonly delivered form of team training because its content is broadly applicable, but its impact depends heavily on whether the four-step minimum (Information, Demonstration, Practice, Feedback) is honored. When team-interaction training is reduced to lecture and discussion, it converges with generic communication training and loses most of its team-specific value.

Type 3: Guided Team Self-Correction

Guided team self-correction trains a team to debrief itself. The curriculum teaches the team a structured debriefing protocol, anchored on behavioral markers, and gives the team supervised practice running its own debriefs. Over repeated cycles, the facilitator’s role diminishes and the team takes increasing responsibility for analyzing its own performance and planning its next attempt (Smith-Jentsch et al., 2008).

This type is most useful for mature teams that cannot stop working for an extended workshop but need a sustained capacity for continuous improvement. Surgical teams, emergency response crews, and high-tempo project teams often fit this profile. The advantage is that the training builds a self-renewing learning capability rather than a finite set of skills; the team continues to develop after the formal training has ended. The disadvantage is that the type requires a baseline of psychological safety and a leader willing to share evaluative authority with the team, which not every team has.

Type 4: Crew Resource Management (CRM)

Crew Resource Management originated in commercial aviation in the late 1970s and has since been adapted for surgery, emergency medicine, military operations, nuclear operations, and other high-reliability domains (Salas et al., 2006). CRM is not a single curriculum but a family of programs unified by a shared structure: high-fidelity simulation, behaviorally-anchored facilitation, and an explicit focus on the non-technical skills that distinguish safe high-reliability teams from unsafe ones (situation awareness, decision-making under uncertainty, leadership, communication, teamwork, and management of stress and fatigue).

CRM is most useful for high-stakes teams in domains where coordination failures cause catastrophic harm. The defining property is the use of isomorphic simulation (training scenarios that closely mirror the real operational context) combined with rigorous post-scenario debriefs. Adaptations for healthcare, such as TeamSTEPPS, draw heavily on CRM principles. CRM is expensive and time-intensive, which limits its use in lower-stakes settings where a less elaborate type would produce sufficient gains.

Type 5: Metacognition Training

Metacognition training develops a team’s capacity to think about its own thinking: to notice when it is working from incomplete information, to recognize when a member’s expertise should override the team’s default decision process, and to question its own framing of a problem before committing to a course of action. The mechanism is a higher-order cognitive skill: not solving the problem, but evaluating how the team is approaching the problem (Cohen et al., 1996; Salas et al., 2008).

Metacognition training is most useful for high-stakes decision teams whose work is dominated by judgment under uncertainty: executive teams, intelligence analysis teams, mergers and acquisitions teams, clinical diagnostic teams in complex cases. The training typically uses scenario-based exercises with explicit prompts that interrupt the team’s reasoning and force surfacing of assumptions and alternative framings. It is among the most demanding types to facilitate well, because the facilitator must intervene at the cognitive level rather than the behavioral level, which requires more skill than simply naming observed behaviors.

A Short Selection Heuristic

A practitioner choosing among the five types can begin with three diagnostic questions.

What kind of breakdowns is the team experiencing? If the breakdowns center on members not understanding what others are doing, the type is cross-training. If they center on basic coordination behaviors, the type is team-interaction training. If they center on the team’s inability to learn from its own experience, the type is guided self-correction. If they center on high-stakes coordination failures with catastrophic consequences, the type is CRM. If they center on poor judgment under uncertainty, the type is metacognition training.

What is the team’s developmental stage? Newly formed teams typically benefit most from team-interaction training and cross-training. Mature teams typically benefit more from guided self-correction or metacognition training. CRM is appropriate at any stage when the operational context demands it.

What does the operating context allow? Teams that cannot stop working need guided self-correction or short embedded practice rather than extended workshops. Teams in high-fidelity simulation environments can use CRM. Teams without simulation budgets but with stable composition can use team-interaction training.

A Note on Combination

The five types are not mutually exclusive. Many effective programs combine elements of two or more (for example, cross-training within a CRM curriculum, or guided self-correction layered onto an initial team-interaction training base). The goal is not to pick exactly one; it is to pick the primary type that matches the team’s most pressing gap and to add elements from other types where the diagnosis indicates. The most common error is the inverse: deploying a single favored type on every team the practitioner encounters, regardless of fit.

Selection is the first design decision, and it is among the most consequential. A well-executed cross-training program delivered to a team that needed metacognition training will produce far less change than a moderately-executed metacognition program would have produced for the same team.