Comparison of barriers self-efficacy and perceived behavioral control for explaining physical activity across 1 year among adolescent girls.

The absence of longitudinal data about correlates of physical activity during adolescence has hindered the identification of key mediator variables that can be targeted by interventions. Building on the authors' previous report of a cross-sectional analysis, this study compared barriers self-efficacy and perceived behavioral control as predictors of change in 2 levels of physical activity across a 1-year period among a sample of Black and White adolescent girls (N=1,038). Self-efficacy did not predict change in either moderate or vigorous physical activity. In contrast, perceived behavioral control exhibited a longitudinal, independent relationship with change in vigorous physical activity. The authors concluded that perceived behavioral control is an independent predictor of change in vigorous physical activity levels across a 1-year period among a sample of Black and White adolescent girls and warrants study as a potential mediator variable in interventions designed to increase or maintain physical activity.     

Putting Communication into the Study of Group Memory

In the spirit of viewing human memory as a social process, Pavitt (2003) revisited the idea of group memory—recollection of information that occurs collaboratively in groups. This paper outlines and supports Pavitt's 5 major points about group remembering: (a) groups outperform individuals, (b) communication allows the group to know all knowledge of its members, (c) groups do not perform as well as a rational model predicts, (d) process losses increase as group size increases, and (e) group interaction does not guarantee optimal performance. The research on which these conclusions are based examines group memory in a simple input‐output function. Communication scholars may extend the group memory literature by studying the structure and content of discussion during collective remembering.     

Self‐Construal Scales Lack Validity

Self‐construal is thought to mediate and explain the effects of culture on a wide variety of outcome variables. A meta‐analysis of published cross‐cultural self‐construal research is reported in this article, and the results across studies suggests that the evidence for the predicted cultural differences is weak, inconsistent, or nonexistent. The results of 3 priming experiments (N = 121, N = 99, and N = 361) suggest that (a) priming does not account for the inconsistent results observed in the meta‐analysis, (b) that scores on a self‐construal scale appear to be measuring trait‐like constructs that are not sensitive to priming, and (c) that measures of self‐construals lack convergent validity. The results of several measurement studies (N = 121, 223, 230, 323, 214, 206, 126, 204, 148, 141, and 150) were inconsistent with the a priori two‐factor measurement model in every case. Self‐construal scales were found to be radically multidimensional and highly unstable within and across cultures. These results lead us to conclude that catastrophic validity problems exist in research involving the use of self‐construal scales in cross‐cultural research.