A common practice for counteracting response acquiescence in psychological measures has been to employ both negatively and positively worded items. However, previous research has highlighted that the reliability of measures can be affected by this practice (Spector, 1992). The purpose of the present study was to examine the effect that the presence of negatively worded items has on the internal reliability of the Group Environment Questionnaire (GEQ). Two samples (N = 276) were utilized, and participants were asked to complete the GEQ (original and revised) on separate occasions. Results demonstrated that the revised questionnaire (containing all positively worded items) had significantly higher Cronbach a values for three of the four dimensions of the GEQ. Implications, alternatives, and future directions are discussed.
Eys is with the School of Human Kinetics, Laurentian University; Carron is with the School of Kinesiology, University of Western Ontario; Bray is with the Department of Kinesiology, McMaster University; and Brawley is with the College of Kinesiology, University of Saskatchewan.