How Come So Many Hypotheses in Educational Research are Supported? (A Modest Proposal)

Cohen, S. Alan; Hyman, Joan S.

doi:10.3102/0013189x008011012

Cited by 28 publications

(10 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Given these small effect sizes, the relations between femininity and self-esteem and the interaction and self-esteem may have little practical significance despite their statistical significance (J. Cohen, 1977;S. Cohen & Hyman, 1979).…”

Section: Congruence Androgyny and Masculinity Hypothesesmentioning

confidence: 99%

Sex role orientation and self-esteem: A critical meta-analytic review.

Whitley¹

1983

Journal of Personality and Social Psychology

298

190

View full text Add to dashboard Cite

Research on the relation between sex role orientation and psychological well-being has been guided by one of three models. The traditional congruence model holds that psychological well-being is fostered only when one's sex role orientation is congruent with one's gender; the androgyny model proposes that well-being is maximized when one's sex role orientation incorporates a high degree of both masculinity and femininity regardless of one's gender; the masculinity model posits that well-being is a function of the extent to which one has a masculine sex role orientation. The adequacy of these three models was tested by means of a meta-analysis of 35 studies of the relation between sex role orientation and self-esteem, the indicator of psychological well-being most widely used in sex role studies. The results of the meta-analysis were most supportive of the masculinity model and showed that the strength of observed relations between sex role orientation and self-esteem varied as a function of both the sex role measure and the type of self-esteem measure used in the studies. In addition, a number of methodological issues was identified that should be taken into consideration in future research.

show abstract

Section: Congruence Androgyny and Masculinity Hypothesesmentioning

confidence: 99%

Sex role orientation and self-esteem: A critical meta-analytic review.

Whitley¹

1983

Journal of Personality and Social Psychology

298

190

View full text Add to dashboard Cite

show abstract

“…Similarly, Carver (1978) argued that estimates of treatment effects "should be interpreted with respect to the research hypothesis regardless of its statistical significance" (p. 394). It is clear that following Cohen and Hyman's (1979) and Carver's (1978) advice, in the area of psychotherapy efficacy studies in which samples are small and nonequivalence is therefore highly probable (see Table 1 and Figure I), could easily result in the misinterpretation of biased andand misleading estimates of treatment efficacy. Clearly, the worst situations of this type occur with the reversal paradox in which the wrong treatment could be estimated to yield the better outcome.…”

Section: Consequences Of Nonequivalence: Simpson's Paradoxmentioning

confidence: 99%

Random sampling, randomization, and equivalence of contrasted groups in psychotherapy outcome research.

Hsu¹

1992

Methodological Issues &Amp; Strategies in Clinical Research.

View full text Add to dashboard Cite

Simple random sampling and random assignment (randomization) are some of the best and most popular methods of attaining the pretreatment equivalence of contrasted groups in psychological research (see Cook & Campbell, 1979), in medical research (see OFallon et al., 1978), and in the specialized area of psychotherapy efficacy studies (Huesmann, 1982; Kendall & Norton-Ford, 1982; D. A. Shapiro & Shapiro, 1983). In fact, in a recent meta-analysis of comparative psychotherapy outcome research that focused on some of the best studies in this area, D. A. Shapiro andShapiro (1982, 1983) noted that unconstrained randomization was used in 57% of the client groups.One of the most appealing characteristics of random sampling and randomization is that these methods can equate groups on several nuisance variables simultaneously and that these methods do not require the researcher to be aware of (a) how the important nuisance variables are related Reprinted from the Jmml of Consulting and Clinical Psychdogy, 57, 131-137.

show abstract

“…Tversky and Kahneman's (1971) points appear worth repeating because of the numerous criticisms of significance tests that have recently appeared in psychological journals (see Kupfersmid, 1988, for a summary). Cohen and Hyman (1979), for example, stated that "even if. .…”

Section: Consequences Of Nonequivalence: Simpson's Paradoxmentioning

confidence: 99%

“…Similarly, Carver (1978) argued that estimates of treatment effects "should be interpreted with respect to the research hypothesis regardless of its statistical significance" (p. 394). It is clear that following Cohen andHyman's (1979) andCarver's (1978) advice, in the area of psychotherapy efficacy studies in which samples are small and nonequivalence is therefore highly probable (see Table 1 and Figure 1), could easily result in the misinterpretation of biased and misleading estimates of treatment efficacy. Clearly, the worst situations of this type occur with the reversal paradox in which the wrong treatment could be estimated to yield the better outcome.…”

Section: Consequences Of Nonequivalence: Simpson's Paradoxmentioning

confidence: 99%

Random sampling, randomization, and equivalence of contrasted groups in psychotherapy outcome research.

Hsu¹

1989

Journal of Consulting and Clinical Psychology

103

View full text Add to dashboard Cite

Random sampling and random assignment (randomization) are some of the most popular methods of equating contrasted groups on pre-existing nuisance variables. However, the small samples typically used in psychotherapy outcome studies raise some questions about the extent to which these methods eliminate the pretreatment nonequivalence of groups in this area of research. This article identifies conditions under which equivalence is likely (and unlikely) to be attained with simple random sampling and randomization in psychotherapy efficacy studies of the kind examined in recent meta-analyses. Some consequences of nonequivalence are viewed as manifestations of Simpson's paradox. Misinterpretations of estimates of the relative efficacy of treatments are expected in view of belief in the law of small numbers. The minimum sample sizes needed to protect against nonequivalence are compared with those needed to satisfy several other criteria.

show abstract

How Come So Many Hypotheses in Educational Research are Supported? (A Modest Proposal)

Cited by 28 publications

References 8 publications

Sex role orientation and self-esteem: A critical meta-analytic review.

Sex role orientation and self-esteem: A critical meta-analytic review.

Random sampling, randomization, and equivalence of contrasted groups in psychotherapy outcome research.

Random sampling, randomization, and equivalence of contrasted groups in psychotherapy outcome research.

Contact Info

Product

Resources

About