Concerning least squares analysis of experimental data.

Overall, John E.; Spiegel, Douglas K.

doi:10.1037/h0028109

Cited by 403 publications

(227 citation statements)

References 9 publications

Supporting

Mentioning

222

Contrasting

Unclassified

Order By: Relevance

“…It can be appreciated by analogy to analysis of variance. In the use of analysis of variance to determine whether the weights and scale values are confounded in an averaging task, the significance of the interactions are Cohen, 1968;Overall & Spiegel, 1969 Cohen (1968) has shown how nonlinear interactions can be included in the model as well as linear ones. The most elegant test of Equation 3 as a derivation of averaging would be to show that all of the interaction variance was predicted by the simple Linear X Linear component.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Decision making and information integration in the courts: The setting of bail.

Ebbesen¹,

Konečni²

1975

Journal of Personality and Social Psychology

145

View full text Add to dashboard Cite

Two studies were conducted to determine how real felony court judges decide the amount of bail to set. In the first, the judges were presented with fictitious case histories containing the relevant information in a factorial design. In the second, multiple regression techniques were used to examine the impact of different kinds of information on decisions made by judges in actual bail hearings. In the simulated cases, the judges seemed to be influenced most by the degree to which the accused was tied to the area and whether or not he had a prior criminal record. However, the judges' actual bail decisions were not at all affected by these variables. Instead, their decisions were almost exclusively determined by the district attorneys' recommendations. Both the district attorneys' and the defense attorneys' actual recommendations were found to be primarily based upon the severity of the crime. These results were discussed in terms of the utility of simulation and observational research for drawing applied and theoretical conclusions.

show abstract

Section: Resultsmentioning

confidence: 99%

“…Furthermore, once these cases are selected, it is possible to analyze them using dummy-variable multiple regression, a technique that is identical to the least-squares analysis of variance (Applebaum & Cramer, 1974;Cohen, 1968;Overall & Spiegel, 1969;Wolf & Cartwright, 1974).…”

Section: Comparison Of Experiments 1 Andmentioning

confidence: 99%

Decision making and information integration in the courts: The setting of bail.

Ebbesen¹,

Konečni²

1975

Journal of Personality and Social Psychology

145

View full text Add to dashboard Cite

show abstract

“…However, to rule out the potential bias due to defining the mating categories as continuous variables with equidistant phenotypes, we also repeated the analyses comparing between two discrete character states of mating system [monogamous (N ϭ 61) vs polygynous (N ϭ 17)]. The alternative would have been to use ANOVA, equivalent to multiple regression on similar dummy variables, but constrained to sum up to zero (Overall and Spiegel, 1969;Dobson, 1990). Because the data are unbalanced with respect to the factors analyzed, no matter which analysis is used, the problem of colinearity and thus, deciding which variable is more influential remains.…”

Section: Control For Phylogenymentioning

confidence: 99%

Comparative analysis of male androgen responsiveness to social environment in birds: the effects of mating system and paternal incubation

Hirschenhauser¹,

Winkler²,

Oliveira³

2003

Hormones and Behavior

138

View full text Add to dashboard Cite

Male androgen responses to social challenges have been predicted to vary with mating system, male-male aggressiveness, and the degree of paternal investment in birds ("challenge hypothesis," Am. Nat. 136 (1990), 829). This study focused on the interspecific predictions of the challenge hypothesis. Comparative methods were used to control for effects of the phylogenetic relatedness among the sampled species. Male androgen data of 84 bird species were collected from literature records on seasonal androgen patterns. From these, the androgen responsiveness (AR) was calculated as described in the original challenge hypothesis (i.e., maximum physiological level/breeding baseline). Scatterplots of AR versus mating strategy, male-male aggressiveness, and the degree of paternal care confirmed the expected interspecific patterns. When phylogenetic analyses were performed among all of the sampled species, the effects of paternal investment disappeared while the AR remained covarying to a high degree with mating system and male-male aggressiveness. Although these mechanisms may be different at the intraspecific level, this suggests that interspecific differences of AR in male birds may have evolved in response to changes of mating strategies, rather than in response to altered paternal duties. However, control for phylogeny among the subsample of 32 passerine species revealed that if any paternal investment contributed to the observed variance in AR, then the change from "no male incubation" to "male shares incubation duties" represented the most effective, whereas the male's contribution to feeding offspring did not explain the observed variation of AR.

show abstract

“…For example, consider the mainframe version of SPSS's general linear model command in the 1980's. Option 9, a contrast coding least squares regression approach due to Overall and Spiegel (1969), was subsequently shown to test no known statistical hypothesis (see, e.g., Blair & Higgins, 1978a;Blair, 1978;Blair & Higgins, 1978b). Another example is the R aov() function "when conducting analysis of covariance" which "does not work correctly" (Schumacker, 2015, p. 288).…”

Section: Introductionmentioning

confidence: 99%

Caution for Software Use of New Statistical Methods (R)

Lorenz

Markman²,

Sawilowsky³

2015

J. Mod. App. Stat. Meth.

View full text Add to dashboard Cite

Open source programming languages such as R allow statisticians to develop and rapidly disseminate advanced procedures, but sometimes at the expense of a proper vetting process. A new example is the least trimmed squares regression available in R's lqs() in the MASS library. It produces pretty regression lines, particularly in the presence of outliers. However, this procedure lacks a defined standard error, and thus it should be avoided.

show abstract

Concerning least squares analysis of experimental data.

Cited by 403 publications

References 9 publications

Decision making and information integration in the courts: The setting of bail.

Decision making and information integration in the courts: The setting of bail.

Comparative analysis of male androgen responsiveness to social environment in birds: the effects of mating system and paternal incubation

Caution for Software Use of New Statistical Methods (R)

Contact Info

Product

Resources

About