Abstract
Theoretical models of individual motivation in groups represent overt effort intentions as precursors of observable effort expenditure in a group context. We examined established triggers of group motivation gains in a scenario-based paradigm, exploring which of these triggers are already manifested at the level of effort intentions. We expected higher effort intentions during teamwork as compared with individual work when teamwork enabled one of the following processes: social compensation, social comparison, or social indispensability. Fifty-seven basketball players (Study 1) and 97 adolescents (Study 2) were asked to imagine individual and team sports situations and to indicate their intended effort in these situations. Features of the team situations were manipulated following a 2 (task demands: conjunctive vs. additive) × 4 (partner performance: inferior, equally strong, moderately superior, very superior) design. Results showed that social compensation, social comparison, and social indispensability were already at work at the level of overt effort intentions.
Keywords
In addition to the demotivating effects of teamwork (e.g., social loafing; cf. Karau & Williams, 1993, for a review), the motivating effects of teamwork are established and well documented (see Larson, 2010; Weber & Hertel, 2007, for reviews). Most of these studies have used performance data as a reliable indicator of motivational differences between teamwork and individual work. Indeed, performance indicators have various benefits as compared with self-report measures or physiological measures, such as higher robustness against social demands or self-deception and better generalizability to applied settings. However, performance indicators reflect the end product of a complex process of motivation. To understand the motivating effects of group settings on different phases of individuals’ action regulation process (e.g., Heckhausen & Gollwitzer, 1987), it might be useful to focus on different stages of individual team members’ action regulation. The current study provides an initial contribution focusing on the motivating effects of teamwork on the level of initial effort intentions that are activated by social cues (either strategically or heuristically) even before the actual group interaction has begun.
Indeed, apart from explanations of motivating team effects that refer to psychological processes during teamwork (e.g., mere presence effects or social support from fellow team members; see, for instance, Zajonc, 1965, and Hüffmeier & Hertel, 2011b), many if not most explanations of motivating (and demotivating) effects of teamwork focus on team members’ intentions to invest personal resources (i.e., effort, time) as a precursor to their actual effort expenditure in the team task. For instance, expectancy × value models of motivation in group contexts (e.g., Karau & Williams, 1993; Shepperd, 1993) assume that group members either deliberatively analyze the costs and benefits of the group situation or react heuristically to the social cues associated with preexisting scripts of effort expenditure in groups (e.g., norms of social responsibility or group loyalty; cf. Hertel & Kerr, 2001). In both cases, individual team members are assumed to build an intention that specifies how much effort they are willing to invest in the following team task. Thus, at least a major part of the motivating process is assumed to happen before the actual teamwork begins. However, there is a lack of research that has explicitly tested this theoretical assumption. Moreover, showing that (and which) established triggers of motivating team effects have already been manifested at the level of overt effort intentions might also have interesting implications for the ease of applying these principles in occupational teams. However, if the motivating effects of teamwork are based mainly on conscious (e.g., social support) or subconscious processes (e.g., mere presence) during teamwork, the established triggers of motivating effects of teamwork should have no effect on effort intentions before the group interaction begins.
In the current study, we systematically manipulated established triggers of both motivation gains and motivation losses in groups as compared with individual work (such as social compensation, social comparison, social indispensability, and the free-rider effect), and examined which of these factors already affected individuals’ motivation at the level of overt effort intentions. Moreover, we considered different trigger factors in the same study, enabling a comparison of the relative strengths of these factors on individuals’ effort intentions. Apart from its theoretical implications (e.g., initial hints on cognitive precursors of effort expenditure in groups), this comparison might also provide useful guidance for managers of existing teams regarding which trigger factors might be most fruitful for open discussions, for motivating communication, and also for designing motivating team tasks.
Established Triggers of Motivating Effects of Teamwork
The collective effort model (CEM; Karau & Williams, 2001, 1993), as perhaps the most influential model of motivation in groups, specifies how team members integrate various (deliberate or automatic) assessments of action consequences to decide how much effort they want to expend during team work. It thereby predicts and explains the emergence of effort losses and gains when people work in teams as compared with working individually (Karau, Markus, & Williams, 2000). Rooted in a general expectancy × value framework (e.g., Atkinson, 1958; Vroom, 1964), the CEM proposes that team members’ intentions to expend effort on a task is determined by three independent factors. Expectancy designates the expectation that high individual effort will lead to high individual performance, whereas instrumentality comprises three expectation components: namely, the expectations that high individual performance will lead to high team performance, that high team performance will lead to a desired outcome, and that desired outcomes for the team will lead to desired individual outcomes. Valence finally designates the perceived value of the achievable outcome. According to the CEM, team members will exert their highest level of effort when all three factors are high (Karau & Williams, 1993).
In the current research, we specifically focused on the role of perceived instrumentality for team members’ effort intentions. Indeed, research has indicated that team members will be highly motivated if their individual performance is indispensable for the team and will help the team to achieve a desired outcome (e.g., Hertel, Niemeyer, & Claus, 2008; Kerr et al., 2007). Whether or not team members will perceive their individual performance to be instrumental for high team performance depends on both task demands and the relative abilities of fellow team members (e.g., Hertel, Kerr, & Messé, 2000; Steiner, 1972). According to Steiner, task demands can for instance be classified as additive (i.e., the team’s performance is determined by the sum of all individual performances), conjunctive (i.e., the weakest individual performance solely determines the team’s performance), or disjunctive (i.e., the best individual performance solely determines the team’s performance). Obviously, the relative strength of an individual team member compared with the other team members is thus crucial for his or her instrumentality perceptions and related effort intentions. Indeed, team members’ perceptions of instrumentality have been demonstrated to underlie several motivation loss and gain phenomena, such as the free-rider effect (e.g., Kerr & Bruun, 1983), social compensation (e.g., Williams & Karau, 1991), and social indispensability effects (e.g., Hertel et al., 2000, 2008; Kerr & Hertel, 2011).
The tendency of team members to reduce their effort when they perceive that their individual performance is not instrumental for the team goals has been described as free riding (e.g., Kerr & Bruun, 1983). This effect is most likely to occur among high ability team members when the team’s performance depends on low ability members (i.e., under conjunctive task demands) and among low ability team members when the team’s performance depends on high ability members (i.e., under disjunctive task demands). Free riding can also occur in large teams working under additive task demands because the instrumentality of each individual contribution decreases with increasing team size (cf. Stroebe, Diehl, & Abakoumkin, 1996).
Likewise, it is possible that team members will show gains in effort for additive group tasks when they perceive their individual performance to be instrumental for high team performance (a social compensation effect; Williams & Karau, 1991). This is particularly likely for high ability team members who perceive their contribution to be highly instrumental for the team’s goals because they can compensate for the poor performance of other team members. The resulting effort gain in groups (viz. the social compensation effect) was initially demonstrated by Williams and Karau (1991) and has been replicated in various studies (e.g., Hart, Bridgett, & Karau, 2001; Todd, Seok, Kerr, & Messé, 2006). Social compensation was demonstrated not only when participants expected their partners to be low in motivation but also when they expected their partners to be simply not capable of performing well (Williams & Karau, 1991). Results further revealed high task meaningfulness to be a precondition for the occurrence of social compensation (Williams & Karau, 1991, Study 3).
Another type of effort gain in groups caused by high instrumentality perceptions is described by the Köhler effect (e.g., Kerr & Hertel, 2011; Stroebe et al., 1996). In his initial experiments, Köhler (1926) demonstrated that team members will exert greater effort compared with individual work when performing with a superior partner on a conjunctive group task (Steiner, 1972). In addition, Köhler observed that the relation between relative member ability and effort followed an inverted u-shaped function with maximum increments in effort for moderate ability discrepancies between the inferior and superior team member. To date, various studies have replicated the overall Köhler effect (for reviews, see Kerr & Hertel, 2011; Weber & Hertel, 2007) as well as the Köhler discrepancy effect (e.g., Messé, Hertel, Kerr, Lount, & Park, 2002; Stroebe et al., 1996, experiment 1). Moreover, Messé et al. (2002) have demonstrated that knowledge of one’s partner’s ability is a requirement for the occurrence of the Köhler discrepancy effect.
Regarding the Köhler motivation gain effect, research has identified two basic trigger mechanisms: upward social comparison and perceived social indispensability (e.g., Hertel et al., 2008; Kerr et al., 2007; Weber & Hertel, 2007). According to Festinger’s social comparison theory (SCT; 1954), people have a need to evaluate their own performance, and thus, to compare their performance levels with those of others in the absence of objective evaluation criteria, especially if the setting is evaluative or competitive (cf. Seta, 1982). Festinger further assumed that people have a general tendency to make upward comparisons. People may then apply the performance level of a superior other as a standard or goal to define success and failure. Comparisons with superior others have accordingly been demonstrated to enhance individual performance (e.g., Munkes & Diehl, 2003; Seta, 1982) as long as the difference in performance levels is not too large (e.g., Seta, Seta, & Donaldson, 1991).
The second process contributing to the Köhler motivation gain effect is perceived social indispensability (i.e., high perceived instrumentality of personal contributions for the team output; cf. Hertel et al., 2000). Under conjunctive task demands, inferior team members are likely to perceive that their individual contribution is indispensable for their team because team performance is solely determined by the performance of the weakest member.
Social comparison and social indispensability processes are also invoked as possible explanations for the Köhler discrepancy effect. First, individuals prefer to compare themselves with moderately superior rather than with very superior or similar others (Festinger, 1954). Second, team members experience especially strong social indispensability if their team partner is moderately superior: Team members performing with an equally strong partner do not have to expend extra effort to achieve a desired outcome (e.g., to avoid being seen as the one letting the team down), whereas team members do not expect achieving their superior partner’s performance if the difference in performance levels is too large (cf. Hertel et al., 2000; Seta et al., 1991).
Hypotheses
The present research investigated effects of established triggers of motivating (and demotivating) teamwork on effort intentions in the context of team sports. Participants were asked to imagine different teamwork situations and to indicate their effort intentions. Imagined teamwork situations included either conjunctive or additive task demands, and the participants were asked to imagine that they had to work with an inferior, an equally strong, a moderately superior, or a very superior team partner (see Table 1 for an overviews of the design and hypotheses). The variation of both task demands and partner capability enabled the operationalization of various triggers of motivation gains and losses that have already been well documented in prior research using performance measures as indicators of motivation. However, in the present study, we examined whether these triggers had already led to significant motivating effects of teamwork as compared with an individual work baseline on effort intentions expressed before the actual teamwork had begun. In general, we expected that both instrumentality and social comparison processes would already show significant motivating group effects at the level of overt effort intentions. Such manifestations would confirm the cognitive tenet of expectancy × value models of motivation (e.g., Karau & Williams, 1993; Shepperd, 1993). The effects predicted in Hypotheses 1 to 4 share the quality that one single process is the driving force responsible for team members’ levels of motivation.
Design and Hypotheses of This Research.
When the team’s performance is solely determined by the performance of the weakest member (i.e., under conjunctive task demands; cf. Kerr & Bruun, 1983), superior team members should perceive their contributions as not instrumental for achieving a good performance and should thus reduce their effort relative to working alone (i.e., when working individually, one’s own performance is more instrumental for a good performance). Thus, we hypothesized:
Hypothesis 1: Persons expecting to work with an inferior team partner under conjunctive task demands will intend to invest less effort as compared with working alone (i.e., the free-riding effect).
Based on an expectancy × value framework (e.g., Karau & Williams, 1993; Shepperd, 1993), it can be expected that superior team members will perceive their performance as particularly instrumental when working with an inferior team partner if all team members’ contributions’ jointly determine the team’s performance (i.e., additive task demands). When the task is sufficiently meaningful (i.e., the valence component of the framework is positive), team members are expected to compensate for another team member’s expected low contribution (cf. Todd et al., 2006; Williams & Karau, 1991). Thus, we predicted:
Hypothesis 2: When the task is meaningful, persons expecting to work with an inferior team partner under additive task demands will intend to invest more effort as compared with individual work (i.e., social compensation).
In the absence of more objective performance information, team members can be expected to compare their performance with their fellow team members according to social comparison theory (Festinger, 1954) and to try to do as well as or even outperform them, especially in evaluative or competitive settings (e.g., Seta, 1982). Prior research has demonstrated that team members gain motivation under additive and conjunctive task demands even if they are similar in ability (e.g., Messé et al., 2002; Seta, 1982; Stroebe et al., 1996). It is important to keep in mind that no additional increase in perceived instrumentality (e.g., Karau & Williams, 1993; Shepperd, 1993) should occur under conjunctive task demands when the capabilities of both team partners are perceived to be about the same because both team members can be expected to feel equally responsible for the team’s output and no team member feels as though he or she is holding the team back. Thus, increases in effort that occur when a person works with an equally strong partner should be mainly due to social comparison.
Social comparisons are also responsible for the expected intention of team members to increase their effort when working with a moderately superior fellow team member under additive task demands. This prediction holds only for small teams working under additive task demands when the team’s performance is thus determined by only a few individual contributions. Therefore, motivation losses of inferior team members due to low perceived instrumentality are unlikely to occur (Stroebe et al., 1996). In such situations, individuals should rather compare their performance levels with others whom they perceive as slightly superior and increase their effort to reach this higher level of performance or even to outperform the superior members (e.g., Seta, 1982; Stroebe et al., 1996). Thus, we predicted:
Hypothesis 3: Persons expecting to work with an equally strong team partner (a) under conjunctive or (b) additive task demands and (c) persons expecting to work with a moderately stronger team partner under additive task demands will intend to invest more effort as compared with individual work (i.e., social comparison).
Social comparison theory (Festinger, 1954) predicts that social comparisons become less likely with an increasing divergence in people’s abilities. Thus, effort should be less pronounced when inferior team members work with very superior as compared with moderately superior team members under conjunctive task demands (cf. Messé et al., 2002) because inferior team members will compare their performance less with very superior team partners. Inferior team members should, however, still show gains in effort as they perceive their contribution to be indispensable for the team’s outcomes because they are still determining the performance of their respective team. Thus, we hypothesized:
Hypothesis 4. Persons expecting to work with a very superior team partner under conjunctive task demands will intend to invest more effort as compared with individual work (i.e., social indispensability).
Following the same theoretical argument, comparisons with very superior fellow team members are also unlikely when working under additive task demands. When comparing themselves with their very superior fellow team members, inferior team members do not seem to benefit motivationally (cf. Seta et al., 1991). As team members under additive task demands do not perceive indispensability of their efforts, we predicted:
Hypothesis 5: Persons expecting to work with a very superior team partner under additive task demands will intend to invest levels of effort that are similar to what they would expend when performing individual work.
In contrast to the previous predictions, we expected that inferior team members’ effort intentions would be fueled by two different motivational processes when expecting to work under conjunctive task demands: These conditions not only afford social comparison effects vis-à-vis a superior fellow team member but also provide indispensability effects for those team members who are inferior in strength (Weber & Hertel, 2007). In other words, when working with a moderately superior fellow team member, positive comparison effects on motivation should occur according to social comparison theory (e.g., Festinger, 1954). Independent of the positive effects of social comparison, the perception that one’s own individual performance determines the team’s output under conjunctive task demands should additionally increase instrumentality perceptions (e.g., Karau & Williams, 1993; Shepperd, 1993). Thus, we predicted:
Hypothesis 6a. Persons expecting to work under conjunctive task demands with a moderately superior team partner will intend to invest particularly high effort as compared with individual work (i.e., an effort increase due to both social comparison and social indispensability).
Social comparison and social indispensability were accordingly expected to function as additive, independent effects (cf. Hertel et al., 2008; Weber & Hertel, 2007). We therefore expected that inferior team members’ effort intentions would be more pronounced when working with a moderately superior partner under conjunctive task demands (i.e., when two motivating processes operate) than under all conditions in which effort intentions are positively influenced by only one such process. Thus, we predicted:
Hypothesis 6b. Persons will intend to exert higher effort in the conjunctive task/moderately superior partner condition (due to two additive processes, i.e., social comparison and social indispensability) as compared with conditions in which effort gains are expected to be driven by only one process (additive task/inferior partner, additive task/equally strong partner, conjunctive task/equally strong partner, additive task/moderately superior partner, conjunctive task/very superior partner).
Study 1
Method
Participants and procedure
In Study 1, we approached 70 basketball players from four men’s and two women’s teams from the first to third amateur leagues within one federal state of Germany during their regular training sessions. We asked them to participate in a study on behavioral tendencies in practice situations by completing a questionnaire at home, which took between 20 and 25 min. All players agreed to participate in the study. They were handed questionnaires in prepaid return envelopes. Fifty-seven players (14 female) sent back the questionnaire (a response rate of 81%). Their average age was 25.6 years (SD = 6.8).
Measures and design
The study followed a 2 (task demands: conjunctive vs. additive) × 4 (partner performance: inferior vs. equally strong vs. moderately superior vs. very superior) × 2 (scenario order: order 1 vs. order 2) design with the first two factors as within-subjects factors and participants’ effort intentions as the dependent variable. The questionnaire included the description of nine similar basketball practice scenarios in which participants were asked to imagine running a 1,000-meter run, a typical activity in the preparatory stage of a basketball season. The first scenario in both scenarios was an individual scenario, which was used to provide the participants with a reference point of their regular effort intention level for individual practice situations. The ratings for this scenario were not included in the analysis. The scenario read as follows: Please imagine now that you are in the preparatory stage of a new basketball season. You are aware that the performance of the players in the preparatory stage is particularly important to your coach. A conditioning session is on today’s schedule. Your coach has decided that each player has to individually complete a 1,000-meter run. This means that your personal success is determined solely by your individual time. How will you behave in this run?
The remaining eight scenarios described team practice situations in which the independent variables task demands (conjunctive vs. additive) and partner performance (inferior vs. equally strong vs. moderately superior vs. very superior) were manipulated in the instructions. For each of these situations, participants were asked to imagine that they were performing with a concrete and alternating member of their basketball team. The relevant information was printed in bold to attract participants’ attention. For instance, the scenario with a conjunctive task structure and an inferior team partner read as follows (formulations for an additive task structure and for an equally strong, a moderately superior, and a very superior team partner are in parentheses): At this time, please imagine: Your coach divides the team into dyads, which have to complete a 1000-meter run as a team. Your coach decides for this 1,000-meter run that the time of the slower dyad member will count as the dyad’s performance. This means that the second person to reach the finish line will determine the time and thus, the success of the dyad (that the accumulated time of both dyad members will count as the dyad’s performance. This means that the time of both dyad members together will determine the time and thus, the success of the dyad). During the run, you notice that your dyad partner is running considerably behind you (directly next to you, a short distance ahead of you, considerably ahead of you). How will you behave in this run compared to a run in which you are running individually for yourself?
To ensure that the team partner’s performance represented the only available performance information, we ensured that the scenarios did not provide any objective performance standard (e.g., interim times). After each scenario, we measured participants’ self-reported effort intentions as our main dependent variable with two items (“How much effort will you expend in this run?” and “How much dedication will you show during this run?”). Given that individual (non-team) performance was our central baseline, the two items were rated on 7-point scales ranging from 0 (much less compared with running individually) to 6 (much more compared with running individually) with a scale midpoint of 3 (as much as when running individually). Both items were adapted from earlier studies on motivation gains in teams (e.g., Hertel et al., 2000) and showed satisfactory reliabilities for all scenarios in the present study (rs between .80 and .96). They could thus be averaged to yield an effort intention score for each scenario. At the end of the questionnaire, participants had to indicate their gender, age, and league. To control for effects of scenario order, scenarios were put into two different random orders resulting in a 2 (task demands: conjunctive vs. additive) × 4 (partner performance: inferior vs. equally strong vs. moderately superior vs. very superior) × 2 (scenario order: order 1 vs. order 2) study design with the first two factors as within-subject factors and the last as a between-subjects factor. The questionnaire was presented in German language.
Results
To examine differences in effort intention ratings between the experimental conditions, we first conducted a 2 (task demands) × 4 (partner performance) × 2 (scenario order) repeated-measures analysis of variance (ANOVA) of participants’ effort intention ratings with the first two factors as within-subjects factors. This analysis revealed significant main effects of task demands, F(1, 56) = 6.15, p < .05, f = 0.33, and partner performance, F(3, 168) = 43.09, p < .001, f = 0.88, as well as a significant interaction of these two factors, F(3, 168) = 25.80, p < .001, f = 0.68. Moreover, no order effect occurred in this ANOVA, F < 1 (for cell means and standard deviations, see Table 2).
Means and Standard Deviations of Effort Intentions as a Function of Task Demands and Partner Performance (Study 1; N = 57).
Note. Effort intentions were rated on 7-point scales (0 = much less compared to running individually, 3 = as much as when running individually, 6 = much more compared to running individually).
Given our specific contrast predictions, we tested our hypotheses via a series of planned contrasts (Rosenthal & Rosnow, 1985). To test for team effort losses and gains, we compared the level of indicated effort in each team condition with the scale midpoint (i.e., 3 = as much as when running individually). Mean scores below the scale midpoint indicate an intention to reduce one’s effort when working on a team, whereas mean scores above the scale midpoint indicate an intention to expend more effort during teamwork.
First, our results indicated no reduction in effort intentions in the conjunctive task/inferior partner condition as compared with running individually. The related ratings (M = 3.1) were not significantly different from the scale midpoint (3; t < 1). Thus, Hypothesis 1 was not supported. However, consistent with Hypotheses 2 to 4 and 6a, t tests revealed significant increases in intended effort compared with the individual baseline in the following teamwork conditions: conjunctive task/equally strong partner (M = 4.1), t(56) = 6.98, p < .001, d = 0.93; conjunctive task/moderately superior partner (M = 5.0), t(56) = 16.09, p < .001, d = 2.14; conjunctive task/very superior partner (M = 5.1), t(56) = 17.16, p < .001, d = 2.27; additive task/inferior partner (M = 4.3), t(56) = 7.64, d = 1.02; additive task/equally strong partner (M = 4.3), t(56) = 9.79, p < .001, d = 1.2; and additive task/moderately superior partner (M = 4.8), t(56) = 12.61, p < .001, d = 1.67. Moreover, supporting Hypothesis 6b, a planned contrast showed that effort intentions in the conjunctive teamwork condition with a moderately superior partner (M = 5.0) were higher than in the conditions in which one rather than two additive underlying processes was predicted to drive the expected motivation gains (additive task/inferior partner, additive task/equally strong partner, conjunctive task/equally strong partner, additive task/moderately superior partner, conjunctive task/very superior partner), M = 4.5, t(56) = 4.53, p < .001, d = 0.61. However, inconsistent with Hypothesis 5, we also found intentions to increase effort during teamwork in the additive task/very superior partner condition, M = 4.7, t(56) = 11.76, p < .001, d = 1.56.
To further explore this pattern of results, we conducted Bonferroni post hoc tests with an adjusted significance level of α = .01. These comparisons showed that effort intentions were significantly higher under conjunctive than under additive task demands when the participants imagined working with a very superior partner, t(56) = 3.68, p ≤ .001, d = 0.49. However, effort intentions were not different at the adjusted significance level under conjunctive and additive task demands with a moderately superior partner, t(56) = 2.16, p = .035, d = 0.29. Moreover, under additive task demands, effort intentions were significantly higher when participants imagined working with a moderately superior compared with an equally strong partner, t(56) = 3.45, p ≤ .001, d = 0.46. Differences in effort intentions between the moderately superior and very superior conditions were not observed under additive task demands, t < 1, ns, or under conjunctive task demands, t < -1.1, ns.
Discussion
In line with our hypotheses, we found empirical evidence for the assumption that established triggers of motivation gains are preceded by effort intentions that are conscious and openly communicated. More precisely, we found three different types of increases in effort intentions during teamwork in a sports training context with amateur basketball players. First, participants reported increased effort intentions as compared with the individual training baseline when envisioning that they were exercising with an inferior team partner under additive task demands. This effect is consistent with Hypothesis 2 and with theories of social compensation (e.g., Williams & Karau, 1991). Second, participants also reported increased effort intentions as compared with the individual training baseline when imagining exercising with an equally strong partner under both conjunctive and additive task demands as well as with a moderately superior partner under additive task demands. This effect is consistent with Hypothesis 3 and with theories on social comparison processes (e.g., Festinger, 1954; Seta, 1982). Third, participants reported increased effort intentions as compared with an individual training baseline under conjunctive task demands both with a moderately and a very superior team partner. This effect is consistent with Hypotheses 4 and 6a and with theories on social indispensability (e.g., Hertel et al., 2000, 2008).
However, no evidence emerged for a Köhler discrepancy effect because the effort intentions were not less pronounced when participants imagined exercising conjunctively with a very superior as compared with a moderately superior team partner. We further found that participants indicated unexpected intentions to increase their effort during teamwork when team members envisioned exercising with a very superior team partner under additive conditions. On the other hand, in contrast to Hypothesis 1, we found no evidence of free riding (cf. Kerr & Bruun, 1983) in the reported effort intentions as the only predicted type of group-related motivation losses in this research.
Together, our results suggest that team-related motivation gains due to the different processes demonstrated in laboratory research (i.e., social compensation, social comparison, and social indispensability) can also be observed in individuals’ conscious intentions to invest effort in a given task. This finding provides support for the assumption that motivation gains in groups are at least partly driven by consciously reflected decision processes (Karau & Williams, 1993; Shepperd, 1993).
The absence of observed motivation losses in groups in the current study might be interpreted as an indication that motivation losses in groups are generally less pronounced in field settings than in laboratory settings. However, it might also signal that—different from motivation gains in groups—motivation losses in groups are based on unconscious or automatic regulation processes instead of reflected effort decisions. A possible methodological explanation might finally focus on the within-subjects design of this study. As a consequence of this approach, participants might have transferred their general response tendency that was a result of the predominantly motivating scenarios to the one scenario in which we expected motivation losses. As the absence of motivation losses was unexpected, a replication of this finding seemed warranted by all means.
It should be further noted that the study’s results were observed in a sample of amateur basketball players who devote a considerable amount of their time to team (sports) activities. Individuals who have a long history in team sports are supposedly characterized by a specific set of values, attitudes, and behavioral tendencies. Thus, a follow-up study was conducted to replicate our findings with a nonathlete sample.
Study 2
In Study 2, we replicated Study 1 by testing students from different high schools recruited during visiting days in the psychology department of a German university.
Method
Participants and procedure
Participants were 97 adolescents and young adults (69 female and 26 male; 2 participants did not specify their gender) with an average age of 18.4 years (SD = 1.7). The unequal number of males and females in Study 1 as well as in Study 2 reflects the skewed distribution of these subsamples in the two populations: Whereas more males than females play organized basketball in Germany, more females than males are interested in (and end up) studying psychology. Participants were asked to take part in a study on behavioral tendencies in physical education lessons. Participants were randomly assigned to the order conditions and filled out questionnaires in the presence of an investigator (completion time between 20 and 25 min).
Measures and design
The questionnaire used in Study 2 was identical to the questionnaire used in Study 1, except that the basketball practice scenarios were adapted to a physical education school context. The first paragraph of all team scenarios read as follows: Please imagine that you are in a physical education lesson at school. A conditioning session is on today’s schedule. You know that student performance in the endurance run is particularly important to your teacher. You know also that your performance in this run is important for the physical education mark on your school certificate.
The following paragraphs of the eight team scenarios that included information about the independent variables were identical to Study 1 except that the term coach was replaced by the term teacher and that participants were asked to imagine that they were performing with different members of their class in each scenario. The same two items and the same response format as in Study 1 were applied to measure effort intentions and showed satisfactory reliabilities for all eight scenarios (rs between .79 and .92). As in Study 1, participants indicated their gender and age at the end of the questionnaire. In addition, participants were asked whether they were active members of sports teams. Forty-three participants (43.3% of the sample) reported that they were not active members of a sports team. Finally, participants rated their preference for group versus individual work on two items of the relevant subscale from the Beliefs about Groups Scale (Karau & Elsaid, 2009; “I prefer group work to individual work” and “Whenever possible, I like to work with others rather than by myself”; r = .77).
Results
Forty-seven randomly missing values (3.0%) on the 16 effort intention items were estimated via expectation maximization methods. Similar to Study 1, a 2 (task demands: conjunctive vs. additive) × 4 (partner performance: inferior vs. equally strong vs. moderately superior vs. very superior) × 2 (scenario order: order 1 vs. order 2) analysis of variance (ANOVA) with the rated effort intentions as dependent variable was conducted. The ANOVA revealed significant main effects of task demands, F(1, 95) = 21.92, p < .001, f = 0.47, and partner performance, F(3, 285) = 61.09, p < .001, f = 0.80. These main effects were qualified by a task demands × partner performance interaction, which is consistent with our hypotheses, F(3, 285) = 26.77, p < .001, f = 0.52, and a significant three-way interaction, F(3, 285) = 2.73, p < .05, f = 0.17. The three-way interaction signals a significant influence of scenario order on our results (for cell means and SDs, see Table 3). We thus report further results separately for orders 1 and 2 if there is a difference between orders.
Means and Standard Deviations of Effort Intentions as a Function of Task Demands, Partner Performance, and Scenario Order (Study 2).
Note. Effort intentions were rated on 7-point scales (0 = much less compared to running individually, 3 = as much as when running individually, 6 = much more compared to running individually).
Moreover, we conducted an additional ANOVA that included current membership on a sports team (yes vs. no) and the preference for group versus individual work as additional factors. However, because neither membership on a sports team nor the preference for group versus individual work showed any main or interaction effects, Fs < 1.4, data were aggregated across these two factors in the following analyses.
As in Study 1, we tested our hypotheses with a series of planned contrasts. We again compared the level of effort intentions in each team condition with the scale midpoint (i.e., 3 = as much as when running individually).
For both orders 1 and 2, results did not confirm the predicted free-rider effect (Hypothesis 1). For order 1, effort intentions were even higher in the conjunctive task/inferior partner condition as in the individual control condition (M = 3.8), t(48) = 4.37, p < .001, d = 0.63. For order 2, no such differenced occurred (M = 2.7), t < –1.1, ns.
Consistent with Hypotheses 2 to 4 and 6a, one-sample t tests revealed overall intended effort increases across both orders in the following conditions: conjunctive task/equally strong partner (M = 4.3), t(96) = 10.51, p < .001, d = 1.07; conjunctive task/moderately superior partner (M = 4.9), t(96) = 20.85, p < .001, d = 2.12; conjunctive task/very superior partner (M = 5.1), t(96) = 21.65, p < .001, d = 2.20; additive task/inferior partner (M = 4.4), t(96) = 11.19; p < .001, d = 1.14; additive task/equally strong partner (M = 4.4), t(96) = 13.42, p < .001, d = 1.36; and additive task/moderately superior partner (M = 4.8), t(96) = 18.56, p < .001, d = 1.88.
For order 1 and inconsistent with Hypothesis 6b, intended effort increases were only descriptively stronger in the conjunctive task/moderately superior partner condition (M = 5.0) than in the other five conditions that allowed for social compensation only, social comparison only, or social indispensability only, but did not differ statistically (M = 4.8), t < 1.3, ns. For order 2, however, the respective contrast revealed empirical support for Hypothesis 6b by showing significantly stronger effort intentions in this condition (M = 4.9) than in the other five conditions in which we predicted effort increases (M = 4.4), t(47) = 5.58, p < .001, d = 0.81. Moreover, inconsistent with Hypothesis 5 and parallel to Study 1, we found overall significant effort increases in the additive task/very superior partner condition (M = 5.0), t(96) = 19.76, p < .001, d = 2.01.
As in Study 1, we conducted Bonferroni post hoc tests with an adjusted significance level of α = .01 to further explore our pattern of results. In contrast to Study 1, no significant differences occurred between effort intentions under conjunctive and additive task demands when participants imagined working with a very superior partner. Moreover, whereas some further contrasts were significant for one of the two order conditions, none of these contrasts was significant for both order conditions. Therefore, we refrain from reporting the latter because robustness can be questioned.
Discussion
In Study 2, we successfully replicated the basic results of Study 1 by again finding empirical evidence for the assumption that established triggers of motivation gains are preceded by overt effort intentions. We demonstrated intended effort increases in groups as compared with individual work due to social compensation, social comparison, and social indispensability processes. Remarkably, these results were not restricted to members of existing sports teams or individuals who preferred teamwork to individual work (Karau & Elsaid, 2009) as we did not find significant interactions between task demands, partner performance, and any one of these factors.
Further replicating the results of Study 1, we found that participants reported particularly strong gains in effort intentions during group work when they imagined exercising with a very superior partner under additive task demands. In addition, we did not find evidence for the Köhler discrepancy effect: Participants’ effort intentions were similar when they envisioned exercising with a very superior partner as when they envisioned exercising with a moderately superior partner under additive as well as conjunctive task demands. The replication of this pattern in Study 2 implies that the absence of the Köhler discrepancy effect is not specific to athletes in our scenario-based paradigm. Although this unexpected absence should not be overinterpreted, different theoretical and methodological explanations seem possible: Theoretically, participants might have anticipated that working with a superior fellow team member who is relatively familiar might be less frustrating—as was the case in our current study—because one’s own self-efficacy beliefs may be positively influenced by the performance of this familiar team partner (cf. Lent & Lopez, 2002). Methodologically, responses to hypothetical scenarios may well differ from actual effort considerations. An intention to expend high effort levels, for instance, may be a less valid indicator of motivation when merely imagining working with a highly superior fellow team member as compared with actually being in that very situation or even truly expending this effort over time. Investigations focusing on effort intentions in real team situations thus represent a crucial next step for further establishing the validity of these findings.
In addition, there may be a pronounced difference between effort intentions and actual expended effort: Motivation gains of team members who actually perform with a very superior partner might possibly not endure continued cooperation because the experience of permanent inferiority might trigger dissatisfaction (Locke & Latham, 2002), frustration (Seta et al., 1991), or decreased self-efficacy (e.g., Bandura, 1997; Schunk, 1984), and might thus decrease motivation (e.g., Hertel, 2002; Messé et al., 2002). Further research is required to differentiate between these and other plausible explanations for positive social comparison effects of team members who perform with a very superior team partner.
There are plausible methodological and theoretical reasons for the absence of the predicted free-rider effects in both studies. From a methodological point of view, it may have been easier for our participants to indicate effort intentions comparable with the individual scenario than actually expending this kind of effort when actually working with an inferior partner under conjunctive task demands. Alternatively, as a consequence of the within-subjects design of this study, participants may not have sufficiently adjusted their relatively high levels of motivation from previous scenarios to the scenario in which we predicted a free-rider effect (i.e., they may have transferred a general response tendency across scenarios). Apart from these explanations for the absence of motivation losses in this study, there are also theoretical explanations: Participants may have simply been motivated to present themselves in a favorable light (cf. Paulhus, 1984) and may thus have avoided reporting motivation losses, which could have left the impression that they were not dependable or likeable team members. Furthermore, it may be possible that the absence of self-reported motivation losses might also reflect that—different from motivation gains from working in group—motivation losses are rather based on nonconscious or automatic regulation processes instead of consciously reflected effort decisions. Importantly, both explanations are in accordance with previous findings demonstrating that participants who had shown motivation losses in terms of performance did not show social loafing in their self-reports (Karau & Williams, 1993; Williams & Karau, 1991). Finally, a last theoretical explanation focuses on the specific context of our study: In our scenarios, we employed meaningful tasks in a sports context. Such an argument is in line with recent studies (Hüffmeier & Hertel, 2011a; Hüffmeier, Kanthak, & Hertel, 2012; Hüffmeier, Krumm, Kanthak, & Hertel, 2012; Osborn, Irwin, Skoksberg, & Feltz, 2012) that had shown gains in motivation—and no incidents of losses in motivation—in field studies with existing teams in a sports context. Future research should test these different explanations.
Consistent with previous meta-analytic findings (Weber & Hertel, 2007), we showed social indispensability effects when participants imagined working with a moderately superior partner in Study 1 and for order 2 in Study 2. However, we were not able to demonstrate this effect for order 1 in Study 2. We suggest that this nonsignificant finding may be directly due to the specific orders in which scenarios were presented: For instance, the scenario in which we expected an additional increase in motivation due to social indispensability (i.e., conjunctive task/moderately superior partner), was presented first in order 1. Participants might have been reluctant to indicate maximum intentions to expend effort this early in the questionnaire. In order 2, however, participants encountered this scenario after having reported their intentions for two scenarios with less pronounced effort intentions. Thus, they might have been more willing to indicate maximum levels of effort. This and further rather subtle differences between the two scenario orders (e.g., motivation gains in the conjunctive task/inferior partner condition for order 1) may also explain why we found slightly inconsistent findings for orders 1 and 2.
The absence of a significant interaction between task demands, partner performance, and the preference measure indicates that differences in personal preferences for group work are not a moderator of team members’ motivation. However, to eventually exclude a preference for group work as an explanation for our unexpected findings, research that replicates our findings with the full preference subscale of the Beliefs about Groups Scale (Karau & Elsaid, 2009) is needed.
General Discussion
The main objective of this research was to explore whether and which established triggers of motivating (and demotivating) team effects are already manifested at the level of overt effort intentions of team members before the actual team interaction. Although this assumption is (implicitly or explicitly) included in existing theoretical explanations of the motivating effects of teamwork (e.g., Hertel et al., 2000; Karau & Williams, 1993, 2001; Kerr, 1983; Shepperd, 1993; Williams & Karau, 1991), it had not yet been empirically tested. In the present study, we manipulated context conditions of various established motivation gains and losses in teams as compared with individual work. However, instead of measuring performance indicators of individuals’ resulting effort expenditure, we measured participants’ overt effort intentions before (in fact without) implementing the actual teamwork. The results of both studies consistently showed significant increases in effort intentions in group work as compared with individual work due to social compensation, social comparison, and social indispensability for both amateur athletes as well as for young adults with and without a team sports background. Moreover, we also demonstrated summative effects of the different processes. The effort intentions of participants in the condition that allowed both social indispensability and social comparison processes (i.e., conjunctive task demands and a moderately superior partner) were—except for order 1 in Study 2—stronger than those in conditions that allowed for social compensation only, social comparison only, or social indispensability only.
Whereas performance data have specific properties that make them preferable for tests of motivation compared with self-reports (e.g., lower susceptibility to memory and social desirability biases), we were able to gain initial insights into participants’ decision-making processes concerning their effort expenditure during teamwork through the use of self-reports. In accordance with the respective theoretical models (e.g., Karau & Williams, 1993; Shepperd, 1993), our research indicates that at least part of the motivation gains demonstrated in previous research may be preceded by effort intentions that occur before the actual teamwork has begun—due to either deliberate considerations of gains and losses, heuristic reactions to social cues associated with preexisting scripts, or norms regarding effort expenditure. To gain more insight into these different processes, future research that combines behavioral, self-report, and physiological data as indicators of motivation (e.g., Blascovich, Mendes, Hunter, & Salomon, 1999; Cleveland, Blascovich, Gangi, & Finez, 2011; Wittchen, Krimmel, Kohler, & Hertel, 2012) is needed, further differentiating conscious and reflected regulation strategies from nonconscious impulsive regulation strategies (e.g., Strack & Deutsch, 2004).
Interestingly, in contrast to the different triggers of motivation gains found in teams, the only established trigger of motivation losses in this study (the free-rider effect) was not preceded by related conscious effort intentions across the two studies. There is a need for future research that empirically differentiates between social desirability, conscious self-deception, and automatic (as opposed to conscious) effort expenditure decisions as the most obvious explanations for the absence of motivation losses at the level of effort intentions.
Limitations, Practical Implications, and Future Research
The present research is limited in several ways. First, we detected a scenario order effect in Study 2, but not in Study 1. This difference between our studies may well be due to the specifics of our data collection approach: The basketball players filled out the questionnaire at home, and only those players who were sufficiently motivated to do so sent their questionnaires back. By contrast, all adolescents in Study 2 agreed in principle to participate in a study, but they may have grown tired of filling out a rather long questionnaire over time. The presence of an instructor may have prevented them from aborting the study before the end, but they may have been less motivated to devote attention to filling out the questionnaire. This difference in motivation is reflected in a greater share of missing data in Study 2 (3% vs. 0% in Study 1), but could also be reflected in a transfer of response tendencies from a particular scenario to the others that followed, thereby causing order effects. To allow for unambiguous conclusions regarding each part of the observed results, we recommend that future research employs a between-subjects design.
The use of a within-subjects design represents a further limitation of this research. Asking participants to imagine different teamwork situations together with the required responses to identical items may have biased participants’ answers toward motivation gains rather than losses. In particular, the repeated presentation of teamwork scenarios may have elicited demand characteristics and led to hypothesis guessing among participants. Moreover, we cannot exclude the idea that imagining a single teamwork situation could have led to different answers from the participants than the repeated imagining of several such situations: Our approach may have elicited participants’ tendency to compare the scenarios with each other, and they may have accordingly arrived at relative rather than absolute assessments of intended effort.
Our findings have different implications for motivation management for teams in general and for sports teams in particular. First, to effectively increase their athletes’ overt effort intentions, sports coaches could let team members perform in team situations that are known to systematically induce motivation gains (e.g., social comparison, social compensation, and social indispensability). This recommendation holds in particular if the task is not appealing in itself and it especially holds for single athletes who are normally not members of sports teams (e.g., sprinters or swimmers). Moreover, the recommendation presumably also holds for teachers or managers who assign their pupils or employees to teams for working on routine tasks. Second, the choice of the appropriate task structure appears to be crucial because it needs to fit the supervisor’s goals and the team members’ needs: If the goal is to simultaneously enhance the effort intentions of inferior and superior team members, an additive task structure is the best choice because it allows both team members to experience motivation gains. If the priority is to maximize the effort intention of inferior team members, a conjunctive task structure has been demonstrated to be the most effective strategy. In light of our results, supervisors might keep in mind that their team members may be particularly motivated to make an effort if they can compare their performance with fellow team members and if their individual performance is indispensable for the team. Finally, of course, trainers and team managers should keep in mind that the present study addressed effort intentions instead of actual effort expenditure. Thus, subconscious or autonomous and deliberate processes later in the action regulation process, such as reactions to the mere presence of (e.g., Zajonc, 1965) or the specific behaviors of fellow team members (e.g., Hüffmeier & Hertel, 2011b), might significantly change the overall performance of an individual team member. However, the reported findings may provide initial insights into the motivational processes that exist at the beginning of a team task.
Conclusion
In this study, we tested whether established triggers of motivation gains and losses in groups are already manifested at the level of effort intentions before teamwork actually begins. We found support for this assumption for effort gains that are triggered by social compensation, social comparison, and social indispensability. No support was found for effort losses that might have been triggered by free-riding effects. In addition to providing a better theoretical understanding of the motivating effects of teamwork on different stages of individuals’ action regulation process, the current findings might also provide helpful guidance for team managers by suggesting that established triggers of motivation gains in groups are more easily included in the overt effort intentions of team members than established triggers of motivation losses in groups.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
This research has been supported by a grant from the Deutsche Forschungsgemeinschaft (He 2745/8-4) to the last author.