Person:

Rosenthal, Robert

The purpose of this chapter is to provide a context for thinking about the role of ethics in quantitative methodology.We begin by reviewing the sweep of events that led to the creation and expansion of legal and professional rules for the protection of research subjects and society against unethical research. The risk–benefit approach has served as an instrument of prior control by institutional review boards. After discussing the nature of that approach,we sketch a model of the costs and utilities of the “doing” and “not doing” of research.We illustrate some implications of the expanded model for particular data analytic and reporting practices.We then outline a 5 × 5 matrix of general ethical standards crossed with general data analytic and reporting standards to encourage thinking about opportunities to address quantitative methodological problems in ways that may have mutual ethical and substantive rewards. Finally,we discuss such an opportunity in the context of problems associated with risk statistics that tend to exaggerate the absolute effects of therapeutic interventions in randomized trials.

The purpose of this article is twofold: first, to present the strategic advantages of focused over omnibus tests of statistical significance in counseling research and, second, to demonstrate the utility of interpreting the magnitude of the effect as a way of assessing practical significance (i.c., in addition to computing p levels). Simple procedures are given for computing effect size using the product-moment r in the context of focused significance testing. We show that recasting the effect size into a binomial display is a convenient way of revealing the practical significance of r in terms of whatever equivalency measure is deemed appropriate (e.g., improvement rate, success rate, survival rate).

This reply to Meyer explains again that cell means, although usually the results of greatest interest, should not be confused with interaction effects. Unless all main effects are 0, one cannot accurately interpret an interaction by plotting the cell means. To interpret an interaction, it is the residuals remaining after removal of constituent effects (e.g., row and column effects in 2-factor analyses) that must be examined.

To describe a systematic quantitative approach to assessing the predictions made by competing theories using contrasts and correlational indices of effect sizes. We illustrate the use of the contrast F and t to compare and combine predictions when the raw data are continuous scores, and z contrasts when working with frequencies in 2 x k tables of counts. The traditional effect size correlation indicates the magnitude of the effect on individual scores of participants' assignment to particular conditions. The contrast correlation obtained from the contrast F or t is, in some cases, the easiest way of estimating the effect size correlation in designs using more than two groups. The alerting correlation is another way of appraising the predictive power of a contrast and can be used to compute the contrast F from published results when all we have are condition means and the omnibus F from an overall analysis of variance. Omnibus Fs, those with more than 1 df in the numerator, are rarely useful in data analytic work since they address unfocused questions, yielding only vague answers. Asking focused questions using contrasts increases the clarity of our questions and the clarity and statistical power of our answers.

We describe convenient statistical procedures that will enable research consumers (e.g,, professional psychologists, graduate students, and researchers themselves) to reach beyond the published conclusions and make an independent assessment of the reported results. Appropriately conceived contrasts accompanied by effect size estimates often allow researchers to address precise predictions that the authors of the published report may have ignored or abandoned prematurely. We describe the use of t, F, and Z to compute contrasts with different raw ingredients, and we review 3 effect size indices (Cohen's d, Hedges's g, and the Pearson r) and a way of displaying the magnitude of any effect size r. We also describe how to construct confidence limits for the obtained effect as well as its null-counternull interval.

When interpreting an interaction In the analysts of variance (ANOVA), many active researchers (and, In turn, students) often Ignore the residuals defining the interaction Although this problem has been noted previously, It appears that many users ofANOVA remain uncertain about the proper understanding of interaction effects To clear up this problem, we revIew the way In which. the ANOVA model enables us to take apart a table of group means or the individual measurements contributing to the means to reveal the underlying components We also show how (using only published data) to compute a contrast on the question that may be of primary Interest and Illustrate strategies for interpreting tables of residuals We conclude with an exercise to check on students' understanding ofANOVA and to encourage Increased preCISIOn In the specification of research results.

ThIs reply to Abelson (this Issue) and Petty, Fabrtgar, Wegener, and Priester (this issue) is couched within the framework offive basic principles advising that we (1) hang on to what we predicted long enough to test It (Tarzan's leap), (2) be wary of beguiling statistical designs that may not address the question of Interest (the Sirens' song), (3) not allow well worn habits of thinking to ensnare our perceptions (Lavoisier's crease), (4) weigh the possIbIlity of more than one correct hypothesis (the dayyan's decree), and (5) not confuse unplanned wIth planned contrasts (the archer's aim).

The effect size (ES) is the magnitude of a study outcome or research finding, such as the strength of the relationship obtained between an independent variable and a dependent variable. Two types of ES indicators are sampled here: the difference-type and the correlational (or r-type). Both are well suited to situations in which there are two groups or two conditions, whereas the r-type, used in association with focused statistical procedures (contrasts), is also ideal in situations where there are more than two groups or conditions and there are predicted overall patterns to be evaluated. Also discussed are procedures for computing confidence intervals and null-counternull intervals as well as a systematic approach to comparing and combining competing predictions expressed in the form of contrast weights and ES indicators.