Nudging aside Meehl's paradox.

P.E. Meehl (1967) attacked significance tests used to appraise directional theoretical predictions. With perfect statistical power, the probability approaches .5 of significance in the predicted direction, even for meritless theories. Feeble theory corroboration results. This article argues that directional predictions, not significance tests, produce the feeble corroboration. The author reviews previous solutions to Meehl, rejecting all except the multiple corroboration solution by D.T. Lykken (1968) and A. Kukla (see record 1991-31958-001). This solution is defended against Meehl's (see record 1991-31961-001) criticisms and extended. The complexity of a typical research design enables a theory to make multiple directional predictions, each of which can be adequately tested with a significance test. Reasonable theory corroboration results when all or most predictions succeed, and techniques to assess this are discussed here. (French abstract) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Meehl revisited: A look at paradigms in psychology.

Comments on Meehl's (see record 1979-25042-001) article on the nature of psychological debates. The author is in agreement with Meehl that there is a need to reconsider how research is designed and carried out within psychology. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Review of Clinical Versus Statistical Prediction.

Reviews the book "Clinical versus statistical prediction," by Paul Everett Meehl (see record 1996-97896-000). Meehl attempts to plot a course to the port of valid prediction through the "rigorous" channels of statistical methods and the "sophisticated" undercurrents of clinical dynamics. Most of the book is devoted to a thoughtful but discursive analysis of the alternative strategies for prediction of statistical or actuarial methods as compared with clinical or casestudy methods. Meehl believes that much confusion could be avoided if careful distinction were made between two different uses of statistics, the types of data involved in prediction, and the methods of combining data in making predictions. Meehl does an excellent job of raising some of the basic problems and issues which have to be considered in carrying out research in this area. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Limitations of Meehl's MAXCOV-HITMAX procedure.

Comments on P. E. Meehl's (see record 1995-29291-001) article claiming that the MAXCOV-HITMAX procedure (MAXCOV) can be used to determine whether a latent structure, e.g., the structure underlying variation in schizotypal symptomatology, is taxonic or typological rather than dimensional. Miller contends that (1) the covariance curve is not necessarily flat when there are no latent taxa, (2) most published MAXCOV studies used dichotomous variables to compute covariances that may have lead to false detection of taxa, and (3) researchers relying on visual inspection of covariance curves may be misled by their choice of scale and aspect ratio. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A Note on Mixed Group Validation.

Comments on the original article by R. M. Dawes and P. E. Meehl (see record 1966-11283-001) which considered mixed group validation. In the current article, a basic assumption of mixed group validation is questioned and a means of checking the validity of the assumption is suggested. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

"Nudging aside Meehl's paradox": Erratum.

Reports a typesetting error in the original article by L. Leventhal (Canadian Psychology, 1994[Jul], Vol 35[3], 283–298). On page 295, the sentence should read "Compared to the old views, the new views might give less weight to the confidence interval and more weight to the probability value of .007 which indicates that the theory performed far better than chance." (The following abstract of this article originally appeared in PA, Vol 82:8115.) P. E. Meehl (1967) attacked significance tests used to appraise directional theoretical predictions. With perfect statistical power, the probability approaches .5 of significance in the predicted direction, even for meritless theories. Feeble theory corroboration results. This article argues that directional predictions, not significance tests, produce the feeble corroboration. The author reviews previous solutions to Meehl, rejecting all except the multiple corroboration solution by D. T. Lykken (1968) and A. Kukla (see record 1991-31958-001). This solution is defended against Meehl's (see record 1991-31961-001) criticisms and extended.… (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Deeper into the schizotypy taxon: On the robust nature of maximum covariance analysis.

Prior research has suggested that the latent structure of the schizotypy construct (P. E. Meehl, 1990) may be qualitative in nature and have a low base rate (L. Korfine & M. F. Lenzenweger, 1995; M. F. Lenzenweger & L. Korfine, 1992). These studies relied on the application of maximum covariance analysis (MAXCOV) to 8 true–false format items from a schizotypy measure. The current study sought to examine the robustness of those prior findings through MAXCOV analysis of fully quantitative measures of schizotypy. Measures of perceptual aberration, magical ideation, and referential thinking were analyzed using MAXCOV in a sample of 429 persons. The results of these analyses strongly support a latent taxonic structure for schizotypy and a low base rate for the schizotypy taxon. Furthermore, the members of the putative taxon reveal an increased level of deviance on a psychometric measure known to be associated with schizophrenia liability. The possibility that the dichotomous item format of those items analyzed previously with MAXCOV lead to spurious pseudotaxonicity is greatly diminished in light of these results. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Commentary on Meehl and Waller's (2002) Path Analysis and Verisimilitude.

P. E. Meehl and N. G. Waller's (see record 2002-18342-001) proposed method may not yield unique solutions for model parameters nor unique solutions for model lack of fit. The author argues from a naturalistic-cognitive philosophy of science that science seeks objective knowledge and that hypothesis testing is central to achieving that goal. It is also argued that P. E. Meehl and N. G. Waller's proposal blurs the distinction between hypothesis testing and explorations of the data seeking an optimal model to serve as a prospective inductive generalization. But it is noted that inductive generalizations are never unique and must be tested to eliminate those that reflect subjective aspects of the researcher's methods and points of view. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Taxometrics and the revision of diagnostic criteria.

Comments on P. E. Meehl's (see record 1995-29291-001) article describing taxometric (mathematical) procedures used to classify psychiatric patients. Garb claims that Meehl did not cite studies that do not support the use of taxometric procedures, nor discuss how taxometric methods can be used to revise diagnostic criteria. To use taxometric methods, one needs to be correct in conjecturing the existence of a latent taxon, only it is very hard to know for sure when a single taxon exists. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Review of Clinical versus statistical prediction: A theoretical analysis and review of the evidence.

Reviews the book "Clinical versus statistical prediction: A theoretical analysis and review of the evidence" by Paul E. Meehl (see record 1996-97896-000). This book talks about a continuing debate among psychologists regarding the relative accuracy and efficiency of statistical (actuarial) predictions and those made by clinicians on the basis of subjective "understanding" of individual cases. This book represents the author's first published statement of his position. In the reviewers opinion, the author has succeeded admirably. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Identifying schizoid-taxon membership with the Golden-Meehl MMPI items.

Seven MMPI items proposed by R. R. Golden and P. E. Meehl (see record 1979-25088-001) as markers for the schizoid taxon were investigated in a sample of 383 randomly selected psychiatric inpatients assigned to 1 of 8 groups according to the number of items they endorsed (0–7). Ss were then divided into 4 diagnostic groups: schizophrenia, major affective disorder, nonpsychotic, and a residual group. ANOVA revealed overall significant differences among the 4 diagnostic groups, with a posteriori comparisons showing a higher endorsement rate for the nonpsychotics than for the schizophrenics. Further examination of these items, both individually and as a set, and an alternative indicator for the schizoid taxon proposed by Golden and Meehl found schizophrenics to be consistently underidentified as schizoid-taxon members relative to the number of nonpsychotics so identified and to base rates, regardless of the cutting score used. More conventional procedures for selecting inventory items sensitive to taxon variances are recommended. (4 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Improving construct validity: Cronbach, Meehl, and Neurath's ship: Comment.

Smith's article "On Construct Validity: Issues of Method and Measurement" (see record 2005-16347-002) is a fine tribute to L. J. Cronbach and P. E. Meehl (1955) that clarifies the current state and future directions in the understanding of construct validity. Construct validity is a dynamic process, and fit indices need to be used at the service of understanding, not in place of it. The failure of a study or set of studies to support a construct, a measure, or the theory underlying it admits of many explanations, and the ways scientists interpret such failures are prone to cognitive biases and motivated reasoning. This suggests why metrics designed to index the extent to which observations match expectations can be useful prostheses to scientific judgments. As P. E. Meehl (1954) showed decades ago, quantitative, statistical formulas and indices tend to outperform informal, qualitative judgments, and this applies as much to the way researchers evaluate constructs and measures as to judgments in the consulting room. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Comments on the Meehl-Waller (2002) procedure for appraisal of path analysis models.

P. E. Meehl and N. G. Waller (see record 2002-18342-001) proposed an innovative method for assessing path analysis models wherein they subjected a given model, along with a set of alternatives, to risky tests using selected elements of a sample correlation matrix. Although the authors find much common ground with the perspective underlying the Meehl-Waller approach, they suggest that there are aspects of the proposed procedure that require close examination and further development. These include the selection of only one subset of correlations to estimate parameters when multiple solutions are generally available, the fact that the risky tests may test only a subset of parameters rather than the full model of interest, and the potential for different results to be obtained from analysis of equivalent models. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The predictive power of ridge regression: Some quasi-simulation results.

Reviews research indicating that ridge regression tends to improve the mean square error of prediction obtained with ordinary least squares (OLS) regression in a wide range of conditions. MMPI profiles of 861 psychiatric patients and diagnoses from 29 psychologists, as collected by P. E. Meehl (see record 1960-04396-001), were used to examine the gains in cross-validated multiple correlation obtained with ridge regression compared with OLS and equal weights as a function of sample size and ridge constant. A simple formula for estimating the ridge constant was also evaluated. Results that are related to recent developments concerning the use of Bayesian regression procedures show that ridge regression improves both the mean square error of prediction and the cross-validated multiple correlation obtained with OLS when the ratio sample size to number of predictor variables is relatively small. (19 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reflecting interactions among personality items: Meehl's paradox revisited.

Regarding configural scoring, P. E. Meehl (see record 1951-00366-001) provided an artificial example in which 2 binary items were not individually correlated with the incidence of schizophrenia; yet, if item interactions were taken into account, the items could accurately predict schizophrenia. Relative to this paradox, the present study shows how substantial new information can be used by uncovering item interactions in personality scales. The California Psychological Inventory (CPI) was analyzed using data consisting of 472 binary CPI item scores taken from 5,219 high school students residing in small US cities in the 1950's, originally collected by H. G. Gough (1975). Results show that for most subscales, a substantial number of items interacted such that some item weights depended on how some other items were scored. Subsequent analyses indicated that identifying interacting items pointed toward interesting subscales and subgroups of individuals as well as augmented interactive scales that were substantially more reliable than existing scales. (20 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Comments on Meehl and Rosen's paper.

In a paper by Meehl and Rosen (see 30: 2902) a rationale for evaluating the predictive efficiency of psychometric instruments is presented. The purpose of this paper is to emphasize the importance of administrative policy in any consideration of statistical criteria. A discussion of Meehl and Rosen's Case 1 situation is presented. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A good match is hard to find: Comment on Mayer, Tajika, and Stanley (1991).

R. E. Mayer et al (see record 1991-19796-001) reported that US students scored higher on problem-solving tasks than their Japanese peers when matched on computation skills. Contrary to Mayer et al, the authors believe that these results are artifacts that reveal little about the nature of Japanese and US education. Drawing on work by P. E. Meehl (1970), the Mayer et al study is used to illustrate pitfalls of matching cases in an attempt to create equivalent groups of Ss. Besides well-known regression effects, matching (whether by selecting Ss or by statistical correction) is subject to 3 major criticisms: (1) It produces systematically unmatched samples with respect to other meaningful variables, (2) it results in unrepresentative samples from which it is hard to draw meaningful generalizations, and (3) causal interpretation becomes nearly impossible. These pitfalls must be considered when interpreting the results of ex post facto matched designs. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

More theoretical risks.

Responds to the comments by F. Paniagua ("Kuhn's paradigmatic view of psychology and Skinner's theory of behavior." Theoretical and Philosophical Psychology, 11, 1991, 122-125) on the current author's original article, "Meehl revisited: A look at paradigms in psychology" (Theoretical and Philosophical Psychology 9, 1989, 30-36), in which the current author reviewed Paul Meehl's (see record 1979-25042-001) famous article "Theroetical risks and tabular asterisks: Sir Karl; Sir Ronald, and the slow progress of soft psychology." According to the current author, Paniagua takes exception to two casual remarks made in the current author's paper, one about Kuhn and the other about Skinner, but neither remark is related to the actual thesis. Paniagua's comments do not carry the substantive aspects of the article forward, which is unfortunate asserts the current author, because the theory discussed therein may prove useful in understanding the nature and evolution of psychology. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Got a minute? Attentional limits revisited.

Responds to comments by R. C. Tees (see record 1991-03035-001), J. G. Adair (see record 1991-03013-001), J. E. Grusec (see record 1991-03021-001), K. Danziger (see record 1991-03016-001), L. P. Mos (see record 1991-03027-001), H. J. Stam (see record 1991-03033-001), and V. Vikis-Freilbergs (see record 1991-03039-001) on W. Thorngate's (see record 1991-03036-001) contention that the overproduction of psychological literature will lead to fragmentation of the discipline. As the works of psychologists proliterate, their scientific truth may capture less attention than their practical importance or personal interest. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The predictive power of ridge regression: Some quasi-simulation results: Correction to Cattin.

Reports an error in the original article by Philippe Cattin (Journal of Applied Psychology, 1981, Vol. 66, No. 3, pp. 282-290). The fifth sentence in the first paragraph on page 284 contains an error. The sentence should read: "For intermediate values of k?, the ridge regression weights are 'weighted sums' of the OLS regression weights [not models] and of the zero-order sample correlations and tend to decrease in absolute value as k? increases." (The following abstract of this article originally appeared in record 1981-27117-001.) Reviews research indicating that ridge regression tends to improve the mean square error of prediction obtained with ordinary least squares (OLS) regression in a wide range of conditions. MMPI profiles of 861 psychiatric patients and diagnoses from 29 psychologists, as collected by P. E. Meehl (see record 1960-04396-001), were used to examine the gains in cross-validated multiple correlation obtained with ridge regression compared with OLS and equal weights as a function of sample size and ridge constant. A simple formula for estimating the ridge constant was also evaluated. Results that are related to recent developments concerning the use of Bayesian regression procedures show that ridge regression improves both the mean square error of prediction and the cross-validated multiple correlation obtained with OLS when the ratio sample size to number of predictor variables is relatively small. (PsycINFO Database Record (c) 2010 APA, all rights reserved)