The "sleeper effect" has been defined as a delayed increase in the persuasive impact of a communication from a source low in credibility. This effect failed to appear in the present series of seven replications (total N = 656) of a sleeper effect experiment. However, the effectiveness of communications attributed to a source high in credibility did significantly decrease with passage of time when subjects were unprepared to counterargue with the communication. On the basis of these results and a retrospective review of related studies, it was concluded that the latter effect, by itself, could account satisfactorily for previous findings that source credibility effects dissipate or disappear with passage of time.As many undergraduate psychology students have by now learned, the "sleeper effect" is an established phenomenon in the attitude change literature. This effect is said to occur when a communication from a source of low credibility has a greater persuasive impact after some time delay than on original exposure. The communication works, that is, while the audience "sleeps" on it. Descriptions of the sleeper effect may be found in a variety of social psychology texts. Examination of the authors' bookshelves indicated that approximately 75% of recent survey texts make explicit reference to the sleeper effect as an established phenomenon.The negative results of seven replications of a sleeper effect experiment to be reported here p r o m~t e d the authors to reexamine the empirical literature on this phenomenon. Remarkably, this retrospective review led to the conclusion that the sleeper effect has never had a truly satisfactory emnirical basis. The origin21 report of a sleeper effect by Hovland, Lumsdaine, and Sheffield (1949) was based on only a subset of the opinion items they 'This research was supported in part by Grants GS-1601 and GS-3050 from the National Science Foundation and by Grant 1-R01-MH23527-01 from the United States Public Health Service. The authors are grateful to Thomas D. Cook and Walter Weiss for critically reading an earlier draft of this article; to David D. S. Poor for assistance in data analysis; and to James A. Chase, Merry Edison, Shiloh Glass, Peter Jay, Kurt Kahler, and Richard Pr6e for assistance in conducting the studies.2 Requests for reprints should be sent to Anthony G. Greenwald, Department of Psychology, Ohio State University, 404C West 17th Avenue, Columbus, Ohio 43210.used: these authors did not report any criteria that could be used to sort items, a priori, into ones that would show a sleeper effect and ones that would not. I n subsequent reports of a sleeper effect, a statistically significant increase in opinion for a group receiving a communication from a low-credibility source has never been r e p~r t e d !~ Typically, statistical significance has been achieved in sleeper effect studies bv contrasting the nonsignificant (or absent) delayed i m~a c t of a communication from a low-credibility source with a substantial decrease in impact over time in a group receiving ...

This paper describes two related studies. Study 1 is a literature review of existing adult partner domestic violence assessment scales. Results of the review revealed that the scales varied on the available amount of empirical evidence for validity and reliability. More importantly, results showed that the content of the scales focused most heavily on the physical abuse aspects of domestic violence. Study 2 is a factor analysis performed on the results of 64 items taken from the Artemis Intake Questionnaire, a clinically relevant tool constructed by treatment providers used in working with the victims of domestic violence. Results indicate that reported humiliation and blame of the victim accounted for the largest amount of variance, followed by controlling the victim, and then physical violence. Results of this factor analysis suggest that greater emphasis must be put on factors other than physical violence in the construction of future domestic violence scales.KEY WORDS: domestic violence; emotional abuse; assessment; literature review. INTRODUCTIONIn the last 10-15 years, domestic violence against an adult partner (DV) has been identified as a major health and safety concern (Browne & Williams, 1993;Campbell, 1995a;Walker, 1994). Consequently, research in this area has exploded and, as seen in a literature review addressed in this paper, numerous DV assessment scales have been created (e.g., Dutton & Golant, 1995;Dwyer, 1999;Kropp & Hart, 2000;Marshall, 1992;Shepard & Campbell, 1992;Straus et al., 1996). This paper presents two related studies. The overall purpose of these studies is two-fold: (a) to determine the type of psychometric evidence employed in the development of these current domestic violence measurement techniques and the constructs that are commonly incorporated, and (b) to determine whether the constructs of intimidation, power, and control (in addition to physical violence) are supported by victims' responses and should thus be emphasized in future domestic violence scales and models. The first study consisted of a literature review of existing domestic violence scales in order to examine what constructs are commonly incorporated into the scales, and whether they had been psychometrically evaluated. The review focused on two issues: (a) the types of psychometric support used during the development of these scales, and (b) the overall content of these scales to determine what constructs were most heavily represented. There were two general hypotheses. The first hypothesis concerning the review of these scales was that the majority would focus psychometrically on demonstrating construct validity (convergent, divergent, factor analytic), with some providing evidence for content validity and only a few using predictive validity (concurrent and/or future). Further, we expected that the majority of scales (probably greater than 60%) would provide evidence of 340Strauchler, McCloskey, Malloy, Sitaker, Grigsby, and Gillig reliability consisting of internal consistency and/or temporal stability. The ...

No abstract

Masters-level social work clinicians at two urban psychiatric emergency services self-rated their cognitive reactions and conscious feelings toward a total of 783 patients. Negative self-reported affect toward patients was related to their displaying overt psychotic symptoms, having a final diagnosis of substance abuse, being impulsive, unemployed, or having a history of criminal behavior. However, clinicians' reactions did not enter into the model predicting hospitalization, which included opinion of family and friends, degree of psychopathology, suicidal ideation and lack of social supports as the major predictors of whether or not a patient would be hospitalized.

No abstract