3.2 Relevant measures
Using exploratory factor analysis, we examined the underlying factor structure of the
items in the questionnaire. The analysis was conducted using PASW/SPSS version 20.
We used the maximum likelihood method and oblique rotation to evaluate the underlying factors along with scree plot to determine the number of factors to include in the model. A 0.35 loading was cut-off point for deciding to include items in a scale.
The scree plot and Eigen values support the proposed dimensions (see discussion on
results of factor analysis later in this section).
Commitment to team goals. This variable was defined as the extent to which team
members were committed to a common goal and supported each other in accomplishing that goal. We assessed goal commitment using a four-item measure developed by Pazos et al. (2011). A sample item is “Our team was united in trying to reach its goals for performance”. The coefficient alpha for goal commitment was 0.91 indicating high reliability.
We evaluated Conflict Management as a process mediator of the relationship between the input variable and team outcomes. Conflict management measured the extent to which the team was able to prevent negative conflict and solve emerging conflict. We assessed conflict management using a four-item measure developed by Tekleab et al. (2009) with a five-point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree). A sample item for this scale is “Conflict is dealt with openly on this team.” Exploratory factor analysis resulted on all items loading on a single factor. The reliability coefficient for this scale was a ¼ 0.88.Perceived performance. We adapted a performance measure from Ancona and Caldwell (1992), which included team member’s ratings of their team along five dimensions: efficiency, quality, technical innovation, adherence to requirements, and work excellence. Team members rated each question on a five-point Likert scale in which 1 ¼ poor and 5 ¼ excellent. The five-item scale showed high (a ¼ 0.95)reliability.Satisfaction. We used a measure of satisfaction based on the work by Van der Vegtet al. (1998) and Flynn et al. (2001) to assess the overall team satisfaction with processes and outcomes. The initial measure had 4 items but after factor analysis two items were dropped as they loaded poorly in the factor. The final measure consisted of two questions. The resulting reliability of the measure was a ¼ 0.93. Table I indicates the variable category, name, corresponding items in the survey, and anchors.