Abstract
Sackett et al. (2025) report that deep-level diversity (goal diversity + personality diversity) relates more negatively to initial venture viability than surface-level diversity (age, gender, ethnicity), with this effect amplified in familiar teams. When familiarity is high, deep-level diversity devastates outcomes, while showing no effect when familiarity is low. The authors conclude that “team familiarity accelerates (and exacerbates) the negative effects of deep-level diversity” (p. 11) because familiar teams discover their psychological differences faster. I propose an alternative interpretation: their “deep-level diversity” may reflect formation-contingent misalignment rather than stable pre-existing attributes that teams simply discover over time. Under this view, what appears as diversity could signal problems in how teams are formed rather than immutable individual differences.
This commentary develops two distinct but related challenges. First, a measurement issue: goal ratings and personality assessments may be time-sensitive and relationally constructed, meaning what the authors measure as “deep-level diversity” at their study timepoint may not represent the pre-existing attributes relevant at team formation. Goals, in particular, are negotiated through social interaction (Locke & Latham, 2002); even personality traits, while more stable, require behavioral translation to manifest as work style incompatibility—meaning Big Five differences may not predict team dysfunction without evidence that these differences actually affected interaction patterns; without measuring whether teams discussed objectives before forming or how goals shifted during the program, the authors cannot determine if their diversity scores reflect stable individual differences versus formation-phase misalignment. Second, a mechanism issue: even if differences existed prior to formation, familiarity may change how teams interpret those differences—not through accelerated discovery, but through violated expectations about what familiar relationships entail (Morrison & Robinson, 1997). These interpretations suggest different interventions than the authors propose: rather than encouraging post-formation discussion, platforms might implement pre-formation goal negotiation protocols, particularly for familiar dyads where implicit assumptions may substitute for explicit alignment. Adjudicating between the authors’ discovery account and this formation-artifact account requires evidence that the focal study did not collect: direct measurement of whether goals were discussed before teaming, reasons teams formed, and whether psychological attributes shifted during the program.
The Reverse Causality Problem: When “Diversity” Measures Selection Failure
Sackett et al. treat deep-level diversity as an independent variable—a characteristic of individuals that exists prior to team formation and gets “discovered” through interaction. They measure goal diversity by asking participants to rate seven venture goals (learning, mentorship, cofounder search, idea validation, venture launch, funding) and then calculating profile similarity across team members (p. 15). A team where one member rates “to launch a venture” as 7/7 and another rates it 2/7 gets coded as high goal diversity. The authors assume these are stable individual goals that create coordination problems when misaligned.
This assumption ignores that goals are relationally constructed, not individually pre-determined (Locke & Latham, 2002; Oettingen et al., 2015). When a familiar team shows high goal diversity, this does not mean they “discovered” misalignment—it means they formed a team while bypassing the conversation that trusted partners normally have before committing to a joint venture. The “diversity” is the symptom of broken formation, not the cause of subsequent failure. The authors’ own homophily logic makes this clear: entrepreneurs preferentially select “family, friends, or former coworkers” (p. 10) based on similarity (McPherson et al., 2001; Ruef et al., 2003). If homophily drives selection, then familiar teams with severe psychological differences represent a theoretical anomaly—they violate the very principle that should have prevented their formation.
Several formation dynamics could produce this pattern. Familiar teams with high diversity may reflect asymmetric dependence (one member needs the other more than vice versa), convenience partnering (teaming with available friends rather than optimal matches), social pressure to join a friend’s struggling idea, or friendship types that provide social support but lack foundation for venture collaboration—what Ingram and Zou (2008) distinguish as activity friendships versus intimate friendships. Effective teams typically engage in explicit goal negotiation during formation (Hackman & Wageman, 2005; Mathieu et al., 2008); high diversity in familiar teams may measure the absence of this negotiation. This explains why unfamiliar teams show no diversity effect—strangers negotiate explicitly because they lack a prior relationship to leverage. Goal diversity in unfamiliar teams reflects differences the team accepted and planned to manage; in familiar teams, it may reflect a formation that assumed prior friendship guaranteed alignment. The practical implication depends on the formation context: in some cases, where teams formed through convenience or pressure rather than genuine partnership alignment, dissolution may be appropriate; in others, where teams show willingness to renegotiate roles and objectives, structured coaching could help bridge initial misalignment.
Familiarity as Diagnostic, Not Moderator: What “Acceleration” Actually Reveals
Sackett et al. frame familiarity as a moderator that “accelerates the discovery of misaligned goals and personality differences” (p. 11), contrasting this with their rejected hypothesis that familiarity might mitigate diversity through trust (H3a). The interaction pattern—devastating effects in familiar teams (B = −3.03, p < .001), null effects in unfamiliar teams (B = −1.08, p = .33)—leads them to conclude familiarity makes diversity worse by speeding up when teams learn about their differences.
This interpretation misidentifies what familiarity reveals. Familiarity does not make pre-existing diversity more harmful—it exposes whether “diversity” represents normal variation versus pathological misalignment. When unfamiliar teams discover goals or personality differences, they lack the context to judge whether these differences are problematic. A stranger learning that their new cofounder prioritizes “learning” over “launching” has no baseline: maybe this is normal entrepreneurial variation, maybe it’s a red flag. Research on transactive memory systems shows that unfamiliar partners lack the metacognitive knowledge to evaluate whether differences represent complementarity versus incompatibility (Lewis & Herndon, 2011). An unfamiliar team displaying what the authors code as “high personality diversity” based on Big Five scores might simply reflect measurement noise—they don’t yet know each other well enough to determine if personality profiles translate into incompatible work styles.
Familiar teams possess this context. When long-time friends or former colleagues discover through the incubator program that they have divergent venture goals, they recognize this as anomalous. The shock is not “we have different goals” but “how did we agree to start a company together without realizing we have different goals?” This constitutes a psychological contract violation—a breach of implicit expectations about what the relationship entails (Morrison & Robinson, 1997; Rousseau, 1995). Such violations trigger strong negative emotional reactions and withdrawal behaviors that exceed the instrumental harm of mere goal misalignment. The authors’ data show a pattern consistent with this interpretation: diversity has no effect when familiarity is low but severe negative effects when familiarity is high. This discontinuous pattern suggests familiarity changes what diversity means, not just when it gets discovered. If familiarity merely accelerated discovery timing, we would expect diversity’s negative effect to strengthen gradually as familiarity increases, rather than showing this qualitative shift between low and high familiarity.
The consequence is that familiarity should not be studied as a moderator of diversity effects but as a diagnostic of whether measured diversity reflects meaningful individual differences or team dysfunction. This challenges how moderation is theorized in entrepreneurial team research more broadly. The field often treats moderators as variables that strengthen or weaken fixed causal effects—familiarity “accelerates” discovery, task interdependence “amplifies” conflict, or time pressure “magnifies” coordination costs. But when team composition emerges through voluntary selection, moderators may instead indicate qualitatively different underlying processes. High familiarity + high diversity does not represent “the same diversity discovered faster”—it represents a fundamentally different phenomenon (formation failure) than low familiarity + high diversity (accepted differences). Meta-analytic evidence supports this interpretation: diversity effects vary substantially across contexts, with particularly inconsistent findings in entrepreneurial versus organizational settings (Triana et al., 2021). In organizational teams where composition is assigned, diversity truly represents individual differences; moderators genuinely moderate. In entrepreneurial teams where composition is chosen, “diversity” conditional on formation may primarily capture formation failure; “moderators” diagnose which type of diversity researchers actually measured. This means interaction effects in voluntary team contexts may require fundamentally different interpretation than in assigned team contexts.
Implications: Toward a Formation-Contingent Theory of Diversity
If deep-level diversity in familiar teams reflects formation-contingent misalignment rather than stable individual differences, research design and measurement must change before intervention. First, studies must distinguish diversity as an attribute from diversity as formation byproduct by measuring formation processes: Did teams form through active mutual selection versus passive assembly? Did they negotiate roles and goals during formation (Wageman et al., 2012)? Researchers should not simply control for familiarity as a covariate—they should use it to split samples and test whether diversity operates through different mechanisms in familiar versus unfamiliar teams. Goal diversity that emerges despite explicit negotiation differs fundamentally from goal diversity reflecting absent negotiation, yet current measurement treats them identically. Second, research should examine why certain familiar dyads form high-diversity teams: Do these represent asymmetric power relationships, or friendship types lacking a foundation for venture partnership Third, having established these measurement distinctions, interventions should target formation processes. The authors’ incubator context featured a “networking event” in which “teams form organically around the most popular ideas” (p. 13)—a process that may inadvertently create convenience partnerships in which less attractive ideas recruit reluctant familiar others. Platform designers might require explicit goal alignment discussions before teams can officially form, particularly for familiar dyads where implicit assumptions may be strongest. For teams that have already formed with misalignment, the appropriate intervention depends on venture stage and goal reconcilability: early-stage ventures with low switching costs and irreconcilable differences may benefit from dissolution, while teams showing willingness to renegotiate roles could benefit from structured facilitation.
Sackett et al. have surfaced a crucial empirical pattern: familiar teams with deep-level diversity experience severe initial venture viability problems. Their interpretation—that familiarity accelerates discovery of pre-existing differences—represents one plausible account. An alternative interpretation is that what appears as “deep-level diversity” in familiar teams may reflect formation-contingent processes: violated expectations, absent negotiation, or convenience partnering that substitutes friendship for strategic alignment. These accounts lead to different research priorities. The discovery account suggests studying how quickly teams learn about differences; the formation-artifact account suggests studying how teams form and whether measured diversity reflects stable attributes or emergent misalignment. Distinguishing these mechanisms matters because they imply different interventions—accelerating discovery versus preventing problematic formations—and different boundary conditions for when deep-level diversity predicts team outcomes in voluntarily-formed entrepreneurial contexts.
