Psychological safety in teams:
The influence of power, diversity and
identification
Master Thesis MSc Human Resource Management
University of GroningenFaculty of Economics and Business
The influence of power, diversity and identification
TABLE OF CONTENTS
1. INTRODUCTION ... 1
2. THEORY AND HYPOTHESES ... 1
2.1 VARIABLES ... 1
2.2 HYPOTHESES ... 2
3. METHOD ... 5
3.1 PARTICIPANTS AND DESIGN ... 5
3.2 MEASURES ... 6 3.3 DATA ANALYSES ... 7 4. RESULTS ... 7 4.1 DESCRIPTIVE STATISTICS ... 7 4.2 HYPOTHESIS TESTING ... 7 5. DISCUSSION ... 10 5.1 FINDINGS ... 10
5.2 STRONG AND WEAK POINTS ... 11
5.3 THEORETICAL IMPLICATIONS ... 12
5.4 DIRECTIONS FOR FUTURE RESEARCH ... 12
5.5 PRACTICAL IMPLICATIONS ... 14
5.6 CONCLUSIONS ... 14
REFERENCES ... 15
APPENDIX A. ILLUSTRATION OF PARTICIPANT’S TEAM POSITION ... 17
APPENDIX B. EXAMPLE SCENARIO... 19
ABSTRACT
Psychological safety, defined as a perception that the team is safe for interpersonal risk taking (Edmonson, 1999), is an important concept in research on teams. The present
research expands the knowledge of the antecedents of psychological safety by relating it to two popular topics in team research: power and diversity. Through an empirical study among 307 university students, it is shown that individual-level power increases psychological safety. Furthermore, team diversity decreases psychological safety, via reduced team identification. These findings indicate that HR managers and team leaders of highly diverse teams should put effort in increasing identification with the team if they wish to establish a psychologically safe work climate.
1
1. INTRODUCTION
Teams are a crucial part of organizational performance and hence the subject of teams, their functioning and outcomes, has attracted researchers‟ interests. In the mass of team research, the concept of psychological safety is a relatively under-exploited domain. Psychological safety is a perception that the team is safe for interpersonal risk taking (Edmondson, 1999). It has been found to have a beneficial influence on team learning (Edmondson, 1999;
Kostopoulos, Bozionelos & Prastacos, 2009; Tucker, Nembhard & Edmondson, 2007) and job involvement (Brown & Leigh, 1996). Ultimately, psychological safety has a positive effect on team effectiveness (Kostopoulos et al., 2009), team success (Edmondson, Bohmer & Pisano, 2001; Tucker et al., 2007) and team or firm performance (Baer & Frese, 2003; Brown & Leigh, 1996; Edmondson, 1999; Faraj & Yan, 2009).
Whereas there is a rich body of evidence on the advantageous effects of psychological safety, research into the antecedents of psychological safety is less extensive. Kahn (1990)
determined four factors that influence psychological safety: interpersonal relationships, group dynamics, management style, and organizational norms. Edmonson (1999) found that context support and team leader coaching have a positive effect on psychological safety. Finally, Faraj and Yan (2009) concluded that boundary reinforcement – meaning that a team sets its
boundaries by emphasizing team identity – is necessary for the establishment of psychological safety.
The present research aims to expand the knowledge of the antecedents of psychological safety by relating it to two popular topics in team research: power and diversity. In short, the central question of this study is: how do power and diversity, both independently and through their interaction, affect psychological safety? Through an empirical study, I will attempt to show that power increases and diversity decreases the level of one‟s individual psychological safety. In addition, I will examine the interaction effect of power and diversity on psychological safety and a potential mediator of this relationship (as well as of the main effects), identification.
I begin with defining the research variables and justifying the hypotheses. Then, in the method section I will explain the research design, followed by the results section, which will show the outcomes of the investigation. Finally, I discuss the results and propose directions for future research.
2. THEORY AND HYPOTHESES
In this section, I will define the key variables of the present research. Next, I will present the hypotheses.
2.1 Variables
Psychological Safety. Team psychological safety is defined as a “belief that the team is safe
2
Power. Many authors have made an effort to define power and the result is a variety of
definitions that share similarities and differences. The present study defines power as “the ability to control resources” (Galinsky, Gruenfeld, & Magee, 2003, p. 454). In many
definitions of power, control over resources is present (for example Galinsky, Magee, Inesi, & Gruenfeld, 2006; Keltner, Gruenfeld & Anderson, 2003). The more a person is able to control resources, the more he or she is able to influence other people, (Galinsky et al. 2003; Galinsky et al. 2006; Keltner et al. 2003), because those others depend on the resources controlled by the powerful person. This reasoning demonstrates that power is not a characteristic of a person, but a social construct - a property of a relationship between two or more individuals (Emerson, 1962).
Diversity. The term diversity can be specified as the distribution of differences among team
members in regard to a common characteristic (Harrison & Klein, 2007). Diversity is frequently divided into two types. The first type is diversity on observable or visible
characteristics, which are often not task-related, for example gender, age and ethnicity. The second are differences in less visible, task-related characteristics such as education and functional background (Jackson, Joshi, & Erhardt, 2003; Milliken & Martins, 1996). The present study includes both types of diversity.
Identification. The concept of identification stems from social identity theory, which states
that individuals derive part of their personality from the membership of a particular group, the so-called, social identity (Ellemers, Kortekaas, & Ouwerkerk, 1999). Hence, identification with a team means that an individual regards the team as part of his or her self-concept (Griffith & Neale, 2001). Three elements are distinguished that add to an individual‟s social identity. The first is self-categorization (a cognitive element), which is an awareness of one‟s group membership. Group self-esteem is the second (an evaluative element) and refers to the value a person attaches to this membership. The third is group commitment (an affective element), which is defined as emotional involvement with the team (Ellemers et al., 1999). Together these three components determine to what degree a person identifies with the team. The current study includes all three components of identification.
2.2 Hypotheses
Main effect of power. The power-approach theory states that high power triggers
approach-related behavior and low power or a lack of power leads to inhibition-approach-related behavior. High power is associated with positive affect and sensitivity to rewards. On the contrary, low power is related to negative affect and sensitivity to threat and punishment (Keltner et al., 2003). Since individuals in a low power position pay more attention to threats and punishments, they seem to experience an environment that is less psychologically safe than individuals in a high power position. People in a high power position on the other hand, experience the freedom to act without interference or negative social consequences, which indicates that they feel psychologically safe.
3
In general, people experience relationships with persons higher in hierarchy as more threatening than relations with persons at the same level. People feel more psychologically safe in relationships with peers (Kahn, 1990). Drawing on this reasoning, I assume that the same applies for relationships with people lower in hierarchy than themselves: these relationships are less threatening than are relationships with people higher in power. Thus, people are expected to feel more comfortable in relationships with peers and with people lower in hierarchy than they are in relationships with people in higher hierarchical positions. This reasoning indicates that in a relationship between both high- and low-power individuals, high-power individuals feel more psychologically safe than low-power individuals, because they do not experience the threat of a higher-power other in the environment.
In conclusion, I propose that power positively predicts psychological safety.
Hypothesis 1. Power is positively related to psychological safety.
Main effect of diversity. Diversity is a double-edged sword. On the one hand, it has positive
cognitive consequences, such as improved decision making, problem solving, innovation, and eventually performance. On the other hand, diversity is associated with negative affective reactions, such as, decreased commitment, cohesion, and satisfaction (Jackson et al., 2003; Milliken & Martins, 1996). The latter phenomenon can be explained from the similarity-attraction perspective: in general, people have a preference for people who are similar to themselves (Byrne, 1997). Similarity increases interpersonal liking, which in turn, is assumed to increase psychological safety. Thus, in a diverse team, psychological safety is likely to be lower than in a similar team.
Furthermore, diversity has a negative effect on team processes such as
communication, conflict and cooperation; however, this pattern is complex and inconclusive (Jackson et al., 2003; Milliken & Martins, 1996). Diverse teams are characterized by a work climate of increased conflict and decreased cooperation and cohesion. In such a climate, I would expect psychological safety to be lower than in an environment where less conflict occurs. This is because this kind of work climate will discourage team members to express themselves due to the threat of negative consequences for such behavior.
Therefore, I hypothesize that diversity is negatively associated to psychological safety, because diversity increases conflict and decreases cohesion and cooperation, and in such a climate I expect people to feel less psychologically safe.
Hypothesis 2. Diversity is negatively related to psychological safety.
Interaction of power and diversity. Previous research by Rink (2009) examined interactive
effects of power and diversity within teams and revealed how power differences influenced reactions to diversity within teams. First, she found a main effect of power on turnover intentions: the powerless were more likely to turnover than were the powerful. Furthermore, diversity was strongly related to turnover for individuals in a high power position but not for individuals in a low power position. Thus, powerful people in a diverse team were more likely to turnover than powerful people in a non-diverse team. Thus, individuals in a low power position appeared to be less affected by diversity.
4
members better and might discover shared interests or characteristics. High power individuals do not engage in such individuation, and thus, never realize the shared characteristics of a diverse team. Such shared characteristics will overshadow the apparent differences, such as gender, race or age. High-power individuals pay less attention and use more stereotypes, thereby reinforcing differences. Besides, power increases prejudices (Guinote, Willis & Martellotta, 2010), so high-power individuals in a diverse team are expected to be guided by prejudices, eliciting negative affective reactions.
Therefore, I propose that power and diversity will interact in a similar way in affecting psychological safety.
Hypothesis 3: Power will interact with diversity in predicting psychological safety. The nature of this relationship will be such that the relationship between diversity and psychological safety is negative when power is high, but becomes non-significant when power is low.
Identification as a mediator. The same study by Rink (2009) showed a relationship between
power and diversity on the one hand, and identification at the other. People in diverse teams identified less with the team than did people in a group with people who were similar to them. Likewise, the more power a person had, the more he or she identified with the team. The research also found an interaction effect of power and diversity on identification: for high power individuals the relationship between diversity and identification was strongly negative, whereas diversity did not affect identification for low power individuals.
A central point of social identity theory is that in many situations people tend to think of themselves and others in terms of members of a particular group, rather than as
independent individuals. The more people identify with the team, the more likely they are to adopt group norms as guidelines for their own behavior (Ellemers et al., 1999). Identification has many positive effects for the team: increased cohesion, satisfaction, cooperation,
commitment and trust (Griffith & Neale, 2001). All these factors are ingredients of a
psychologically safe environment, and thus it can be argued that identification contributes to higher levels of psychological safety.
In sum, power and diversity affect psychological safety. I propose that these effects occur through their impact on identification with the team, such that powerful people and people in a team with similar others identify more with the team and as a result experience higher levels of psychological safety.
Hypothesis 4: the interaction effect of power and diversity on psychological safety is mediated by identification.
5
FIGURE 1 Research Model
3. METHOD 3.1 Participants and Design
Participants (Mage = 21.55, SD = 1.99, 56% female) were 307 students from a university in
The Netherlands. They participated in this study as part of a classroom exercise. The
participants came from two different classes and completed the surveys on different days. One class received the packet in paper-and-pencil form and completed it during the lecture. All 124 students who were present during the lecture completed the survey. During this procedure, I had concerns over whether participants were completing the survey
independently or if they were observing each others‟ work, leading to potential insights in the power and diversity manipulation. This observation of others‟ work might have compromised the research. Therefore, I decided to replicate the research using a virtual platform in another class. This class received an email with an invitation to fill in the survey online outside of class. Of the 428 students who were invited, 183 students completed the survey (43 percent response rate). Thus, I combined samples and performed the final data analyses on N = 307 participants.1
Using a 2 (diversity: similar, different) x 2 (power: high, low) design, I randomly assigned participants to one of the six conditions. I then told participants that they may be assigned to work groups for an upcoming class project. Specifically, I explained:
We may be forming working groups for an upcoming project about organizations. These groups will be formed based on the responses that you and your classmates gave to the personality and demographics (e.g., age, race, gender) questions that you answered after the first class lecture. We are trying to get people’s feelings about the
1
6 potential groups that they might be put in. Therefore, please read the scenario below and answer the following questions.
Thus, participants believed that their group assignments would be based on individual differences (or similarities) between class members. This introduction was followed by a scenario, in which I manipulated the focal variables, power and diversity.2 I also included a figure that illustrated the participant‟s position in the potential team (see Appendix A). The scenario provided participants with information on their power position and the level of diversity within the team. For example in the high diversity condition, I told participants that they were highly different from the other team members. In contrast, in the low diversity condition, I told them that they are highly similar to the other team members and that all team members “have a lot in common.” I manipulated power by indicating if they did or did not have the ability to control resources within the team. In the high power condition, I told participants that they had control over resources and more “say” in important issues. In contrast, in the low power position I told participants that other team members controlled resources and had more “say” in important issues. Appendix B contains an example scenario. After reading the introduction and about their potential team assignment, participants
completed several dependent measures, including a measure of psychological safety (Edmondson, 1999) and a demographics questionnaire.
3.2 Measures
Psychological safety. To assess psychological safety I used a seven-item scale adapted from
Edmondson (1999). Items included, “It will be safe to take a risk on this team” and “If you make a mistake on this team, it will be held against you” (reverse coded). Three items were reverse-coded. Participants answered the items on a 7-point Likert scale, ranging from 1 (not
at all) to 7 (very much). I averaged these items to compute a total psychological safety score.
The internal reliability of the scale was good (Cronbach’s alpha = 0.73).
Identification. To measure identification, I used a nine-item scale adapted from Ellemers,
Kortekaas, and Ouwerkerk (1999). Items included, “You will be proud to belong to this team” and “You will fit well within your team”. Participants answered the items on a 7-point Likert scale ranging from 1 (not at all) to 7 (very much). I calculated the average of the nine items to compute a total identification score. The scale‟s internal reliability was very good
(Cronbach’s alpha = 0.94).
Manipulation check. To check whether the power manipulation had been successful, I
included three items that measure sense of power adapted from Anderson and Berdahl (2002). These items are: “You will be dominant in this team,” “You will be in control in this team,” and “You will lead discussions in this team.” I accompanied the items with a 7-point Likert scale ranging from 1 (not at all) to 7 (very much). I averaged the items to calculate a sense of power score (Chronbach’s alpha = 0.91).
Control variable. I included gender as a control variable.
2
7
3.3 Data Analyses
I tested my hypotheses using hierarchical regression. I transformed power and diversity into dummy variables (dichotomously coded as -1 and 1). To compute an interaction term, I multiplied the power dummy variable with the diversity dummy variable. To increase ease of interpretation, I mean-centered the identification scale by subtracting 4 from each individual‟s total scale score.
4. RESULTS 4.1 Descriptive Statistics
Table 1 presents means, standard deviations and correlations of all the variables included in the current research. Diversity correlates negatively with psychological safety and
identification. Power correlates positively with psychological safety and identification. Finally, there is a positive correlation between psychological safety and identification. Only diversity and power do not correlate, which is expected since these two variables were experimentally manipulated and randomly-assigned.
TABLE 1
Means, standard deviations and intercorrelations of all variables
Variable Mean SD 1 2 3 4 1. Psychological Safety 4.40 0.82 - 2. Power - - 0.37*** - 3. Diversity - - -0.29*** 0.02 - 4. Identification 4.42 1.18 0.58*** 0.47*** -0.45*** - *** p < 0.001, N = 203 - 3073 4.2 Hypothesis testing
Manipulation check. To assess whether the manipulation of power was successful, I
regressed one‟s sense of power (Anderson & Berdahl, 2002) on to the power manipulation. My power manipulation significantly predicted participants‟ sense of power (b = 0.60, t = 6.65, p < 0.001). To rule out other variables, I also checked whether gender and diversity were significant predictors of sense of power. These variables were non-significant (p = 0.69 and p =0.25, respectively). Thus, the power manipulation can be considered successful.
Control variable. In order to rule out any of influence of the control variable, gender, I
entered gender in the first step of the regression model. Since this variable was non-significant (b = 0.14, t = 1.15, p = 0.25), I removed it from further analyses.
Main effects. The main effects concern the relationship between power and diversity on the
one hand and psychological safety on the other (Hypothesis 1 and 2). To assess the main effects, I ran a regression analysis with psychological safety as the dependent variable. I entered power and diversity as independent variables. Results show that power is positively related to psychological safety (b = 0.31, t = 5.78, p < 0.001) and that diversity is negatively
8
related to psychological safety (b = -0.18, t = -3.33, p < 0.001). In both cases the directions of the relationships are consistent with the hypotheses, such that power increases and diversity decreases psychological safety. Thus, both Hypothesis 1 and 2 are confirmed.
Interaction. The third hypothesis concerned the interaction effect of power and diversity on
psychological safety. To investigate this hypothesis, I entered power and diversity in the first step and the interaction term in the second step of the regression analysis. Results were non-significant (p = 0.75). Thus, I rejected Hypothesis 3, which stated that power and diversity interact to affect psychological safety.
Table 2 summarizes the results of this regression analysis.
TABLE 2
Tests of main effects and interaction
Step 1 (main effects) Step 2 (interaction effect)
Variable b b Constant 4.33*** 4.33*** Power 0.31*** 0.31*** Diversity -0.18*** -0.18*** Power*Diversity - 0.02 R2 0.18 0.18 R2 - 0.00 F 21.98 14.62 ***p < 0.001
Mediation. Since I found no relationship between the interaction term and psychological
safety, Hypothesis 4 (the interaction effect of power and diversity on psychological safety is
mediated by identification) was ruled out. However, because I found a main effect of both
power and diversity, I did a post hoc examination to see if identification mediated these two main effects on psychological safety.
First, I examined whether identification is a mediator in the power to psychological safety relationship. I already showed (see Table 2) that power is positively related to psychological safety (b = 0.31, t = 5.59, p < 0.001). To establish a mediation effect, three further conditions must be met: (1) power should predict identification, (2) identification should predict
psychological safety and (3) the relationship between power and psychological safety should become non-significant when identification is entered in the model.
I ran a series of regression analyses to examine this mediation effect. The results of these analyses are presented in Table 3. First, results show that power is positively related to identification (b = 0.56, t = 7.52, p < 0.001). Second, identification is positively related to psychological safety (b = 0.40, t = 12.28, p < 0.001). Finally, power remained a significant predictor of psychological safety (p = 0.05) when identification was entered into the model, concluding that, identification does not mediate the relationship between power and
9
TABLE 3
Tests of identification as mediator between power and psychological safety
Conditions to establish mediation* Independent variable b t P R2 F
Main effect power on psychological safety Power 0.31 5.59 < 0.001 0.14 31.27
1. Does power significantly predict
identification? Power 0.56 7.52 < 0.001 0.22 56.57
2. Is identification significantly related to
psychological safety? Identification 0.40 12.28 < 0.001 0.33 150.89
3. Does the effect of power on psychological safety become non-significant when identification enters the model?
Power 0.11 2.01 0.045 0.33 49.06
Identification 0.35 7.62 < 0.001
*Dependent variables are in italics.
In a similar way, I examined whether identification mediates the relationship between diversity and psychological safety. Table 4 shows the results of these analyses. As shown above (see Table 2), the relationship between diversity and psychological safety is negative (b = -0.24, t = -5.28, p < 0.001). First, diversity is negatively related to identification (b = -0.52, t = -8.68, p < 0.001). Secondly, as shown in Table 3, identification is positively related to psychological safety (b = 0.40, t = 12.28, p < 0.001). Finally, the negative relationship between diversity and psychological safety becomes non-significant when identification is entered into the model (p = 0.45), and identification remains significant (p < 0.001). Thus, identification mediates the relationship between diversity and psychological safety.
TABLE 4
Tests of identification as mediator between diversity and psychological safety
Conditions to establish mediation* Independent variable b t p R2 F
Main effect of diversity on psychological
safety Diversity -0.24 -5.28 < 0.001 0.08 27.83
1. Does diversity significantly predict
identification? Diversity -0.52 -8.68 < 0.001 0.20 75.36
2. Is identification significantly related to
psychological safety? Identification 0.40 12.28 < 0.001 0.33 150.89
3. Does the effect of diversity on psychological safety become non-significant when identification enters the model?
Diversity -0.03 -0.76 0.449 0.33 75.62
10
5. DISCUSSION 5.1 Findings
The present study aimed to expand the knowledge of the antecedents of psychological safety through an empirical study. This goal was achieved by demonstrating that both power and diversity affect psychological safety.
First, power was positively related to psychological safety (Hypothesis 1). Team members in a high power position felt more psychologically safe than did those is a low power position. This finding is in line with the power-approach theory (Keltner et al., 2003), which poses that people in a high power position experience the freedom to act without the fear of negative social consequences.
Second, I found that diversity is negatively related to psychological safety (Hypothesis 2). The more diverse a team is, the less psychologically safe the team members feel.
Lau and Murnighan (2005) investigated the relationship between diversity and psychological safety. In their study, Lau and Murnighan (2005) conceptualized diversity in the form of faultlines. A faultline is an imaginary dividing line within a team. If several faultlines such as age and gender converge, for example a team with young females and middle-aged men, then the potential of subgroup formation is largest (for a full discussion of the faultline-model see Lau & Murnighan, 1998). In contradiction with their hypothesis, Lau and Murnighan (2005) found that groups with strong faultlines reported higher psychological safety than groups with weak faultlines. They explained this finding as the result of
generalizations: people might have attributed positive social effects within the subgroup to the entire group.
The results of Lau and Murnighan‟s (2005) study seem to contract the results of the present study. Lau and Murnighan‟s (2005) found that groups with strong faultlines have high levels of psychological safety; the current study found that highly diverse groups have low levels psychological safety. However, it is very difficult to compare the results of both studies, since they use a completely divergent conceptualization of diversity. The faultline-model concentrates on the potential of subgroup formation within teams, whereas the current research measures diversity based on the degree to which team members differ. In other words, the faultline-model measures diversity in terms of weak or strong faultlines, the present study measured diversity in terms of similar or diverse (that is, little or many
differences among team members). The contrast between weak and strong faultlines does not equate with low and high diversity. This is because similar teams generally have weak faultlines, but highly diverse teams can have either strong or weak faultlines, depending on whether the dividing lines within the team converge or not. For example, when all HR specialists within a team are young females, and all Marketing specialists are older males, than three potential dividing lines (age, gender and function) converge. In this case, faultlines are strong, and the potential of subgroup formation is large. According to the Lau and
Murnighan (2005) research, psychological safety is high in a team with strong faultlines, which might be due to the fact that team members feel safe within their subgroup and attribute these feelings to the entire group. However, it is also possible that within an organization, both HR specialists and Marketing specialists are comprised of both young and old males and females with a diversity of backgrounds. Thus, in this situation, faultlines are weak and are likely to result in lower psychological safety.
11
commitment, cooperation and cohesion and increased conflict (Jackson et al., 2003; Milliken & Martins, 1996). A work climate that is characterized by conflict and decreased cohesion and cooperation will discourage team members to take the risk of seeking feedback or expressing concerns, because they fear negative consequences of such behavior. The third hypothesis stated that diversity and power would interact in predicting
psychological safety. Contrary to what I hypothesized, power and diversity did not interact. This finding is unexpected, especially since previous research (Rink, 2009) did find an interaction between power and diversity.
One difference between the research of Rink (2009) and the current study is the scenario. Rink (2009) asked students to imagine they were working in a management team of a consultancy company. Whereas in the current study, participants believed they would be put in the proposed teams, and were asked to imagine that they were working in a team “to
complete consulting projects about organizational issues and give advice to local
organizations”. The difference between the two is thus that the research of Rink (2009) is entirely hypothetical, whereas the current study possessed more realism.
The fact that the two studies produced inconsistent results might indicate that people are poor predictors of how they would respond when actually facing the possibility of being part of a group. To answer the question of whether power does influence people‟s responses to diversity field or lab research is needed. I will return to this topic in paragraph 5.4
„directions for future research‟.
Finally, I examined identification as a possible mediator (Hypothesis 4). Since no interaction effect was found, I performed post hoc analyses to examine whether identification mediated the main effects of power and diversity on psychological safety.
Results show that identification does not mediate the power to psychological safety relationship. This null finding raises the question of what variables do mediate the
relationship between power and psychological safety. One possible mediator is self-esteem. Wojciszke and Struzynska-Kujalowicz (2007) demonstrated that power increases self-esteem, while acknowledging that power and self-esteem are reciprocal influences. I would expect that higher self-esteem would in turn increase the perceived level of psychological safety, because if people have more positive attitudes towards themselves, I expect them to feel more free to speak up without being afraid of negative consequences.
Whereas identification was not a mediator in the relationship between power and psychological safety, identification did mediate the effect of diversity on psychological safety. Thus, when diversity increases, people identify less with the team and as a result they feel less psychologically safe. This finding is in line with what I hypothesized based on existing
theory.
5.2 Strong and weak points
Some limitations of the research need to be acknowledged. First, because of the low
variability in the personal homogeneity of the sample (all participants are university students), one should be cautious in generalizing the results to other samples. For example, results might be different if I were to examine individuals with significant business experience or
individuals with experience working with diverse others. Also, results might be different among individuals with actual leadership experience, as opposed to students who generally have little experience with being in a high power position.
12
differently than task-related diversity, because the first type is more visible and observable than the latter. In general, demographic diversity leads to negative affective reactions such as conflict and commitment, whereas task-related diversity leads to positive cognitive effects, such as innovation and the presence of a range of different perspectives (Milliken & Martins, 1996). I would expect both types of diversity to affect psychological safety, but the results of demographic diversity might be stronger than for task-related diversity. This is because demographic diversity elicits negative affective reactions among team members that are very closely related to psychological safety, such as increased conflict and decreased commitment and cohesion.
Finally, since power and diversity were dichotomous variables, no conclusions can be drawn about how varying degrees of power and diversity affect psychological safety. Power and diversity can be continuous constructs. In the real world, power is often not an all- or nothing-variable, but can range from absolute powerlessness to absolute power. For example, in a management team, the CEO has the highest power position; the CFO has less power than the CEO but still more power than the marketing manager; and the marketing manager in his turn has less power than the CEO and CFO but more power than the HR manager. The same holds for diversity. Some teams are highly diverse meaning that team members differ on many characteristics. An example is a medical surgery team, with people of different disciplines, functional background, age, gender, race, etc. Other teams share many
similarities, for example a team of young, female nurses. In between is a whole range of other teams, each with a different degree of diversity. Again, it is not simply a dichotomy of diverse or similar.
On the other hand, some strong points of the present study also deserve to be acknowledged. The first is the large sample size, which strengthens the power of the research. Furthermore, I used two different methods of data collection (i.e., virtual and paper-and-pencil), thereby reducing any variance due to the method of data collection. Another strong point is the level of realism in the study; students actually thought they would be placed in a team. Lastly, the power and diversity manipulations were both verbally and graphically illustrated, making it very clear for the participants to imagine their position in the team.
5.3 Theoretical implications
Existing research on psychological safety has primarily focused on the advantageous effects of psychological safety such as improved learning and team effectiveness or success (for example Edmondson, 1999; Kostopoulos et al. 2009). Less attention has been paid to factors affecting psychological safety in teams. The current study has extended theory on the
antecedents of psychological safety by demonstrating that psychological safety is affected by power and diversity, in latter case through identification.
I showed that an individual level variable (power) influences psychological safety. This implies that team members from the same team but holding different power positions within that team, might hold different perceptions of the level of psychological safety within the team. This finding contradicts the common assumption that all members of a team share a similar perception of the level of psychological safety (Edmondson, 1999). The results of the present study suggest that psychological safety can be treated as an individual-level variable within a team context. However, because no actual teams were formed in the current study, this is a question for future research.
5.4 Directions for future research
13
factors, besides power, might be uncovered that cause team members to have different perceptions of the level of psychological safety within the team.
I propose that role conflict and role ambiguity could affect psychological safety and cause team members to experience different levels of team psychological safety. Role conflict occurs when an individual receives conflicting requests and role ambiguity refers to an
individual not having enough information to carry out his or her job; both lead to reduced commitment and satisfaction (Anton, 2009). I expect people who experience the deleterious role-related attributes to feel less secure in expressing their thoughts and concerns, because they are already in a problematic situation and do not want to take the risk of aggravating their position.
Second, as mentioned above, the interaction between power and diversity might be examined in a field or lab study. Since two experimental scenario studies provide mixed results on whether power influences responses to diversity, the next step would be to study existing teams in organizations or do a lab study in which people are put in teams and power and diversity are manipulated.
In a field study, objective measures can be combined with subjective measures. For example, demographics and personnel records can be used to measure the level of diversity (objective data), and team members can rate the level of psychological safety within the team (subjective data). Besides, observations of participants can provide additional information. For example, in team meetings one can observe behaviors that are a consequence of psychological safety such as asking questions, asking for feedback, admitting mistakes, voicing dissent and expressing concerns. These observations can be combined with the subjective measurement of psychological safety. On the one hand, I expect results of such a research design to be less strong since existing teams have already developed a group culture and group norms, which might blur the effects of diversity and power. If so, this finding would imply that teams with a longer history together might show decreased negative effects of diversity. On the other hand, the combination of objective, observational and subjective data might strengthen the results of the investigation compared to the current study. Another advantage of such a field study is that people do not need to predict their own response to an imaginary situation, as was the case in the present study.
In a lab study, people might be put in teams with differing levels of diversity and be assigned to different levels of power. I expect results of such an experience to be strong, since people are placed in teams with people they have never met before. Accordingly, first
impressions would play a central role and since power and diversity are very visible factors, I expect that power and diversity elicit strong responses.
Another avenue for additional research is to examine self-esteem as a mediator in the relationship between power and psychological safety. As discussed above, I predict that people with high self-esteem perceive higher levels of psychological safety.
Lastly, future research might dive into the question how different types of diversity, observable or underlying task-related differences, affect psychological safety. As argued above, I would predict that demographic diversity would have a stronger effect on
14
5.5 Practical implications
Psychological safety has many advantageous effects for teams, such as improved learning and team performance (Baer & Frese, 2003; Brown & Leigh, 1996; Edmondson, 1999; Faraj & Yan, 2009; Kostopoulos et al., 2009; Tucker et al., 2007). High levels of psychological safety are desirable, especially in teams where innovation and learning are important. The results of the present study show that HR managers and team leaders of highly diverse teams might consider taking measures to increase the level of psychological safety. I found that identification mediates the diversity-psychological safety relationship, so team building activities that increase the level of identification are good means to increase the level of psychological safety within the team, for example a team excursion, an outdoor team exercise, or a workshop on team communications, brainstorming or creativity. Results also indicate that special attention might be paid to people in a low power position, who have lower levels of psychological safety. If team leaders want to increase the level of psychological safety, then they can allocate more responsibilities and influence to low-power individuals.
5.6 Conclusions
15
REFERENCES
Anderson, C. & Berdahl, J. L. 2002. The experience of power: examining the effects of power on approach and inhibition tendencies. Journal of Personality & Social Psychology, 83 (6): 1362-1377.
Anton, C. 2009. The impact of role stress on workers' behaviour through job satisfaction and organizational commitment. International Journal of Psychology, 44 (3): 187-194. 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 (1): 45-68.
Brown, S.P. & Leigh, T.W. 1996. A new look at psychological climate and its relationship to job involvement, effort, and performance. Journal of Applied Psychology, 81 (4): 358-368.
Byrne, D. 1997. An overview (and underview) of research and theory within the attraction paradigm. Journal of Social and Personal Relationships, 14 (3): 417-431.
Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams.
Administrative Science Quarterly, 44 (2): 350–383.
Edmondson, A. C., Bohmer, R. M. & Pisano, G. P. 2001. Disrupted routines: team learning and new technology implementation in hospitals. Administrative Science Quarterly, 46 (4): 685–716.
Ellemers, N., Kortekaas, P. & Ouwerkerk, J. W. 1999. Self-Categorisation, commitment to the Group and Group self-esteem as related but distinct aspects of social identity.
European Journal of Social Psychology, 29 (2-3): 371-389.
Emerson, R. M. 1962. Power dependence relations. American Sociological Review, 27 (4): 31-41.
Faraj, S. & Yan, A. 2009. Boundary work in knowledge teams. Journal of Applied
Psychology, 94 (3): 604-617.
Fiske, S. T. 1993. Controlling other people: The impact of power on stereotyping. American
Psychologist, 48 (6): 621-628.
Galinsky, A.D., Gruenfeld, D.H & Magee, J.C. 2003. From power to action. Journal of
Personality and Social Psychology, 85 (3): 453-466.
Galinsky, A. D., Magee, J. C., Inesi, M. E. & Gruenfeld, D. H. 2006. Power and perspectives not taken. Psychological Science, 17 (12): 1068-1074.
Griffith, T. L. & Neale, M. A. 2001. Information processing in traditional, hybrid, and virtual teams: From nascent knowledge to transactive memory. Research in Organizational
16
Guinote, A., Willis, G.B. & Martellotta, C. 2010. Social power increases implicit prejudice.
Journal of Experimental Social Psychology, 46 (2): 299-307.
Harrison, D. A., & Klein, K. J. 2007. What‟s the difference? Diversity constructs as
separation, variety, or disparity in organizations. Academy of Management Review, 32 (4): 1199–1228.
Jackson, S.E., Joshi, A. & Erhardt, N.L. 2003. Recent research on team and organizational diversity: SWOT analysis and implications. Journal of Management, 29 (6): 801-830. Kahn, W. A. (1990). Psychological conditions of personal engagement and disengagement at
work. Academy of Management Journal, 33 (4): 692–724.
Keltner, D., Gruenfeld, D.H, & Anderson, C. 2003. Power, approach, and inhibition.
Psychological Review, 110 (2): 265–284.
Kostopoulos, K. C., Bozionelos, N. & Prastacos, G. P. 2009. Team learning activities and team effectiveness: the role of psychological safety and conflict. Academy of
Management Proceedings, 1-6.
Lau, D. C. & Murnighan, J. K. 1998. Demographic diversity and faultlines: The
compositional dynamics of organizational groups. Academy of Management Review, 23 (2): 325–340.
Lau D.C. & Murnighan J. K. 2005. Interactions within groups and subgroups: The effects of demographic faultlines. Academy of Management Journal, 48 (4): 645-659.
Milliken, F., & Martins, L. 1996. Searching for common threads: Understanding the multiple effects of diversity in organizational groups. Academy of Management Review, 21 (2): 402–433.
Tucker, A. L., Nembhard, I. M., & Edmondson, A. C. 2007. Implementing new practices: An empirical study of organizational learning in hospital intensive care units.
Management Science, 53 (6): 894–907.
Rink, F. 2009. Power and diversity in teams. Findings presented at a lecture on power in teams, December 2009, University of Groningen.
Wojciszke, B. and Struzynska-Kujalowicz, A. 2007. Power influences self-esteem. Social
17
APPENDIX A. ILLUSTRATION OF PARTICIPANT’S TEAM POSITION
FIGURE A1
High power, high diversity
FIGURE A2 High power, low diversity
18
FIGURE A4 Low power, low diversity
FIGURE A5 High diversity
19
APPENDIX B. EXAMPLE SCENARIO High power, high diversity
We may be forming working groups for an upcoming project about organizations. These groups will be formed based on the responses that you and your classmates gave to the personality and demographics (e.g., age, race, gender) questions that you answered after the first class lecture. We are trying to get people‟s feelings about the potential groups that they might be put in. Therefore, please read the scenario below and answer the following
questions.
Read the following instructions before answering the questions that follow.
Imagine – as best as you can – that you work in a team to complete consulting projects about organizational issues and give advice to local organizations. You work closely together and meet on a regular basis. You need each other in order to successfully complete team projects. Within your team, you control the resources and have more say in important issues than the other members do.
You are also highly different from the other team members. They themselves differ from each other, as well, meaning that all of you do not have a lot in common.
The figure depicted below represents your place in the team.
