Suchergebnisse

AbstractMany purposes can be served by using multiple dependent variables in a study, and the application of path analysis to multiple measures can increase both conceptual and statistical power.

Three models of a three-nation arms race are presented: an all-against-all model, a model in which two nations are allied against the third, and a model in which one of the nations has a secret alliance with the other two. In the all-against-all situation, the model predicts an infinitely explosive arms race as each country fears an alliance of the other two. In the two-sided alliance, the model predicts that there will be an mitial increase until the weaker side approaches parity, after which there will be a mutual deescalation. If one nation signs secret treaties with each of the others, then it immediately deescalates. The other two nations then enter a two-sided arms race in which there will be an initial increase until the weaker nation reaches parity followed by mutual deescalation. These models suggest that two-way arms races would not infinitely escalate if it were not for the fear of third party alliances.

chapter 1. Integrating research finding across studies -- chapter 2. Study artifacts and their impact on study outcomes -- chapter 3. Meta-analysis of correlations corrected individually for artifacts -- chapter 4. Meta-analysis of correlations using artifact distributions -- chapter 5. Technical questions in meta-analysis of correlations -- chapter 6. Treatment effects : experimental artifacts and their impact -- chapter 7. Meta-analysis methods for d values -- chapter 8. Technical questions in meta-analysis of d values -- chapter 9. General technical issues in meta-analysis -- chapter 10. Cumulation of findings within studies -- chapter 11. Different methods of meta-analysis and related software -- chapter 12. Locating, evaluating, selecting, and coding studies and presentation of meta-analysis results -- chapter 13. Availability bias, source bias, and publication bias in meta-analysis -- chapter 14. Summary of psychometric meta-analysis

In a recent meta-analysis of attitude-behavior research, the authors of this article found a strong overall attitude-behavior relationship (r = .79) when methodological artifacts are eliminated. The trend in A-B research, however, is to conceive of behavioral intentions (BI) as a mediator between attitudes (A) and behaviors (B). In this study, it is hypothesized that (a) A-BI correlation would be higher than A-B correlation, (b) BI-B correlation would be higher than A-B correlation, (c) A-BI correlation would be higher than BI-B correlation, (d) the variation in BI-B correlations would be greater than that of A-BI, and (e) attitudinal relevance would affect the magnitude of the A-BI correlation. A series of meta-analyses, integrating the findings of 92 A-BI correlations (N = 16,785) and 47 B-BI correlations (N = 10,203) that deal with 19 specified categories and a variety of miscellaneous topics was performed. The results were consistent with all five hypotheses. The theoretical and methodological implications are discussed.

Uncertainty reduction theory argues that certain trait variables influence the process of information gathering. This article examines the validity of scales designed to tap three such trait variables. First- and second-order confirmatory factor analyses were conducted to assess the psychometric adequacy of the Emotional Empathy Scale (Mehrabian & Epstein, 1972), the Self-Consciousness Scales (Fenigstein, Scheier, & Buss, 1975), and the Self-Monitoring Scale (Snyder, 1974). The Emotional Empathy Scale was shown to be construct invalid in that it was multidimensional at both the first- and second-order factor levels. The subscales may be useful, but the composite score is meaningless. The data also showed that the two Self-Consciousness Scales could be reduced to a single second-order factor. This raises some questions about how the scales should be treated, but does not pose any problems for interpreting earlier research. Finally, the Self-Monitoring Scale yielded four primary factors and two second-order factors. Consistent with earlier investigations, this study shows that the scale exhibits very little correspondence to the self-monitoring construct. Its continued use is strongly discouraged. Recommendations for the use and interpretation of research that has employed the scales are offered, and the implications for past and present communication theory are examined.

A reanalysis is given of data presented by Gillmore in which a paired-item design permitted differentiation between reliability and communality. This analysis yielded results "inconsistent" with traditional reliability theory. A new model based on assumptions about memory is presented which fits Gillmore's data. The implications of this theory are indicated for reliability artifacts, measurement, factor analysis, and the measurement of change.