Diversity in Team Composition and Performance and Creativity

1
Diversity in Team Composition and Performance and
Creativity

• Jill A. Marsteller, PhD MPP
• Assistant Professor,
• Johns Hopkins Bloomberg School of Public Health
• 2007 Health Workforce Interest Group Meeting,
  Orlando, FL Saturday, June 2, 2007

2
Background

• Interdisciplinary teams are used often in
  healthcare
• Team members vary in occupation, training,
  experience, and demographic background
• Too many cooks in the kitchen vs. Two heads
  are better than one

3
Definitions

• Diversitytwo types
• Visible age, gender, (race/ ethnicity)
• Invisible occupation, tenure, education
• Chronic care QI team--three or more nurses, PCPs,
  specialists, administrators, etc. who work to
  improve management of asthma, diabetes,
  depression, or congestive heart failure

4
Performance Vs. Creativity

• Performancereaching a set target
• Creativityproduction of new ideas, testing and
  modifying procedures

5
How diversity works

• Diversity--gtSocial Categorization,
  Identification, Impersonal Attraction--gtBias,
  Negative Conflict, Low Cohesion--gtNegative
  Outcomes
• Diversity--gtAiring Multiple Perspectives--gt
  Positive Conflict--gtImproved Decision-making--gtPos
  itive Outcomes
• Literature--Jury is out on diversity

6
Research Questions

• Does occupational diversity have different
  associations with performance as compared to
  creativity?
• Are there mediators of the effects of
  occupational diversity?

7
Fig. 1--Proposed Conceptual Model for the
Relation of Nonracial Diversity to Performance
(unmeasured process)
(-)
Diversity
Team Performance
-Visible diversity -Invisible diversity
-Interaction variables
()
(-)
Unifying Phenomena
Social categorization, identification, group
attraction conflict coordination
problems
-Goal attainment -Patient outcomes -Met
expectations
(Mediator)
Based on Milliken, Bartel and Kurtzberg (2001),
Tsui and Gutek (1999) and Jackson, May and
Whitney (1995)
8
Fig. 2--Proposed Conceptual Model for the
Relation of Nonracial Diversity to Creativity and
Productivity
(unmeasured process)
(-)
Diversity
Team Productivity and Creativity
-Visible diversity -Invisible diversity
-Interaction variables
Phenomena Stressing the Individual
()
Social categorization, identification, group
attraction conflict coordination
problems
(-)
-Number of actions taken -Novelty of actions
(Mediator)
Based on Milliken, Bartel and Kurtzberg (2001),
Tsui and Gutek (1999) and Jackson, May and
Whitney (1995)
9
Study Design and Population

• Secondary data analysis
• OLS and Logistic regression analysis
• 40 teams from hospitals, physician groups,
  clinics, health plans, or health systems
  participating in 3 Improving Chronic Illness Care
  Collaboratives (evaluation-ICICE)
• Surveyed members of multi-disciplinary teams
• Size 1 to 14 members
• As many as 12 occupational categories

10
Data

• Level of analysis--the team
• Some measures constructed for the team from
  individual-level data
• means
• CVs (Allison 1978)
• Heterogeneity index, H 1-S pi2 (Blau 1977)
• Some measures collected at team level

11
Variables
12
Analytic Methods

• Mediated regression analysis using repeated OLS
  (Baron and Kenny 1986)
• Examines, in additive steps
• controls
• main effects of variables of interest
• moderating effects
• mediators (to see if relationships change)
• Conditions for pure mediation
• Independent variables affect the mediator
• Independent variables affect the DV
• Mediator affects the DV when included with
  independents, effect of independent variables is
  reduced
• An exploratory analysis

13
Results--Simple Statistics
14
Results--T1 Self-Assessed Team Performance
15
Results--Number of Changes in Care Practices
16
Results--Innovativeness
17
Results--T2 Self-Assessed Team Performance
18
What have we learned?

• Occupational diversity had a positive effect on
  T1 self-assessed performance
• But no effect on T2 self-assessment
• And no effect on the 2 creative outcomes
• So its good for starters and makes no difference
  after a while, when the positive start-up effect
  is accounted for

19
Limitations

• Small N means that sophisticated analysis of all
  effects is impossible. Models are stretched to
  their limits.
• Multicollinearity likely mutes some potentially
  significant effects of diversity measures.

20
Conclusions

• Diversity can be one possible barrier to team
  quality improvement
• But sometimes is a boon
• Insight into how to maximize benefits of teamwork
• Keep teams together longer (change in diversitys
  effects over time)
• Seek to manage negative effects of diversity with
  cohesion-building efforts
• Encourage individual participation and grant
  autonomy