How team forgiveness climate influences innovative performance in college student innovation teams: a moderated mediation model Authors: Honghao Hu, Chao Li, Jonas Valkonen, Haoluan Fu Published: 2026 ABSTRACT AbstractBy integrating team design theory and organizational learning theory within an input–process–output (IPO) framework, this study examines the relationships among team forgiveness climate, team learning from mistakes and team innovative performance. We propose a moderated mediation model in which team learning from mistakes serves as the mediating mechanism and team shared vision functions as a strategic boundary condition. Survey data were collected from 145 college student innovation teams from a comprehensive university in East China. The results indicate that team forgiveness climate is positively associated with team innovative performance, and team learning from mistakes mediates this relationship. Team shared vision positively moderates the relationship between team forgiveness climate and team learning from mistakes. Furthermore, when team shared vision is high, team learning from mistakes significantly mediates the relationship between team forgiveness climate and team innovative performance. These findings extend the application of forgiveness research to the team level within student innovation contexts and offer preliminary practical insights for fostering innovation in student-led research projects.Keywords: Team forgiveness climate, Team shared vision, Team learning from mistakes, Student innovation teams Introduction Under China’s current innovation-driven development strategy, college student research and innovation teams, such as those participating in Student Research Training Program (SRTP), have become important platforms for cultivating innovative talent [1]. These teams are characterized by high levels of exploration, interdisciplinarity, and inherent uncertainty [2]. Consequently, during technical breakthroughs and iterative solution development, members inevitably encounter various forms of mistakes and setbacks, including experimental deviations, communication breakdowns, and collaborative inefficiencies [3–5]. Therefore, enabling student innovation teams to manage mistakes effectively is critical for enhancing their innovative performance [6]. However, as recent scholarship suggests, mistakes per se do not necessarily diminish innovative potential; instead, their impact on innovation depends on whether teams can transform mistakes into intellectual resources via learning mechanisms that foster resilience and knowledge transformation [7–9]. Student innovation teams often lack prior teamwork and research experience, as members may be engaging in complex tasks collectively for the first time [10, 11]. Previous studies have noted that such teams are more prone to interpersonal stress or risk aversion triggered by mistakes and failures [12, 13]. Consequently, a core imperative for elevating the quality of college innovation training is the cultivation of a forgiveness climate, defined as a collective perception that benevolent responses to mistakes and failures are supported [14, 15]. Such a climate is inherently restorative and relational, centering on the collective capacity to repair interpersonal relationships and emotional states following a transgression or failure [16, 17]. As a shared social norm, this climate reduces defensive group-based reactions and enhances constructive engagement in collective settings [18, 19]. Despite the recognized benefits of a forgiveness climate, its application within team dynamics remains underexplored, particularly in the context of college student innovation [20]. This study aims to address three key gaps in the extant literature. First, while previous scholarship has examined forgiveness at the individual or organizational level [21, 22], the team as a distinct collective unit has been relatively overlooked. Recent evidence [23, 24] indicates that team innovation is shaped strongly by collective internal perceptions and team-level interaction patterns, rather than only by broad institutional policies. Second, the pathway linking forgiveness climate to innovation remains inadequately explained [17]. Drawing on team design theory [25, 26], we introduce team learning from mistakes, a collective process through which teams identify, analyze, and respond to mistakes [27–29], as a pivotal mediator within the IPO framework: forgiveness climate as the input activates team learning from mistakes as the process, which in turn contributes to team innovative performance as the output. This aligns with the view that a forgiveness climate may facilitate proactive learning [30] and encourage teams to treat mistakes as potential sources of new knowledge and strategic value, as acknowledging mistakes can unlock new knowledge and strategic value [7, 8]. Third, the boundary conditions of team forgiveness climate remain understudied. Scholars caution that forgiveness without strategic guidance may lead to complacency or new harm [17, 20, 31], yet little is known about what can act as a strategic filter to mitigate such risks [32], especially in uncertain student innovation contexts where teams must prioritize mistakes with functional value [2]. Organizational learning theory posits that a shared vision serves as a vital learning orientation, directing members’ attention toward effective information screening [33–35]. Thus, we propose team shared vision as a moderator that may channel forgiveness toward intelligent mistakes aligned with long-term goals [36], leading to a moderated mediation model (see Fig. 1). The theoretical construct model This study makes three main theoretical contributions. First, we introduce team forgiveness climate into the collegiate innovation context, extending its application from individual and organizational levels to team-level dynamics. Although this does not constitute a novel construct, we position team forgiveness climate as a collective resource in a setting where its role remains empirically underexplored. Second, by empirically testing the mediating role of team learning from mistakes, we provide preliminary evidence for a learning-based pathway linking team forgiveness climate to innovation. While forgiveness climate shares conceptual space with related constructs such as psychological safety and error management climate, our contribution lies in integrating these theoretical perspectives within the specific and uncertain environment of student innovation teams. Finally, this study addresses the boundary conditions of team forgiveness climate by identifying team shared vision as a moderating condition, consistent with organizational learning perspectives suggesting that supportive climates may be more effective when paired with clear collective direction. Theory and hypotheses Team forgiveness climate and team innovative performance Recent scholarship has suggested that forgiveness can transcend individual dispositions to become a normative social process at the collective level [18, 19, 37]. Specifically, Borinca et al. [18] found that normatively supported apologies and forgiveness-related processes can facilitate positive interaction and social reintegration, suggesting that these mechanisms function as shared social norms rather than isolated interpersonal acts. This perspective aligns with organizational behavior research, which regards forgiveness climate as a shared emotional culture that shapes team dynamics [38]. Accordingly, we define team forgiveness climate not as a mere aggregate of individual benevolence, but as a normatively structured group-level resource. Within the IPO framework, this climate serves as a contextual input that may preserve the team’s cognitive and emotional resources. By reducing the social and reputational costs of mistakes, it establishes a relational repair mechanism necessary for adaptive and collaborative experimentation, which is important for team innovative performance. Contemporary research suggests that judgmental and punitive responses to mistakes often inhibit learning and adaptation, whereas individuals are more likely to grow in environments free from reprisals for errors [3, 8]. Similarly, recent evidence suggests that supportive relational contexts can promote creativity through psychological safety, further indicating that climates that reduce interpersonal threat are conducive to creativity-related outcomes [39]. It is important to note that team forgiveness climate is conceptually distinct from psychological safety, tolerance for mistakes, and error management climate in terms of temporal orientation and functional domain. Specifically, psychological safety [3] reflects a shared belief that the team is safe for interpersonal risk-taking before members speak up or act. By contrast, team forgiveness climate is a post-hoc relational repair mechanism that focuses on extending collective benevolence after a transgression or mistake has occurred or has been attributed to a team member. While tolerance for mistakes [40] and error management climate [41] emphasize cognitive strategies and task-level error correction, forgiveness climate uniquely prioritizes restoring interpersonal bonds and the reduction of negative affect. Unlike the anticipatory orientation of psychological safety or the task-oriented focus of error handling, team forgiveness climate represents a reactive, emotional, and social process that transforms the aftermath of a mistake from a cycle of blame and anger into one of mutual acceptance and psychological recovery, thereby preserving the team’s social fabric during the mistakes and setbacks that may occur in innovation. We contend that team forgiveness climate may contribute to team innovative performance for three reasons. First, forgiveness climate provides collective benevolence and an emotional buffer that mitigates the perceived risks of exploration. Because innovation is inherently uncertain, mistakes are often unavoidable parts of the research process. Chen et al. [23] found that team forgiveness climate encourages voice behavior, which is associated with better team performance than a blame culture. Unlike professional R&D teams with established protocols, college student teams often need to develop their own interaction norms during collaboration [2]. Therefore, the initial forgiveness climate may influence whether teams engage in adaptive persistence or succumb to premature abandonment following early setbacks [16]. Second, a forgiveness climate facilitates a cognitive shift from individual culpability to collective learning. By decoupling mistakes from professional identity, teams may be more willing to treat intelligent failures as learning opportunities rather than mere losses [9, 20]. This orientation enables more transparent identification and analysis of mistakes, which in turn supports the iterative cycles required for complex innovation. Third, mistakes often trigger interpersonal conflicts that deplete cognitive resources and hinder innovation. A forgiveness climate serves as a relational repair mechanism that fosters mutual trust and organizational identification [15, 30]. By discharging the anger and blame that typically restrain collaboration, a forgiveness climate may help teams sustain the long-term, coordinated effort required for innovation. Therefore, we hypothesize: Hypothesis 1: Team forgiveness climate is positively associated with team innovative performance. Hypothesis 1: Team forgiveness climate is positively associated with team innovative performance. The mediating effect of team learning from mistakes We integrate team design theory [25, 26] and organizational learning theory [33, 35] within an IPO framework to explain the learning-based process of team innovation. In this framework, team forgiveness climate serves as an important contextual input, or enabling condition, that may alleviate threat rigidity and reduce the psychological costs associated with mistakes [16]. By providing a relational repair mechanism that helps transform experiential failure into technical knowledge [42], this input may guide team interaction processes away from defensive blame and toward learning from mistakes, thereby supporting team innovative performance. We propose that team forgiveness climate may promote team learning from mistakes for three reasons. First, it facilitates a cognitive reappraisal of mistakes by decoupling them from blame and interpersonal retaliation. In such a climate, team members may be more likely to maintain an objective stance toward mistakes and perceive them as informative signals rather than personal failures. By reappraising mistakes as learning opportunities rather than threats, members can mobilize greater cognitive and emotional resources for the team [43]. This reappraisal process may enable team members to treat mistakes as valuable sources of new knowledge [27]. Second, a forgiveness climate can serve as an emotional regulation mechanism that shifts team energy from blame to reflection. Mistakes often trigger negative emotions, such as resentment or anxiety, which can deplete cognitive resources within a team [5]. A forgiveness climate may help normalize mistakes and sustain positive collective affect [44]. Instead of devoting energy to scapegoating, team members may focus more on extracting meaning from mistakes. Recent evidence also indicates that forgiveness climate can foster positive organizational attitudes; for example, it has been found to positively predict nursing staff’s organizational identification [30]. Third, because mistakes can strain interpersonal trust and threaten team cohesion, a forgiveness climate may function as a relational repair mechanism [42]. This function is particularly relevant in student innovation teams, where social bonds are still developing and may be vulnerable to interpersonal animosity following failure [22]. By decoupling technical errors from relational friction, a forgiveness climate may support reflective learning, help repair damaged working relationships, and sustain long-term collaborative momentum. Therefore, we hypothesize: Hypothesis 2: Team forgiveness climate is positively associated with team learning from mistakes. Hypothesis 2: Team forgiveness climate is positively associated with team learning from mistakes. Empirical research has identified team learning from mistakes as an important factor associated with improved teamwork processes and performance [27–29]. Across various high-stakes domains, learning from mistakes has been positively associated with collective effectiveness [9, 32]. For instance, in healthcare settings, learning from errors has been linked to lower error rates and improved patient safety [45, 46]. By treating mistakes as potential learning resources, teams may refine their knowledge structures and reduce uncertainty in exploratory work [3]. Research on high-tech ventures also suggests that failure-based learning can support innovative success by helping teams transform setbacks into strategic knowledge [9]. For student innovation teams, this learning process may be particularly relevant, as it allows members to move beyond the emotional burden of setbacks and focus on the technical and procedural improvements needed for innovation [1]. By analyzing mistakes and deriving actionable insights, teams may iterate their innovation processes more effectively and generate stronger innovative outcomes. Consequently, we propose: Hypothesis 3: Team learning from mistakes is positively associated with team innovative performance. Hypothesis 3: Team learning from mistakes is positively associated with team innovative performance. Building on this foundation, we argue that in student innovation teams, team forgiveness climate may provide a relational repair mechanism that supports learning from mistakes. By reducing the psychological costs and interpersonal risks associated with mistakes, this climate may encourage team members to discuss and reflect on mistakes rather than suppress them. In this way, team forgiveness climate as the input may facilitate team learning from mistakes as the process, which in turn may contribute to team innovative performance as the output. We therefore hypothesize: Hypothesis 4: Team learning from mistakes mediates the relationship between team forgiveness climate and team innovative performance. Hypothesis 4: Team learning from mistakes mediates the relationship between team forgiveness climate and team innovative performance. The moderating role of team shared vision Team forgiveness climate is not unconditionally beneficial. Excessive forgiveness may foster complacency and accountability neglect, causing members to overlook the adverse consequences of their mistakes [4, 16, 20]. Thus, its constructive value may depend on a strategic filter that aligns benevolence with collective goals [17, 20, 31]. Bosk’s [47] research on medical teams illustrates that forgiveness may function as a selective tool: certain technical and judgmental mistakes can be tolerated to facilitate learning, whereas normative or irresponsible errors should not be treated in the same way. In student innovation projects, which are characterized by uncertainty and developing social bonds [10], such selectivity is particularly important. Some mistakes, including technical rework, insufficient literature reviews, and immature research designs, may stem from limited research experience and provide opportunities for learning and innovation. By contrast, preventable mistakes, such as free-riding and procrastination, are more likely to arise from attitudinal lapses or poor internal coordination rather than exploratory risk [11, 48]. Distinguishing between these two types of mistakes is important because not all mistakes offer equal learning value. As organizational learning theory posits, learning is goal-directed, and teams use information to refine collective action [33, 35]. Team shared vision, defined as a collective mental model of a team’s future state and direction, may guide how teams process experiential information. Although overly rigid accountability may restrain innovation under some conditions [49], shared vision may help teams distinguish intelligent mistakes from preventable errors and thereby direct attention toward mistakes with learning value. We propose that team shared vision provides a teleological framework that may help translate team forgiveness climate into team learning from mistakes. First, shared vision may act as a strategic filter for identifying intelligent mistakes. Effective innovation management requires differentiating preventable errors from intelligent failures; the former stems from deviance or inattention, whereas the latter may yield substantive new knowledge [7, 8]. When shared vision is high, student teams may be better able to focus their limited cognitive resources on analyzing mistakes with high learning potential for their research objectives. In this way, the collective benevolence and mutual acceptance provided by a forgiveness climate may be directed toward productive learning. As Kantabutra [36] observes, shared vision is an important predictor of effective team learning because it helps establish criteria for what information is worth processing. Second, shared vision may increase responsiveness and reduce internal friction. Sitkin [50] argues that effective learning requires alacrity, characterized by a rapid and adaptive response to failure. In teams with a high level of shared vision, members may spend less time on interpersonal blame or debating the meaning of a setback, because the common goal provides a clearer behavioral framework for recovery [34]. This consensus-driven environment may facilitate the transition from emotional distress to analytical reflection. Finally, shared vision may mitigate opportunism and enhance collective responsibility [36, 51]. In student teams that often lack formal accountability systems, forgiveness climate may create a risk of individual shirking of responsibility or apathy toward quality [20]. However, a strong shared vision may foster collective ownership and altruistic behavior [52], making it more likely that forgiveness is accompanied by responsibility rather than opportunism. As Tao et al. [32] suggested, learning from mistakes may be more transformative when guided by a strategic orientation that aligns individual efforts with common innovative goals. Conversely, when team shared vision is low, a forgiveness climate may be less likely to translate into productive learning. Without a clear direction, teams may find it difficult to distinguish high-value mistakes from preventable errors. In such contexts, a benevolent atmosphere may lack a responsibility anchor, leading to inefficient discussion or responsibility-shirking. Without shared vision, the safety provided by a forgiveness climate may not be sufficient to support the systematic learning required for innovation. Therefore, the following two hypotheses are proposed: Hypothesis 5: Team shared vision moderates the relationship between team forgiveness climate and team learning from mistakes, such that this relationship is stronger when team shared vision is high. Hypothesis 5: Team shared vision moderates the relationship between team forgiveness climate and team learning from mistakes, such that this relationship is stronger when team shared vision is high. Hypothesis 6: Team shared vision strengthens the indirect relationship between team forgiveness climate and team innovative performance through team learning from mistakes, such that this indirect relationship is stronger when team shared vision is high. Hypothesis 6: Team shared vision strengthens the indirect relationship between team forgiveness climate and team innovative performance through team learning from mistakes, such that this indirect relationship is stronger when team shared vision is high. Method Sample This study used a multi-wave questionnaire survey, and data were collected from student innovation teams participating in the Student Research Training Program (SRTP) at a comprehensive university in East China. Students from two broad disciplinary categories, namely science and engineering and humanities and social sciences, were surveyed. Student innovation teams were identified according to the following criteria: (1) the project involved innovative tasks, such as technology research and development, new product design, solution development, or business plan writing; (2) the team consisted of two or more members; and (3) the team completed the innovation project through collaboration. Data collection was designed to align with the lifecycle of the SRTP and spanned six months, from November 2024 to May 2025. To reduce common method bias and capture the developmental nature of team interaction, the survey was conducted in two stages corresponding to official project milestones. The first stage was conducted during the university’s official mid-term evaluation period. At this point, teams had been collaborating for several months and had encountered both preliminary progress and project setbacks. This stage was appropriate for measuring team forgiveness climate, team shared vision, and team learning from mistakes, because these team-level perceptions and learning-related evaluations had begun to develop through members’ initial project interactions and experimental iterations. Surveys were distributed to 180 teams with the assistance of staff from the University Youth League Committee. To ensure data integrity and confidentiality, questionnaires were sealed in envelopes with clear confidentiality instructions. Team leaders collected these envelopes and distributed them to other team members, who provided evaluations of team climate and learning processes. The second stage took place during the official final evaluation and defense phase after project completion. Rather than relying on self-reported performance data, this study used externally assessed performance scores as the outcome variable. These scores were assigned by an independent university expert panel during the final review. The panel evaluated the teams based on standardized criteria, including innovation quality, technical complexity, and project deliverables. The scores were retrieved from the official academic affairs websites of the respective colleges, thereby enhancing the independence and credibility of the team innovative performance measure. Of the 660 questionnaires distributed, 531 were returned and 430 were deemed usable for analysis. The final sample comprised 145 teams, with an average of 3 members per team and a team size ranging from 2 to 4 members. Among the participants, 72.8% were male and 27.2% were female. In terms of academic discipline, 80.6% of the participants came from science and engineering backgrounds, whereas 19.4% came from the humanities and social sciences. Reflecting the target population of the SRTP, the participants were primarily upper-level undergraduates: third-year students accounted for 82.9% of the sample, followed by fourth-year students at 15.6%, while first- and second-year students together accounted for 1.5%. All members within each team came from the same academic discipline. The disciplinary backgrounds of the teams were grouped into two categories: Science and Engineering, and Humanities and Social Sciences. At the team level, 80.7% of the teams were classified as Science and Engineering, and 19.3% were classified as Humanities and Social Sciences. Measures Team forgiveness climate was measured using a four-item scale developed by Cox [14]. A sample item was “In the team, we can tolerate team members’ faults and mistakes”. Cronbach’s alpha for this scale was 0.711. Team shared vision was measured using a four-item scale adapted from Wang and Rafiq [53]. A sample item was “There is a strong sense of where this team is going”. Cronbach’s alpha for this scale was 0.639. Team learning from mistakes was measured using a six-item scale developed by Tjosvold et al. [29]. A sample item was “Mistakes are very useful for improving our team’s work process”. Cronbach’s alpha for this scale was 0.748. Team innovative performance was operationalized using the assessment scores assigned by the Academic Evaluation Committees of the respective colleges during the project conclusion phase. In each college, three members of the Academic Evaluation Committee independently evaluated each team’s final project using the same university-wide standardized evaluation rubric and procedure, assigning scores ranging from 1 to 100. Prior to the evaluation, all evaluators received unified training and instructions to ensure consistent interpretation and application of the scoring criteria across colleges. The final team innovative performance score was calculated as the mean of the three ratings. Inter-rater agreement supported the aggregation of the ratings, with ICC (1) = 0.599 and ICC (2) = 0.818. Control variables Following prior research [23, 28], this study controlled for five variables: team size, team type, member familiarity, knowledge overlap, and team mean year. Team size was measured by the number of team members. Team type was dummy-coded as Science and Engineering = 0 and Humanities and Social Sciences = 1. Member familiarity and knowledge overlap were each measured using a five-point, single-item Likert-type scale, an approach considered appropriate for concrete and unidimensional control variables [54]. The corresponding items were “Our team members are all familiar with each other” and “The professional knowledge of our team members overlaps considerably”. Team mean year was calculated by averaging the academic year of all team members. Measurement model evaluation and common method variance assessment A confirmatory factor analysis (CFA) was conducted to evaluate the measurement structure of the three focal constructs: team forgiveness climate, team shared vision, and team learning from mistakes. As shown in Table 1, the hypothesized three-factor measurement model showed an acceptable fit to the data: χ2/df = 1.264, RMSEA = 0.043, SRMR = 0.064, CFI = 0.950, TLI = 0.938, and IFI = 0.952. By contrast, an alternative one-factor model, in which all items were loaded onto a single factor, showed a substantially poorer fit: χ2/df = 2.439, RMSEA = 0.100, SRMR = 0.097, CFI = 0.715, TLI = 0.663, and IFI = 0.725. This model comparison provides evidence for the distinctiveness of the three focal constructs at the measurement-model level. Table 1Confirmatory factor analysis model fit indicesModelχ2/dfRMSEASRMRCFITLIIFIThree-factor1.2640.0430.0640.9500.9380.952One-factor2.4390.1000.0970.7150.6630.725 Confirmatory factor analysis model fit indices We used a multi-source design to reduce common method concerns, as team innovative performance was assessed using third-party evaluation scores rather than self-reported questionnaire data. Additionally, Harman’s single-factor test was performed using an unrotated principal components analysis of all perceptual measurement items. The first factor accounted for 28.27% of the total variance, which is below the commonly used 50% criterion [55]. These results suggest that the data are not dominated by a single common factor. However, because team forgiveness climate, team shared vision, and team learning from mistakes were measured using same-source perceptual reports collected at the same project stage, common method variance cannot be fully ruled out. Data aggregation Because team forgiveness climate, team shared vision, and team learning from mistakes were conceptualized as team-level constructs, we conducted a series of aggregation tests. First, we calculated the within-group interrater agreement coefficient, Rwg. The mean Rwg values for team forgiveness climate, team shared vision, and team learning from mistakes were 0.944, 0.915, and 0.938, respectively. All values exceeded the commonly used threshold of 0.70, indicating strong within-team agreement [56]. Furthermore, we examined the intraclass correlation coefficients. The ICC (1) values for team forgiveness climate, team shared vision, and team learning from mistakes were 0.286, 0.184, and 0.308, respectively, all exceeding the 0.12 benchmark and indicating meaningful between-team variance [57]. The ICC (2) values, which assess the reliability of team means, were 0.522, 0.381, and 0.548, respectively. These values did not reach the standards often cited in the literature, and the ICC (2) value for team shared vision fell below the relatively lenient reference level of 0.47. This pattern should be interpreted in light of the small team size in the present study [58, 59]. Because ICC (2) is highly sensitive to group size, teams with only 2 to 4 members tend to yield attenuated estimates of group-mean reliability. In addition, the corresponding Rwg values were high, indicating strong within-team agreement, and the focal constructs were theoretically conceptualized at the team level. Therefore, we retained the aggregated team-level scores for subsequent analyses, while interpreting results involving team shared vision with appropriate caution. Analysis and results IBM SPSS Statistics 22.0 was used for data analysis. The proportion of missing data was less than 5%; therefore, the expectation–maximization algorithm was used to impute missing values. Following Aiken et al. [60], all continuous variables were mean-centered before the interaction term was created, and standardized regression coefficients are reported in the regression analyses. Descriptive statistics Descriptive statistics and correlations are reported in Table 2. As shown in Table 2, team forgiveness climate was positively correlated with team shared vision (r = 0.403) and team learning from mistakes (r = 0.337). Team shared vision was positively correlated with team learning from mistakes (r = 0.376). Team learning from mistakes was positively correlated with team innovative performance (r = 0.324, p < 0.001). In addition, team forgiveness climate was positively correlated with team innovative performance (r = 0.226), and team shared vision was positively correlated with team innovative performance (r = 0.274). Table 2Means, standard deviations, and correlation analysis of variablesVariablesMeanSD123456781. Knowledge overlap3.860.662. Team mean year3.150.370.1163. Team type0.190.400.089− 0.0844. Team size2.970.220.014− 0.052− 0.0035. Member familiarity4.030.780.177*0.0540.1120.0746. Team forgiveness climate4.370.39− 0.0640.0990.0710.170*0.0947. Team shared vision3.820.32− 0.0920.158− 0.0800.0120.0730.403***8. Team learning from mistakes4.040.440.0210.162− 0.0300.1240.0150.337***0.376***9. Team innovative performance81.2310.120.1150.040− 0.0440.268**− 0.0460.226**0.274***0.324**** p < 0.05, ** p < 0.01, *** p < 0.001 Means, standard deviations, and correlation analysis of variables * p < 0.05, ** p < 0.01, *** p < 0.001 Hypothesis testing We followed the testing procedure proposed by Preacher et al. [61] and used the SPSS MODMED macro developed by the same authors for hypothesis testing. Indirect effects and the index of moderated mediation were evaluated using 5000 bootstrap resamples and 95% percentile bootstrap confidence intervals. Forgiveness climate and team innovative performance To test Hypothesis 1, we estimated a regression model with team innovative performance as the dependent variable and team forgiveness climate as the focal predictor. As shown in Model 6, the total effect of team forgiveness climate on team innovative performance was significant and positive (β = 0.207, p < 0.05), supporting Hypothesis 1. The mediating effect of team learning from mistakes The MODMED procedure involved multiple testing steps. In the mediator model (see Table 3), team forgiveness climate was positively related to team learning from mistakes. As shown in Model 2, team forgiveness climate had a significant positive effect on team learning from mistakes (β = 0.319, p < 0.001), supporting Hypothesis 2. In the outcome model, team learning from mistakes was positively related to team innovative performance. As shown in Model 7, team learning from mistakes had a significant positive effect on team innovative performance (β = 0.255, p < 0.01), supporting Hypothesis 3. In addition, the indirect effect of team forgiveness climate on team innovative performance through team learning from mistakes was significant when team shared vision was at its mean level (effect = 1.573, 95% CI [0.341, 3.344]), supporting Hypothesis 4 (see Table 4). Table 3Hierarchical regression analyses resultsVariablesTeam learning from mistakesTeam innovative performanceModel 1Model 2Model 3Model 4Model 5Model 6Model 7Team type-0.015-0.041-0.0130.005-0.041-0.058-0.048Team size0.1330.0780.0930.0670.275***0.240**0.220**Knowledge overlap0.0020.0350.0570.0300.1260.1470.138Team mean year0.168*0.1290.0970.0800.0400.015-0.018Member familiarity-0.003-0.029-0.0460.020-0.086-0.103-0.096Team forgiveness climate0.319***0.206*0.235**0.207*0.125Team shared vision0.284**0.335***Forgiveness climate × Team shared vision0.259**Team learning from mistakes0.255**R20.0440.1390.2040.2600.0960.1360.192ΔR20.0950.0650.0560.0400.056ΔF15.29411.11610.3636.4049.468Coefficients are standardized. * p < 0.05, ** p < 0.01, *** p < 0.001 Hierarchical regression analyses results Coefficients are standardized. * p < 0.05, ** p < 0.01, *** p < 0.001 Table 4Conditional indirect effects at values of team shared visionTeam shared visionConditional indirect effect95% CI (LL)95% CI (UL)Mean − 1 SD0.265-1.2171.722Mean1.5730.3413.344Mean + 1 SD2.8800.8875.544Percentile bootstrap 95% confidence intervals were based on 5000 resamples Conditional indirect effects at values of team shared vision Percentile bootstrap 95% confidence intervals were based on 5000 resamples It is also worth noting that after team learning from mistakes was entered into the regression equation, the coefficient of team forgiveness climate on team innovative performance became non-significant in Model 7 (β = 0.125, n.s.). This pattern is consistent with a full mediation effect. In other words, the significant total effect of team forgiveness climate on team innovative performance shown in Model 6 was no longer significant after team learning from mistakes was included, suggesting that the association between team forgiveness climate and team innovative performance was transmitted through team learning from mistakes. The moderating effect of team shared vision The mediator model (see Table 3) also examined the effect of team shared vision and the interaction between team forgiveness climate and team shared vision on team learning from mistakes. The results showed that team shared vision was positively related to team learning from mistakes (β = 0.284, p < 0.001). In addition, the interaction between team forgiveness climate and team shared vision was significant (β = 0.259, p < 0.001), supporting Hypothesis 5. As shown in Fig. 2, the simple slope tests indicated that the relationship between team forgiveness climate and team learning from mistakes was not significant when team shared vision was low (slope = 0.046, n.s.), but was significant when team shared vision was at its mean level (slope = 0.271, p < 0.01) and when team shared vision was high (slope = 0.497, p < 0.001). The moderating effect of team shared vision The moderated mediation effect of team shared vision Following Preacher et al. [61], the moderated mediation effect was evaluated using the index of moderated mediation. The results showed that the index of moderated mediation was significant (index = 3.517, 95% CI [0.289, 7.931]), indicating that the indirect effect of team forgiveness climate on team innovative performance through team learning from mistakes varied across levels of team shared vision. Conditional indirect effects were probed at the mean and at one standard deviation above and below the mean of team shared vision (see Table 4). When team shared vision was low, the indirect effect of team forgiveness climate on team innovative performance through team learning from mistakes was not significant (effect = 0.265, 95% CI [-1.217, 1.722]). When team shared vision was at its mean level, the indirect effect was significant (effect = 1.573, 95% CI [0.341, 3.344]). When team shared vision was high, the indirect effect was stronger and remained significant (effect = 2.880, 95% CI [0.887, 5.544]). These results support Hypothesis 6 and are consistent with a first-stage moderated mediation pattern. Discussion This study tested a moderated mediation model to examine the relationship between team forgiveness climate and team innovative performance in college student innovation teams. The findings are consistent with a learning-based pathway in which team learning from mistakes mediates this relationship and team shared vision serves as a strategic boundary condition. The results support Hypothesis 1, showing that team forgiveness climate is positively related to team innovative performance. Consistent with existing empirical evidence suggesting that forgiveness climate may reduce the negative consequences of mistakes in the workplace [23, 30], this study suggests that, in student innovation teams where members’ research experience and professional identities are still developing [11], such a climate may serve as an important emotional buffer [12]. Rather than responding to setbacks with blame, teams with a high forgiveness climate may be better able to reduce negative interpersonal tensions, preserve relational bonds, and sustain the psychological resources needed for innovation. The results also support Hypotheses 2–4, providing evidence for a learning-based pathway linking team forgiveness climate to team innovative performance. Specifically, team learning from mistakes mediated this relationship, which is consistent with the IPO framework [25, 26]. A forgiveness climate may not be directly associated with innovative performance once learning from mistakes is considered; instead, it may provide a relational context characterized by the release of anger and blame after mistakes, along with non-punishment and mutual acceptance [17, 42]. This context may help student teams interpret experimental deviations as learning opportunities and support more effective iteration of their technical solutions. The support for Hypothesis 5 and Hypothesis 6 further contributes to the discussion of the potential downside of forgiveness, including complacency and the repetition of low-value mistakes [4, 32]. The findings indicate that the positive indirect relationship between team forgiveness climate and team innovative performance through team learning from mistakes is stronger when team shared vision is high. This suggests that while forgiveness climate may provide a relational basis for learning, shared vision may offer directional guidance that channels learning toward goal-directed action and performance improvement. Theoretical implications This study offers three main theoretical implications. First, this study extends the application and conceptual boundaries of team forgiveness climate. Previous scholarship has examined forgiveness at the individual, interpersonal, and broader organizational levels [14, 17, 22], as well as its formation and effects on interpersonal relationships [16, 20]. However, the team as a distinct collective unit has received relatively limited attention in forgiveness research [23]. This is an important issue because innovation is often carried out in teams, and collective responses to mistakes may shape group-level outcomes. Prior research has shown that psychological safety can support team innovation by reducing members’ fear of interpersonal risk-taking [3]. Our findings complement this evidence by suggesting that team forgiveness climate is also positively associated with team innovative performance. By introducing this construct into the context of student innovation teams, this study suggests that a forgiveness climate may function not only as an interpersonal virtue but also as a collective team resource. This team-level climate may help teams respond to setbacks with problem solving rather than defensive reactions, thereby offering preliminary empirical support for the study of team forgiveness within organizational behavior. Second, this study identifies a learning-based pathway linking team forgiveness climate to team innovative performance. Drawing on the IPO framework, we provide empirical evidence that team learning from mistakes mediates the relationship between team forgiveness climate and team innovative performance. This finding is consistent with contemporary views that treat mistakes as potential learning resources rather than mere losses [8, 9, 20]. Although prior studies on learning from failure have often examined objective failure events, such as project setbacks and R&D failures, and subsequent behavioral responses [44], our study captures learning from mistakes as a collective cognitive appraisal process that reflects team members’ shared interpretation of mistakes as learning opportunities. By empirically testing this mediating role, we provide a more focused account of how perceived team learning from mistakes may help explain the association between forgiveness climate and team innovation, thereby contributing to organizational learning theory by identifying a learning-based pathway in the context of student innovation teams. Third, this study examines team shared vision as a boundary condition of team forgiveness climate. Scholars have cautioned that unguided forgiveness may lead to complacency and accountability neglect [16, 17, 20]. Our findings suggest that the positive indirect relationship between team forgiveness climate and team innovative performance through team learning from mistakes is stronger when team shared vision is high. This moderating role is consistent with research emphasizing that innovation requires a balance between psychological safety and goal-directed accountability [62]. Shared vision may help maintain accountability within a forgiving atmosphere by directing members’ attention toward common goals and mistakes with learning value. This boundary condition also suggests that the positive affect and cognitive flexibility associated with a forgiving climate may be more effectively channeled into innovation-related outcomes when teams have clear collective direction [63]. Taken together, this finding helps specify the functional scope of forgiveness climate and highlights the importance of aligning emotional support with strategic objectives in student innovation teams. Practical implications This study offers three preliminary implications for practice. First, the findings may inform efforts to support student innovators in universities, although these implications are based on correlational evidence from a single context and require further verification through intervention studies. College student innovation teams often exhibit high enthusiasm alongside limited experience, and setbacks are common in innovation processes [20]. The observed relationship between team forgiveness climate and team innovative performance suggests that cultivating such a climate could be beneficial. Educators might consider framing intelligent mistakes as learning opportunities rather than failures, thereby shifting attention away from purely outcome-based evaluation [2]. Second, our results suggest that a forgiveness climate without directional guidance may be associated with complacency risks. The moderating role of shared vision implies that if university faculty advisors or project mentors wish to foster forgiveness, attending to strategic alignment early in a project might help students differentiate between valuable experimental deviations and preventable errors caused by insufficient effort. By connecting project goals to broader academic or societal impacts, teams may be more likely to maintain a benevolent yet disciplined and performance-oriented atmosphere [32]. Third, our results suggest that learning from mistakes does not automatically follow from a positive climate; rather, it may require structured cognitive processing [8]. Universities could consider integrating formal reflection mechanisms, such as post hoc analyses after mid-term evaluations, mistake portfolios alongside final deliverables, or peer-review workshops, to help student teams turn isolated mistakes into shared learning resources [64]. These practices may provide useful directions for future educational practice and intervention design. Limitations and future research This study has several limitations. First, the unique characteristics of the student-led innovation sample warrant further cross-contextual validation. Although student research teams may share some interactive and cooperative features with professional R&D teams, they may exhibit lower innovation intensity and weaker structural continuity than corporate innovation teams. Future research could examine these relationships in professional innovation teams to enhance the generalizability of the findings. Second, the construct of forgiveness climate deserves further multidimensional exploration. This study adopted a scale that captures core psychological elements of team-level forgiveness. However, as theoretical frameworks evolve, some scholars have suggested that forgiveness climate may be a multifaceted construct involving distinct cognitive, affective, and behavioral dimensions [17, 22]. Future research could employ a multidimensional measurement approach to examine the impact of forgiveness climate in teamwork. Third, the research design limits causal inference. The perceptions of team forgiveness climate, team shared vision, and team learning from mistakes were measured at a single developmental milestone, which constrains stronger causal inference regarding the directional relationships among these variables and leaves room for reciprocal influence or broader contextual explanations. In addition, although team innovative performance was assessed using external evaluation scores, the key internal variables were based on same-source perceptual reports. Therefore, common method variance among these perceptual variables cannot be fully ruled out. Future research could collect predictor, mediator, and moderator data from different sources or at different time points to provide stronger temporal and methodological evidence. Fourth, the aggregation and reliability issues should be considered when interpreting the findings. Although the aggregation tests generally supported aggregation to the team level, the ICC (2) estimate for team shared vision fell below the reference level of 0.47 often discussed for small groups, indicating comparatively limited reliability of the aggregated team mean for this construct. In addition, the reliability estimate for team shared vision was comparatively modest, which may have attenuated parameter estimates, particularly the moderating effect involving this construct. Future research could employ larger teams, repeated assessments, or multi-wave designs to improve the reliability of aggregated team-level measures. Fifth, some measurement limitations should be acknowledged. Member familiarity and knowledge overlap were each measured using a single-item indicator for reasons of survey economy. Although this approach may be acceptable for concrete and unidimensional control variables, it cannot fully capture the complexity of interpersonal familiarity or knowledge structure in student research teams. Future research could adopt more comprehensive multi-item measures to improve psychometric precision. Sixth, although team innovative performance scores were obtained from standardized external evaluations, such ratings may still contain evaluator judgment and disciplinary variation. Future studies could combine external ratings with additional objective indicators to further strengthen measurement robustness. Finally, the interpretation of team shared vision deserves further examination. Teams with high shared vision may generally function more systematically and may possess stronger learning and innovation capabilities. Thus, the moderating role of shared vision in the forgiveness climate–performance relationship may be partly attributable to these broader team advantages rather than to shared vision alone. Future studies could adopt experimental or longitudinal designs to disentangle the specific role of shared vision from broader team competence. Conclusion Using an IPO framework, this study proposed and tested a moderated mediation model examining the relationship between team forgiveness climate and team innovative performance in student innovation teams. The empirical analysis of 145 student innovation teams showed that team forgiveness climate was positively associated with team innovative performance, and this association was consistent with a full mediation pattern through team learning from mistakes. Moreover, this indirect relationship was conditional on team shared vision: when team shared vision was high, the positive relationship between team forgiveness climate and team learning from mistakes was stronger. These findings provide preliminary empirical support for the links among collective emotional climates, team learning from mistakes, and innovation outcomes in student innovation teams. Acknowledgements Not applicable. Authors’ contributions HH and HF conceptualized the study, conducted the investigation, developed the methodology, curated the data, drafted the original manuscript, and reviewed and edited it. Additionally, HH acquired resources, and HF provided software. CL contributed to the conceptualization, provided resources, and reviewed and edited the manuscript. JV contributed to the conceptualization and reviewed and edited the manuscript. All authors read and approved the final manuscript. Funding This research was funded by Zhejiang Federation of Social Sciences Research Project (No. 2026N037). Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The process and procedures used in this study adhere to the ethical principles outlined in the Declaration of Helsinki. Ethical approval was obtained from the Ethics Committee of the Zhejiang University Global Entrepreneurship Research Center, Hangzhou, China. The ethical approval protocol number is ZJUGERC-20241010. This approval encompasses all aspects of the study, including data collection, analysis, participant recruitment, informed consent procedures, and confidentiality measures, as outlined in the approved protocol. The research was conducted strictly in line with the ethical framework approved by the Ethics Committee of the Zhejiang University Global Entrepreneurship Research Center. All participants provided voluntary informed consent after fully understanding the research purpose, anonymity provisions, and procedures for ensuring data confidentiality and security. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Honghao Hu, Chao Li and Jonas Valkonen contributed equally to this work. References Associated Data Data Availability Statement The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.