Abstract
In a group context (e.g., athletic team), group-level effects may be present in constructs typically measured at the individual level (e.g., athlete satisfaction, leadership behavior). If a group-level effect is present, constructs should be analyzed using the group as the unit of analysis and failure to do so can lead to skewed relationships with other constructs. The purpose of this study is to examine the existence and magnitude of group-level effects when athletes rate athlete satisfaction and leadership behavior. The authors hypothesize: (a) group-level effects emerge when group members rate a shared property of the group, and (b) group-level effects may be present when group members rate an individual-level construct that exist within the context of the group. A total of 212 team athletes (members of 16 interactive athletic teams; mean age 20.1 ± 1.96 years) completed subscales of the Leadership Scale for Sports and the Athlete Satisfaction Questionnaire. Results show large group-level effects for all leadership behavior dimensions and satisfaction dimensions associated with group-level constructs, whereas smaller group-level effects were found for satisfaction dimensions associated with individual-level constructs. The results support the hypotheses that group-level effects can emerge for constructs previously viewed solely as individual-level constructs when measured in a group setting. Recognition of these effects should play a factor in determining the appropriate unit of analysis. Implications for handling groups without a group-level effect, while the majority of groups show an effect, are also discussed.
Get full access to this article
View all access options for this article.
