The mediating effect of team identity on creative
performance in pop/rock bands
Lotte Schuilenborg
10850864
Supervisor: Machteld van den Heuvel
Research Master’s Psychology
2
Abstract
This research uses a combination of the group creativity model and social identity theory in order to negate some of the inconsistencies in the current research on antecedents of creative performance. We argue that team identity is an important and neglected motivational factor that influences creativity. We hypothesize that team identity mediates both the effect of perceived leadership on creative performance and the effect of psychological safety on creative performance. Consistent with one of these predictions, data based on 134 members of pop/rock bands showed that the positive effect of psychological safety on creative performance is positively mediated by team identity. We elaborate on why these effects were not found for perceived leadership and on the importance of the situational and motivational variables within a team when creative performance is the goal.
3
Introduction
Creativity is a quality ever more valued in contemporary society and today’s business environment. Creativity, or creative performance, is broadly defined as the creation of new ideas, insights, or products that are both novel and appropriate (Baas, Roskes, Sligte, Nijstad & De Dreu, 2013; Shin & Eom, 2014). Due to the importance it has gathered in business settings, its antecedents and facilitating factors have been researched extensively (e.g. Tierney & Farmer, 2011). Paulus and Dzindolet (2008) created a group creativity model (GCM) in which they tried to incorporate both situational and motivational factors that influence creative outcomes. However, within these factors, there are still considerable inconsistencies in the literature. For instance, the group structure factor perceived leadership (i.e. having a central vision about leadership within a group) and its relationship with creative performance is highly debated. On the one hand, a leader is said to be able to facilitate effective task distribution (Tyran, Tyran, & Shepherd, 2003), influence motivational team processes (Zaccaro, Rittman, & Marks, 2002) and create a central focus or vision for the whole team (Sivasubramaniam, Murry, Avolio, & Jung, 2002). On the other hand, it is claimed that the presence of a team leader is not always associated with higher performance (Tyran et al, 2003) and leadership has long been viewed as an obstacle to creativity due to the individual nature of creativity (Vessey, Barret, Mumford, Johnson and Litwiller, 2014). Next to this, the group climate factor psychological safety, defined as “the shared belief that the team is safe for interpersonal risk taking” (Edmondson, Roberto & Watkins, 2003, p. 306), also received inconsistent support regarding its effect on creative performance. On the one hand, psychological safety is said to decrease the fear of negative evaluation which could inhibit individuals’ ability to be creative (Baas, De Dreu & Nijstad, 2008). On the other hand, growing up in hard and traumatic times is shown to be related to creative genius (Paulus & Dzindolet, 2008).
The aim of this study is to combine the GCM (Paulus & Dzindolet, 2008) and the Social Identity Theory (SIT) (Tajfel & Turner, 1979) in order to negate these inconsistencies by introducing the mediating variable Team Identity. It is proposed as a mediating variable between the effect of
4
Creative Performance (see Figure 1). The research question is thus: “Is the motivational process of team identification a mediator between situational factors in a band and the band’s creative performance?” In line with the argumentation of Vessey et al (2014) a sample of highly creative
individuals (i.e. musicians playing in a band) is used in order to investigate this question. In this way the current study adds to the previous literature by explaining the above mentioned inconsistencies through using a sample that is highly indicative of these relationships, as the success of these musicians is contingent on creativity and novelty in their performance.
Figure 1 The proposed model
Theoretical framework
The GCM is one of the first models examining creativity from a team perspective as it acknowledges the importance of the current focus on teamwork and the use of multidisciplinary teams (Kozlowski & Ilgen, 2006). Paulus and Dzindolet (2008) posit that group, task and situational variables all affect group creativity through their effect on motivational processes. In addition to the motivational processes they describe, this study adds team identity as a motivational factor. Team identity comes from SIT which identifies the need for identification, and in this case the need for team identificiation, as a process concerned with “the extent to which individuals define themselves in terms of another individual, relationship or group” (Cooper & Thatcher, 2010, p. 517). This need is considered one of our basic motivational needs (Tajfel & Turner, 1979). When this shared sense of identification is present in a team, a motivational climate arises which fosters an atmosphere in which
5 people can work together (Van der Vegt & Bunderson, 2005). Thus, identifying with a group or team is seen as a motivational process (Tajfel & Turner, 1979) and consequently it is a construct that fits well within the GCM framework.
Perceived leadership, psychological safety and creative performance
It is grounded in the literature of team creativity that perceived leadership and psychological safety are pivotal to team creative performance. First, perceived leadership can create a central vision and focus for the team, which enhances creativity (Sivasubramaniam et al, 2002). Although the GCM considered leadership style instead of perceived leadership as a group structure variable, this study will focus on the central vision about leadership because studying specific leadership styles has not led to substantial results on predicting creative performance (Shin & Eom, 2014; Wang & Rode, 2010). Second, psychological safety, in the GCM model, is proposed to allow group members to express themselves freely and to promote risk-taking, which together facilitate creative outcomes (Paulus & Dzindolet, 2008).
The GCM thus advocates for the positive effect of these group structure and climate factors on creative performance. In line with this model the following relationships are postulated:
Hypothesis 1: Perceived leadership has a positive effect on subjective (H1a) and other-rated (H1b) creative performance.
Hypothesis 2: Psychological safety has a positive effect on subjective (H2a) and other-rated (H2b) creative performance.
Team identity as mediator
As mentioned above, not all current literature conforms to the ideas of the GCM and therefore the SIT is introduced. In many different contexts team identity, or the feeling of oneness with or
6 belongingness to the team (Mael & Ashforth, 1992), has shown to be positively related to performance. For instance, team members with a high sense of identification were more likely to be greater contributors to the outcomes of the organization, thereby enhancing these outcomes (Van Knippenberg, Van Knippenberg, de Cremer, & Hogg, 2004). In the current study, the outcome variable is operationalized as creative performance, as this is an important factor to a band’s success (Sawyer, 2006).
Next to this positive effect on performance, team identity can possibly reverse the effects that a non-central vision of leadership and non-perceived psychological safety can have on creative outcomes. For instance, it is argued that, in a group, individuals may feel apprehensive of what others think of them and thereby limit their sharing of ideas (Camacho & Paulus, 1995). However, when an in-group feeling is created through identifying with the group, sharing of ideas and acceptance of each other is facilitated through a higher sense of collective self-esteem (Crocker & Luhtanen, 1990). Furthermore, perceived leadership could lead to more task conflict, which hurts team performance (De Dreu & Weingart, 2003), but when there is a sense of team identity both task conflict and interpersonal conflict are generally lower (Hinds & Mortensen, 2005). In line with these effects of team identity on the relationship between both perceived leadership and psychological safety on creative performance the following relationships are postulated:
Hypothesis 3: Team identity positively mediates the relationship between perceived leadership
and subjective (H3a) creative performance and the relationship between perceived leadership and other-rated (H3b) creative performance.
Hypothesis 4: Team identity positively mediates the relationship between psychological safety
and subjective (H4a) creative performance and the relationship between psychological safety and other-rated (H4b) creative performance.
7
The present study
In the present study we followed a sample of band members that signed up for a band competition. The individual musicians filled out questionnaires before and after a band competition. Also, jury reports were gathered on the creative performance of the bands during their gig. These data are used to test the proposed hypotheses.
Method
Participants
A sample of 144 band members from 38 bands participated in the study (14.6% female, Mage =
37.92, SDage = 13.4). The jury consisted of a varying group of professional musicians who are
employed in the music industry. In total there were 12 jury members, on average there were 3 judges per evening. The composition of the jury changed depending on the date of the competition. The band members were part of a competition in which they were evaluated by a jury in order to see whether they would advance to the finale. After data analysis 10 participants were excluded from further analyses: because they did not provide any information on which band they were in, which hindered the multilevel analysis (N = 9) or because it was a band that only contained a jury rapport and no individual data (N = 1).
Procedure
When the participants signed up for the competition, they received an e-mail asking whether they would like to participate in research about factors that can predict a band’s success. As an incentive, they were notified that participating offered the possibility to win one free studio day.
If the participant agreed to participate, he or she received a questionnaire assessing, along with some demographic and control questions, different leadership and effectiveness topics, such as the
8 ones used in this study: team identity, psychological safety and perceived leadership. One day after the competition, the participant received the second questionnaire which assessed the individual subjective perception of the band’s success of the performance.
During the competition, the jury, which was asked to judge the participants in light of the competition, received a jury rapport, 1 A4 which held the questionnaire relating to the current research, asking them to evaluate the success of the performance of the band. The jury evaluation rapport (i.e. the measure of other-rated creative performance) was used as well as the subjective evaluation of the participants (i.e. the measure of subjective creative performance) because what has been a common problem in leadership research is that outcomes have a strong evaluative and subjective component, especially when leaders are being assessed (e.g. van Knippenberg & Sitkin, 2013). Therefore, by including both, this study assesses both other-rated and subjective success of performance.
Materials
Perceived leadership. A leader can emerge both from being appointed to that position or from
responding to the characteristics of the team (Judge, Piccolo, & Kosalka, 2009). In this study, perceived leadership is operationalized as indicating that there is a person on who the team as a whole reaches a consensus on that they perceive that person to be their leader (Tyran et al, 2003). A self-constructed scale of 4 items was used to assess perceived leadership. A sample item was: “Everyone agrees on who the band-leader(s) is/are”1. The items were answered on a 6-point Likert (1 = ‘totally disagree’, 6 = ‘totally agree’). The scale was reliable and internally consistent (α = .853).
Team identity. Team identity was assessed through a combination of two scales: the one by
Doosje, Ellemers, and Spears (1995) and the one used by Mael and Ashforth (1992). The combined scale assessed team identity through 5 different items. The scales were combined in order to ensure face validity. A sample item is: “When someone is positive about the band, it feels like a personal
1
9 compliment”. The items were answered on a 6-point Likert (1 = ‘totally disagree’, 6 = ‘totally agree’). The scale was reliable and internally consistent (α = .857).
Psychological safety. Psychological safety was assessed through a shortened version of a scale
used by Edmondson (1999). The scale assessed psychological safety through 5 different items. A sample item is: “If you make a mistake in the band, this will not be held against you”. The items were answered on a 6-point Likert (1 = ‘totally disagree’, 6 = ‘totally agree’). The scale was reliable and internally consistent (α = .758).
Other-rated creative performance. Other-rated creative performance was assessed through an
adapted and build-on version of a scale used by Oldham and Cummings (1996). This will not be assessed by the participant, but by the different jury members. It assesses the creative performance of the gig played during the competition through 9 items. A sample item is: “The songs played by the band were original and strong”. The items can be answered on a 7-point Likert (1 = ‘totally disagree’, 7 = ‘totally agree’). The scale was reliable and internally consistent (α = .949), with a reasonable consistency amongst the jury members (the ICC(2,k) was on average .723, with not one of the items falling below .65).
Subjective creative performance. In order to keep the other-rated creative performance and
subjective creative performance as comparable as possible, the same scale was used for subjective success. However, here the questions took the point of view of the participant, for instance: “We were able to impact the audience”. The scale was reliable and internally consistent (α = .843).
Strategy of analysis
The data provided a complex multilevel structure with variables measured both at level 1 (i.e. the individual in the band) and level 2 (i.e. the band). Due to the time constraint of the current internship, some considerations about the statistical procedures were needed and they will be discussed below.
10 In order to be able to analyse the data, it was first necessary to compute the Cronbach’s alpha of the scales that were used (these values can be found in the methods section). A Cronbach’s alpha coefficient of .7 was used as a cut off value. For the judges that provided the other-rated assessment of the creative performance of bands, the Intraclass Correlation Coefficient of two-way random average measures (ICC(A,k)/ICC(2,k) was used (McGraw & Wong, 1996). A cut-off value of .65 was used to ensure that the jury had a fair agreement on how the band performed. These scores were analysed on the (higher) band level whereas the scores from the band members were analysed on the (lower) individual level. All scales met the cut-off values and were therefore combined into a mean score. The longer questionnaires of the creative performance were subjected to EFA analysis and they both implied one factor and met the Cronbach’s alpha cut-off value, therefore they were also combined into a mean score. Multilevel SEM would have allowed for the items to be separate indicators, where mean scores would not be needed; however this would result in estimating too many parameters.
The sample was rather small for the amount of parameters that are often estimated in multilevel analyses. This led some multilevel models to result in non-convergence because the apparent structure of the data caused the final Hessian matrix to not be positive definite and therefore the validity of the estimates could not be ascertained. Consequently, not all random parameters could be estimated. Fortunately, the fixed effects could still be estimated and analysed, thereby allowing for hypothesis testing whilst accounting for the multilevel structure in the data. Another problem with the small sample was the covariance structure that needed to be specified. Ideally, the model would estimate an unstructured covariance matrix, in order not to make any assumptions about any of the (co)variances. However, the multilevel models specified with an unstructured covariance matrix did not converge at all as it had to estimate too many parameters. The other option was using the variances only structure, which estimated only variances and specified the covariances to be zero. Estimating less parameters was favourable due to the small data set, which will often not converge when many parameters are estimated (Tabachnick & Fidell, 2001). Furthermore, the hypotheses were not mainly concerned with the covariance structure, only with the variance structure, and the model converged
11 using the variances only structure. Thus, this seemed to be the best option given the data at hand and was therefore used in the analyses.
The data was analysed not only for direct effects (Hypotheses 1 and 3), but also for mediation effects (Hypotheses 2 and 4). There is still a lot of research on how to perform mediation in multilevel data (Preacher, Zhang, & Zyphur, 2011) and there are various methods, all with their own advantages and disadvantages. One method that showed little bias in their simulation study was a stacked method by Bauer, Preacher, and Gil (2006). They created a new variable that, in combination with indicator variables, could test the indirect effect of X on Y through mediator M in one multilevel model (instead of different models for the effect on Y and M). However, this method also estimated all the random effects, resulting in the same problem as described above: non-convergence. Although a conceptual understanding of Bauer et al.’s (2010) method was achieved, the mathematics behind the procedure were too complex to be able to change the model in order to only estimate fixed effects. A method that does estimate only fixed effects is described by Zhang, Zyphur, and Preacher (2009). Their method was based on the way of calculating indirect effects in normal linear regression and they added the multilevel structure. Thus, their indirect effect was also a result of the product-of-coefficients method (the effect of X on M times the effect of M on Y). They also distinguished between the within- and between-subjects effects of mediation; therefore this method was suitable for the current multilevel data. As this method did lead to convergence of the model, this method, using SPSS MIXED, was chosen. Significance testing was done using a Sobel test.
The hypotheses about the other-rated creative performance added another complication to the data, as these hypothesized about a dependent variable measured on level 2, whereas all the predictors were measured on level 1. The main idea was to convert the level 1 scores in such a way that they could be analysed at level 2. Croon and van Veldhoven (2007) presented an elegant matter to do so, however, their method only related to analyses of direct effects and not to the analysis of mediation effects. Furthermore, their method was computationally complex and time-consuming. After more research it was found that for a model like the current one (i.e. a 1 – 1 – 2 model): “MLM [traditional multilevel modelling] cannot be used; Level-2 dependent variables not permitted” (Preacher, Zyphur
12 & Zhang, 2010, p. 211 Table 1). In order to still be able to analyse the other-rated creative performance, individual scores within the band were transformed to an aggregate score (the mean) per band. This meant that results could no longer be interpreted on the individual level, but only on the band level.
The above mentioned choices will be reflected in the results section. Whereas some choices were suboptimal, they provided the best solution given the time and knowledge at hand.
Results
Subjective creative performance
In order to investigate hypothesis 1a and 2a, SPSS MIXED was used as this allowed for multilevel modeling of the data. For both hypothesis 1a and 2a, the LRT test statistic indicated significant variability between bands (Perceived leadership: χ2(1) = 14.73, p < .001; Psychological Safety: χ2(1) = 16.45, p < .001). Next to this, the intraclass correlation indicated that 35.37% of the unexplained variance in subjective creative performance was due to the band level. These two statistics thus stress the necessity of multilevel analyses.
A multilevel mixed regression analysis showed that perceived leadership did not significantly predict subjective creative performance (γ10 = .14, SE = .11, t(62.67) = 1.25, p = .22). Hypothesis 1a
could thus not be supported. Testing hypothesis 2a (whether psychological safety positively predicts subjective creative performance) in a fully multilevel manner resulted in non-convergence. However, when the slope for psychological safety with subjective creative performance was set to be fixed, parameters could be estimated. It showed that psychological safety had a positive relationship with subjective creative performance (γ10 = .57, SE = .20, t(80.95) = 2.82, p < .01) thereby supporting
hypothesis 2a.
Multilevel mediation analysis showed that team identity did neither mediate the relationship between perceived leadership and subjective creative performance (H3a) on a within-band level
13 (estimate of indirect effect = .08, Sobel test statistic = 1.51, p = .13) nor on a between-band level (estimate of indirect effect = -.0, Sobel test statistic = -.19, p = .85). Therefore, hypothesis 3a was not supported. Team identity did significantly mediate the relationship between psychological safety and subjective creative performance on a within-band level (estimate of indirect effect = .18, Sobel test statistic = 2.09, p < .05) but not on a between-band level (estimate of indirect effect = .25, Sobel test statistic = 1.21, p = .23) thereby lending support to hypothesis 4a.
Other-rated creative performance
To test hypothesis 1b and 2b, simple linear regression in SPSS was used. This showed that both perceived leadership (B = .35, β = .23, t(33) = 1.33, p = .19) and psychological safety (B = .41, β = .14, t(33) = .79, p = .44) did not significantly predict other-rated creative performance. Thus, hypothesis 1b and 2b were not supported.
In order to test hypothesis 3b and 4b, the bivariate correlations between the predictor, dependent and mediator variable were analysed. For both the variables of hypothesis 3b (perceived leadership, team identity, and other-rated creative performance) and 4b (psychological safety leadership, team identity, and other-rated creative performance) the bivariate correlations were not significant, indicating that mediation was not likely to occur. Subsequent mediation analyses using the ‘Process’ macro by Hayes for SPSS confirmed these predictions. Thus, hypothesis 3b and 4b were not supported.
14
Table 1 Means, Standard Deviations, and Correlations
Variable M SD 1 2 3 4 5 1.Perceived leadership 3.361/3.392 1.12/.75 - .20 -.19 - .23 2.Psychological safety 5.13/5.14 .68/.38 .21* - .29 - .14 3.Team identification 4.93/5.08 .70/.48 .11 .33** - - .19 4.Subjective creative performance 6.89/- 1.22/- .16 .27* .41** - - 5.Other-rated creative performance -/6.02 -/1.13 - - - - -
Note. The upper diagonal presents correlations for the aggregated data used for the other-rated
performance, whereas the lower diagonal shows correlations for the individual data used for the subjective creative performance. 1Mean in individual dataset. 2Mean in aggregated dataset. * p < .05. ** p < .01.
Discussion
This research set out to study the relationships between motivational and situational factors and their effect on creative performance. Situational factors such as psychological safety and perceived leadership were expected to positively influence creative performance. These relationships were expected to be positively mediated by team identity, a motivational factor. We found that psychological safety has both a direct and an indirect effect (through team identity) on subjective creative performance, thereby unraveling some of the potential effects that situational and motivational factors can have on creative performance. We did not find direct or indirect effects of perceived leadership on subjective creative performance nor did we find effects of the perceived leadership and psychological safety on other-rated creative performance. Below we will discuss the theoretical
15 contribution of the study, factors that may explain these results as well as suggestions for future research.
The direct effect of psychological safety on subjective creative performance goes back to classic studies on organizational change (e.g. Schein & Bennis, 1965) in which it is argued that a work environment characterized by psychological safety is essential for individuals to change their behaviour and come up with innovative ideas. Furthermore, the effect of psychological safety on subjective creative performance is also in line with more recent research findings (e.g. Gilson & Shalley, 2004). For instance, Baer & Frese (2003) found that psychological safety positively related to process innovativeness. Even though innovativeness was taken as a measure, it can be compared to individual creativity, as this is a necessity for innovative ideas (Kark & Carmeli, 2009). A study that kept closer to our measures was one by Kessel, Kratzer and Schultz (2012) who examined this effect in health care teams. They used the same measure for psychological safety and they also employed a subjective creativity scale, the only difference being that they had participants evaluate members of the team separately instead of the creativity of the team as a whole. They also found a positive effect of psychological safety on creative performance.
The mediating effect of team identity on the relationship between psychological safety and subjective creative performance has, to our knowledge, not yet been investigated. However, the current finding is in line with the argument that identification fosters a sharing of ideas leading to more creative output (Crocker and Luhtanen, 1990). This positive mediation effect also sheds some light on how situational factors can affect subjective creative performance and on that both the individual (who identifies with the team) and the team (with which the individual identifies) are important in enhancing creative performance.
The theoretical relevance of the current research is that these results shed some light on the inconsistencies present on the effect of situational variables on creative performance by introducing a new motivational mediator. Whereas other mediators of the link between psychological safety and subjective creative performance have been investigated (e.g. vitality and aliveness as mediators; Kark
16 & Carmeli, 2009), this research introduced team identity as a mediator, in line with Social Identity Theory. We have shown that in the case of psychological safety, team identity has a positive mediating effect, which further supports the positive link between psychological safety and subjective creative performance. This research thus succeeded in negating some of the inconsistencies regarding the effect of a situational variable (psychological safety) on subjective creative performance by introducing a motivational variable: team identity has both a positive effect on subjective creative performance and a positive mediating effect. These results also support the general findings of the positive link between (positive) situational variables and creative performance (e.g. Baas et al., 2008). Thus combining the Group Creativity Model and SIT shed some new light on the antecedents of creative performance. Future research should explore more antecedents in light of these theories, maybe even resulting in identity theory being added to the GCM framework as a motivational variable.
Although previous studies did not only find effects of psychological safety but also of perceived leadership on creative performance (e.g. Sivasubramaniam et al, 2002), in this study no such effects were found. The lack of support for these hypotheses about perceived leadership could be due to the nature of the construct ‘perceived leadership’. Whether a leader is perceived by the group, may be too vague a construct to explain creative performance. For instance, leadership in itself might be advantageous in some situations and disadvantageous in others (Judge, Piccolo & Kosalka, 2009). To reduce this problem it would be an option to investigate the quality of leadership instead, as this gives an evaluation about the leadership, just like the creative performance is evaluated. This is in line with current research that shows that quality of leadership has an influence on creative performance. For instance, Amabile, Schatzel, Moneta and Kramer (2004) show that the quality of the leader influences the overall creativity in the work of the subordinates. Also, Bakker and Xanthopolou (2013) explain that positive affect (which is part of quality of leadership) has a positive relationship with creativity.
The inconsistent findings with regard to the mediating effect of team identification (i.e. it does mediate the effect of psychological safety on subjective creative performance and not the effect of perceived leadership even though both are situational variables) is in line with Janssen and Huang’s (2008) notion that individual differentiation rather than team identification could lead to creativity.
17 However, Hirst, Van Dick and Van Knippenberg (2009) argue that leaders help catalyse a motivation to be creative by both rendering the team and its goals salient and by building on the beliefs of the efficacy of the team, which is not in line with our findings. An explanation could be that the leadership structure of the sample should be taken into account. For instance, the teams in Hirst et al.’s study (2009) were research and development teams, which had a clear designated leader. It could be that for bands, the leadership structure is not one of one leader but one of shared leadership. For instance, bands could fit well meso-level model of leadership dynamics (Bligh, Pearce & Kohles, 2006). They posit that self-leadership (which is present in a band, as different musicians are responsible for their respective instruments and job-aspects that apply to their instrument alone) through team trust, potency and commitment leads to a sense of shared leadership. This shared leadership then allows for novel knowledge creation, or creative solutions. Therefore, the inconsistencies could be due to one of the situational variables not fully applying to the current sample. Thus, the current research has mainly highlighted that there is still a level of inconsistency with regard to the mediating effect of team identity and that this should be examined in order to comprehend in which situations team identity takes this mediating role and influences creative performance.
The absence of support for the hypotheses about other-rated creative performance are most likely due to the low power to find effects, as the data had to be aggregated to the band level. This aggregation also led to loss of information on the individual level, which was the level of interest. Even though the decision to aggregate the data was made because it was best for the current project, it is a shame that the data could not be analysed on an individual level. It would be ideal to have an other-rated scale that is also measured at the level of the individual – dyad research would be an interesting option in this case.
Importantly, it should also be noted that other models should be examined in order gain a more universal understanding about the effects of motivational and situational factors on creative performance. It could, for instance, also be that psychological safety instead of having a direct effect on creative performance takes a mediating role when different situational or motivational variables are assessed. To illustrate, it has been found that leader inclusiveness fosters a positive climate, one of
18 psychological safety, and also support for creativity (e.g. Carmeli, Reiter-Palmon & Ziv, 2010). Also, as Detert and Burris (2007) argue, when a leader signals the willingness to listen to the subordinates, the risks of speaking up about your creative ideas diminishes. These effects foster creative outcomes, but argue for a more subtle role of psychological safety. Future research could thus analyse more and diverse motivational and situational variables to further explore the role of psychological safety.
Strengths and limitations
The limitations of this study should be noted in order to facilitate future research. First of all, this study was part of a larger study, which assessed various variables. So at the same time the variables in this study were assessed, others were too. This created a rather long questionnaire with only self-report measures, implying a risk of common method variance. However, there are only moderate correlations between the variables, indicating that common method variance should not be a serious risk in this study. Also, predictors and outcomes were not measured at the same time, reducing the risk (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). Next to this, other-ratings (of the jury) were used as well as the self-report measures, adding to the robustness of the measures.
Secondly, the sample size of the current study was rather small, due to unforeseen circumstances such as the band competition not continuing for another season. This made the sample size too small to use structural equation modelling, with all the items loading on their respective construct. However, the Cronbach’s alphas of all scales were acceptable for creating scale mean scores. Next to this, the longer questionnaires of the creative performance were subjected to EFA analysis and they both loaded on one factor. This in combination with a good Cronbach’s Alpha allowed for creating a mean score for these variables too. The small sample size, however, also resulted in problems with the estimation of all the parameters in the multilevel models. Furthermore, the small sample size made the Sobel test less appropriate as it should ideally only be used with large samples. However, there was not enough time to investigate the bootstrapping method that is advised
19 for smaller samples (Preacher & Hayes, 2008). For future research a power analysis which takes in consideration the multilevel structure is recommended prior to data collection.
Thirdly, it might be that winners of the band competition, or bands who generally played a very good gig, were subdued to the overconfidence effect (Dunning, Griffin, Milojkovic, & Ross, 1990). This overconfidence effect results in a higher subjective confidence of a person in his or her judgment and performance. Unfortunately, in the current dataset we only had data on who the winner of the competition was for a small subset of our data, so we could not analyse whether this effect was present. However, here again it is favourable that we also had an other-rated measure in order to rule out this overconfidence effect.
Despite the lack of support for most hypotheses, the strength of this study lies in its realistic approach. Through the use of actual bands and their actual performance, this is more than just a lab study. For instance, because completing the study was not the common goal of the bands (i.e. their goal was winning the competition) the study will probably not have influenced their performance, thereby assessing a real world phenomenon.
Furthermore, a main benefit of this study is that both subjective and other-rated performance are analysed. By adding other-ratings, many self-report biases can be averted or reduced. Next to this, it provides an interesting angle to see how individual measures can predict such a group measure. This, together with the realistic approach due to the sample makes generalizability more plausible.
Future research should address the black box that affects creative performance, to disentangle exactly which motivational and situational factors are at work. This is needed because creative performance is ever more important in contemporary society; when understanding this process creative performance can be facilitated and thereby enhanced.
20
Practical implications
As this research was conducted with a highly creative and realistic sample, implications can be direct suggestions to teams that are comparable to, or are in fact, bands. However, these implications can also be implemented in teams that want to start focussing on enhancing creativity and take this as their starting point. This research stresses the importance of situational and motivational variables in affecting creative output. Especially psychological safety has been shown to influence creativity. For a band, it would thus be ideal not only to focus on the end-product (i.e. the music and the performance) but also on the relations within the band. Once the band members have the feeling that they are accepted, and are in a psychologically safe environment, they will be enabled to be more creative in their output or performance. It also fosters a sense of identification within the band, which is useful in various situations within a team as shown by previous research, but this research adds that this also enhances creative output. Thus, the focus of a creative team should not merely be on the end-product, but on facilitating an environment that will let creativity blossom.
Conclusion
This research shows that GCM and SIT models can be combined in order to facilitate the understanding of the antecedents of creative performance. Team identity is introduced as a mediating factor between situational variables and creative performance and we find that it mediates the link between psychological safety and subjective creative performance. This suggests a motivational variable to the GCM framework and gives an addition to the current research on unravelling the ‘black box’ that enables creative performance. All in all, this research shows that in order to facilitate creative performance, one has to start at the beginning: with the people and the environment.
21
References
Amabile, T. M., Schatzel, E. A., Moneta, G. B., & Kramer, S. J. (2004). Leader behaviors and the work environment for creativity: Perceived leader support. The Leadership Quarterly, 15, 5-32. doi:10.1016/j.leaqua.2003.12.003
Baas, M., De Dreu, C. K., & Nijstad, B. A. (2008). A meta-analysis of 25 years of mood-creativity research: Hedonic tone, activation, or regulatory focus?. Psychological bulletin, 134(6), 779-806. doi: 10.1037/a0012815
Baas, M., Roskes, M., Sligte, D., Nijstad, B. A., & De Dreu, C. K. (2013). Personality and creativity: the dual pathway to creativity model and a research agenda. Social and Personality
Psychology Compass, 7, 732-748. doi:10.1111/spc3.12062
Baer, M., & Frese, M. (2003). Innovation is not enough: Climates for initiative and psychological safety, process innovations, and firm performance. Journal of Organizational Behavior, 24, 45–68. doi: 10.1002/job.179
Bakker, A. B., & Xanthopoulou, D. (2013). Creativity and charisma among female leaders: the role of resources and work engagement. The International Journal of Human Resource Management,
24(14), 2760-2779. doi:10.1080/09585192.2012.751438
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of
personality and social psychology, 51(6), 1173-1182. doi:10.1037/0022-3514.51.6.1173
Bauer, D. J., Preacher, K. J., & Gil, K. M. (2006). Conceptualizing and testing random indirect effects and moderated mediation in multilevel models: new procedures and recommendations. Psychological methods, 11, 142-163. doi:10.1037/1082-989x.11.2.142
22 Bligh, M. C., Pearce, C. L., & Kohles, J. C. (2006). The importance of self‐ and shared leadership in team based knowledge work. Journal of Managerial Psychology, 21(4), 296–318. doi:10.1108/02683940610663105
Camacho, L. M., & Paulus, P. B. (1995). The role of social anxiousness in group brainstorming. Journal of personality and social psychology, 68(6), 1071-1080. doi:10.1037/0022-3514.68.6.1071
Carmeli, A., Reiter-Palmon, R., & Ziv, E. (2010). Inclusive leadership and employee involvement in creative tasks in the workplace: The mediating role of psychological safety. Creativity
Research Journal, 22, 250-260. doi:10.1080/10400419.2010.504654
Cooper, D., & Thatcher, S. M. (2010). Identification in organizations: The role of self-concept orientations and identification motives. Academy of Management Review, 35, 516-538. doi:10.5465/amr.2010.53502693
Crocker, J., & Luhtanen, R. (1990). Collective self-esteem and ingroup bias. Journal of personality
and social psychology, 58(1), 60-67. doi:10.1037/0022-3514.58.1.60
Croon, M. A., & van Veldhoven, M. J. (2007). Predicting group-level outcome variables from variables measured at the individual level: a latent variable multilevel model. Psychological
methods, 12, 45-57. doi:10.1037/1082-989x.12.1.45
De Dreu, C. K., & Weingart, L. R. (2003). Task versus relationship conflict, team performance, and team member satisfaction: a meta-analysis. Journal of applied Psychology, 88(4), 741-749. doi:10.1037/0021-9010.88.4.741
Detert, J. R., & Burris, E. R. (2007). Leadership behavior and employee voice: Is the door really open?. Academy of Management Journal, 50(4), 869-884. doi:10.5465/amj.2007.26279183
23 Doosje, B., Ellemers, N., & Spears, R. (1995). Perceived intragroup variability as a function of group status and identification. Journal of Experimental Social Psychology, 31, 410-436. doi:10.1006/jesp.1995.1018
Dunning, D., Griffin, D. W., Milojkovic, J. D., & Ross, L. (1990). The overconfidence effect in social prediction. Journal of personality and social psychology, 58(4), 568-581. doi:10.1037/0022-3514.58.4.568
Edmondson, A. (1999). Psychological safety and learning behavior in work teams. Administrative
science quarterly, 44, 350-383. doi: 10.2307/2666999
Edmondson, A. C., Roberto, M. A., & Watkins, M. D. (2003). A dynamic model of top management team effectiveness: managing unstructured task streams. The Leadership Quarterly, 14, 297– 325. doi:10.1016/s1048-9843(03)00021-3
Gilson, L. L., & Shalley, C. E. (2004). A little creativity goes a long way: An examination of teams’ engagement in creative processes. Journal of Management, 30, 453–470. doi: 10.1016/j.jm.2003.07.001
Hinds, P. J., & Mortensen, M. (2005). Understanding conflict in geographically distributed teams: The moderating effects of shared identity, shared context, and spontaneous communication. Organization science, 16, 290-307. doi: 10.1287/orsc.1050.0122
Hirst, G., Van Dick, R., & Van Knippenberg, D. (2009). A social identity perspective on leadership and employee creativity. Journal of Organizational Behavior, 30(7), 963-982. doi: 10.1002/job.600
Janssen, O., & Huang, X. (2008). Us and me: Team identification and individual differentiation as complementary drivers of team members' citizenship and creative behaviors. Journal of
24 Judge, T. A., Piccolo, R. F., & Kosalka, T. (2009). The bright and dark sides of leader traits: A review and theoretical extension of the leader trait paradigm. The Leadership Quarterly, 20(6), 855-875. doi:10.1016/j.leaqua.2009.09.004
Kark, R., & Carmeli, A. (2009). Alive and creating: The mediating role of vitality and aliveness in the relationship between psychological safety and creative work involvement. Journal of
Organizational Behavior, 30(6), 785-804. doi:10.1002/job.571
Kessel, M., Kratzer, J., & Schultz, C. (2012). Psychological safety, knowledge sharing, and creative performance in healthcare teams. Creativity and innovation management, 21, 147-157. doi:10.1111/j.1467-8691.2012.00635.x
Kozlowski, S. W., & Ilgen, D. R. (2006). Enhancing the effectiveness of work groups and teams. Psychological science in the public interest, 7(3), 77-124. doi:10.1111/j.1529-1006.2006.00030.x
Mael, F., & Ashforth, B. E. (1992). Alumni and their alma mater: A partial test of the reformulated model of organizational identification. Journal of organizational Behavior, 13, 103-123. doi:10.1002/job.4030130202
McGraw, K. O., & Wong, S. P. (1996). Forming inferences about some intraclass correlation coefficients. Psychological methods, 1, 30-46. doi:10.1037/1082-989x.1.1.30
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of
Applied Psychology, 88, 879-903. doi:10.1037/0021-9010.88.5.879
Oldham, G. R., & Cummings, A. (1996). Employee creativity: Personal and contextual factors at work. Academy of management journal, 39, 607-634.
Paulus, P. B., & Dzindolet, M. (2008). Social influence, creativity and innovation. Social Influence, 3, 228-247. doi:10.2307/256657
25 Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879-891. doi:10.3758/brm.40.3.879
Preacher, K. J., Zhang, Z., & Zyphur, M. J. (2011). Alternative methods for assessing mediation in multilevel data: The advantages of multilevel SEM. Structural Equation Modeling, 18, 161-182. doi:10.1080/10705511.2011.557329
Preacher, K. J., Zyphur, M. J., & Zhang, Z. (2010). A general multilevel SEM framework for assessing multilevel mediation. Psychological methods, 15, 209-233. doi:10.1037/a0020141 Sawyer, R. K. (2006). Educating for innovation. Thinking skills and creativity, 1, 41-48.
doi:10.1016/j.tsc.2005.08.001
Schein, E., & Bennis, W. (1965). Personal and organizational change through group methods. New York: Wiley
Shin, Y., & Eom, C. (2014). Team Proactivity as a Linking Mechanism between Team Creative Efficacy, Transformational Leadership, and Risk‐Taking Norms and Team Creative Performance. The Journal of Creative Behavior, 48(2), 89-114. doi:10.1002/jocb.42
Tabachnick, B. G., & Fidell, L. S. (2001). Multilevel linear modelling. In Using multivariate statistics (pp. 781-857) San Francisco, CA: Pearson.
Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin, & S. Worchel, The Social Psychology of Intergroup Relations (pp. 94-109). Montery, CA: Brooks-Cole.
Tierney, P., & Farmer, S. M. (2011). Creative self-efficacy development and creative performance over time. Journal of Applied Psychology, 96, 277-293. doi:10.1037/a0020952
Tyran, K. L., Tyran, C. K., & Shepherd, M. (2003). Exploring emerging leadership in virtual teams. Virtual teams that work: Creating conditions for virtual team effectiveness, 183-195.
26 Van Der Vegt, G. S., & Bunderson, J. S. (2005). Learning and performance in multidisciplinary teams: The importance of collective team identification. Academy of Management Journal, 48, 532-547. doi:10.5465/amj.2005.17407918
Van Knippenberg, D., & Sitkin, S. B. (2013). A critical assessment of charismatic—Transformational leadership research: Back to the drawing board?. The Academy of Management Annals, 7, 1-60. doi:10.1080/19416520.2013.759433
Van Knippenberg, D., Van Knippenberg, B., De Cremer, D., & Hogg, M. A. (2004). Leadership, self, and identity: A review and research agenda. The Leadership Quarterly, 15(6), 825-856. doi:10.1016/j.leaqua.2004.09.002
Vessey, W. B., Barrett, J. D., Mumford, M. D., Johnson, G., & Litwiller, B. (2014). Leadership of highly creative people in highly creative fields: A historiometric study of scientific leaders. The Leadership Quarterly, 25(4), 672-691. doi:10.1016/j.leaqua.2014.03.001
Wang, P., & Rode, J. C. (2010). Transformational leadership and follower creativity: The moderating effects of identification with leader and organizational climate. Human relations, 63(8), 1105-1128. doi:10.1177/0018726709354132
Zaccaro, S. J., Rittman, A. L., & Marks, M. A. (2002). Team leadership. The Leadership
Quarterly, 12(4), 451-483. doi:10.1016/S1048-9843(01)00093-5
Zhang, Z., Zyphur, M. J., & Preacher, K. J. (2009). Testing multilevel mediation using hierarchical linear models problems and solutions. Organizational Research Methods, 12(4), 695-719. doi:10.1177/1094428108327450
27
Appendix
Perceived leadership:
Onze band heeft een duidelijke leider.
Het is duidelijk wie meestal de muzikale koers uitzet in onze band. Er is iemand die over het algemeen de leiding neemt in de band. Iedereen is het er over eens wie de band-leider(s) is (zijn).
Team identity:
Ik voel me sterk verbonden met de band. Ik ben blij dat ik tot de band behoor.
Ik voel me verbonden met de leden van de band. Ik ben trots dat ik bij deze band hoor.
Als iemand positief is over de band, voelt dit als een persoonlijk compliment.
Psychological safety:
Als je een fout maakt in de band wordt het nooit tegen je gebruikt.
Mijn bandgenoten en ik zijn in staat om moeilijke kwesties en problemen bespreekbaar te maken. Ik ben niet bang om een risico te nemen tijdens het spelen met de band.
Niemand in de band zal opzettelijk mijn prestaties dwarsbomen.
28 Other-rated creative performance:
In hoeverre zijn de muzikanten technisch sterk?
Hoe authentiek / sterk waren de liedjes, zowel melodie als tekst?
Hoe creatief / origineel / vernieuwend is het optreden? In hoeverre ontwikkelt de band nieuwe ideeën /sounds/ manieren van spelen - die origineel zijn en tegelijkertijd radiowaardig?
Hoe artistiek is het geluid en het optreden van deze band? In hoeverre was de performance mooi, kunstzinnig met eigen geluid?
Hoe radiowaardig is deze band? Hoe goed slagen ze er in om een nieuwe sound toegankelijk/radiowaardig te maken?
In hoeverre heeft het optreden jou en/of het publiek geraakt, ontroerd, positieve gevoelens gebracht? In hoeverre heeft het optreden energie overgebracht, een activerende impact op jou en/of het publiek gehad?
In hoeverre was de band in staat om het publiek te raken/ contact te maken / mensen mee te krijgen? Was men op elkaar ingespeeld? Was er sprake van goed samenspel / luisteren naar elkaar; geheel meer dan de som der delen?
Subjective creative performance:
In hoeverre ontwikkelde de band nieuwe sounds / manieren van spelen - die origineel zijn en tegelijkertijd toegankelijk voor publiek?
Hoe authentiek / sterk waren de (arrangementen van de) liedjes, zowel melodie als tekst? In hoeverre vond je jullie performance artistiek, mooi / kunstzinnig met een eigen geluid?
29 In hoeverre zijn de muzikanten van jullie band technisch sterk?
Hoe radiowaardig vind je jullie als band? Hoe goed slagen jullie er in om een nieuwe sound toegankelijk / geschikt voor de radio te maken?
Waren jullie goed op elkaar ingespeeld? Was er sprake van goed samenspel / luisteren naar elkaar? In hoeverre heeft het optreden het publiek geraakt, ontroerd, positieve gevoelens gebracht?
In hoeverre heeft het optreden energie overgebracht, een activerende impact op het publiek gehad? In hoeverre was de band in staat om contact met het publiek te maken / mensen mee te krijgen?
