The many definitions of test bias.

Argues that the definition of test bias has many widely disparate aspects stemming from different universes of discourse. The status of each of these is reviewed. It is suggested that all the various aspects of test bias be kept in mind so that the risk of losing perspective when one operational definition is settled on is lessened. The concept, being a public one, cannot encompass all the definitions. Research on these issues is some of the more exciting and significant being done today, but confusion over which is the real issue is unnecessary because they are all the real issue. (24 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On differential validity and bias.

Responds to J. E. Hunter and F. L. Schmidt's criticism (see record 1979-22562-001) that the present authors underestimated Type I bias and overestimated Type II bias in their analysis of studies of ethnic differences in employment test validity (see record 1978-02212-001). It is suggested that it is premature to dismiss the issue of differential validity. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A note on minority group test bias studies.

Suggests that under certain conditions, comparisons of majority and minority group regression lines for purposes of assessing test bias can be viewed as comparisons of conditional bivariate distributions. Under conditions of trivariate normality, findings should reveal parallel regression lines except for a special case. One implication is that even when the test is a parallel form of the criterion, lines with equal slopes but unequal intercepts should be found. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A Combi-Stroop test for measuring food-related attentional bias.

There are methodological complexities with the supraliminal–lexical versions of the modified versions of the Stroop tests that could be responsible for inconsistencies across the literature (Field & Cox, 2008). We tested whether a combination of subliminal–pictorial and classic Stroop tests can differentiate between dieters' and nondieters' food attentional bias (FAB). Participants were dieters (n = 30) and nondieters (n = 32) who were tested 3 hr after having a meal. Each picture from among 24 high-calorie and 24 low-calorie food pictures was presented for 32 ms before the appearance of a congruent or an incongruent color word, in response to which participants were required to manually report, via a tagged keyboard, the correct color of the word as quickly and accurately as possible. Color-naming latencies and interference scores were calculated. Dieters showed the highest reaction times to incongruent color words following high-calorie food pictures; overall, dieters showed significantly higher FABs than nondieters. The Combi-Stroop test has differential validity. Moreover, findings suggest that FAB can result from early allocation of dieters' attention to food-related stimuli. (PsycINFO Database Record (c) 2011 APA, all rights reserved)     

Revival of test bias research in preemployment testing.

We developed a new analytic proof and conducted Monte Carlo simulations to assess the effects of methodological and statistical artifacts on the relative accuracy of intercept- and slope-based test bias assessment. The main simulation design included 3,185,000 unique combinations of a wide range of values for true intercept- and slope-based test bias, total sample size, proportion of minority group sample size to total sample size, predictor (i.e., preemployment test scores) and criterion (i.e., job performance) reliability, predictor range restriction, correlation between predictor scores and the dummy-coded grouping variable (e.g., ethnicity), and mean difference between predictor scores across groups. Results based on 15 billion 925 million individual samples of scores and more than 8 trillion 662 million individual scores raise questions about the established conclusion that test bias in preemployment testing is nonexistent and, if it exists, it only occurs regarding intercept-based differences that favor minority group members. Because of the prominence of test fairness in the popular media, legislation, and litigation, our results point to the need to revive test bias research in preemployment testing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On the bias of information estimates.

Considers the problem of the calculation of the bias of the maximum likelihood information estimate H, based on independent choices among k events. The expectation EH is calculated exactly as a function of the probabilities p1, p2, . . . , pkk. The bias H - EH is approximated by using a convergent expansion for a logarithm and using the 1st 2 terms of a finite expansion for the jth moment of a random variable. The resulting approximation is more generally valid, although less concise and simple, than the classical Miller-Madow approximation. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On detecting sex bias in salaries.

Comments that the model used by M. H. Birnbaum (1979, 1981) to derive his diagnostic test makes the implausible assumption that salary depends only on true quality of performance and is not influenced at all by common measures of merit (e.g., number of publications, years of experience). (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Cultural test bias: Comment on Hunter and Schmidt.

Responds to J. E. Hunter and F. L. Schmidt's (see record 1977-20109-001) criticisms of the author's 1971 paper on defining a culture-fair test. The present author identifies 3 classes of solutions for defining culture-fair tests—idiosyncratic, mechanical, and rational. While identifying himself in the rational position and Hunter and Schmidt in the idiosyncratic, the present author criticizes them for the contradictory tone of their article. It is concluded that Hunter and Schmidt are actually in general agreement with the present author with the possible exception of the question of whether value judgments can be usefully quantified. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On belief bias in syllogistic reasoning.

A multinomial model is used to disentangle the respective contributions of reasoning processes and response bias in conclusion-acceptance data that exhibit belief bias. A model-based meta-analysis of 22 studies reveals that such data are structurally too sparse to allow discrimination of different accounts of belief bias. Four experiments are conducted to obtain richer data, allowing deeper tests through the use of the multinomial model. None of the current accounts of belief bias is consistent with the complex pattern of results. A new theory of belief bias is proposed that assumes that most reasoners construct only one mental model representing the premises as well as the conclusion or, in the case of an unbelievable conclusion, its logical negation. New predictions derived from the theory are confirmed in 4 additional studies. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Bias in defining test bias: Reply to Darlington.

Argues that R. B. Darlington's response (see record 1979-22558-001) to the present authors' original criticism (see record 1977-20109-001) is not satisfactory. It is suggested that Darlington's "rational" viewpoint cannot reconcile the issue of individual rights vs quotas and that his contention that the present authors' article was contradictory is incorrect. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Racial and ethnic bias in psychological tests: Divergent implications of two definitions of test bias.

Examines the 2 most widely accepted definitions of unfair test bias. A. Cleary in 1968 defined unfair test bias in terms of consistent under- or overprediction of actual performance levels of minority or majority groups. R. L. Thorndike's 1971 definition holds that a test is unfairly biased whenever the difference between the minority and majority groups is greater on the test than on actual performance. These 2 definitions, which superficially appear to be similar, are shown to be very different in their implications for minority selection. It is demonstrated that whenever test validity is less than perfect and mean majority criterion performance is higher than that of the minority group, Thorndike's definition leads to acceptance of a larger percentage of minority group members than does the Cleary definition. A review and reanalysis of published research on test bias suggests that most conventional test usages are unbiased by Cleary's definition but unfair by Thorndike's criteria. Advantages and disadvantages of both concepts of unfair test bias are discussed. (32 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Selective citation mars conclusions about test validity and predictive bias.

Comments on the article by Vasquez and Jones (see record 2006-01690-003), in which they put forward the argument that standardized tests do not evaluate much of anything worthwhile and do not assess merit. The current authors argue that Vasquez and Jones support their argument only through highly selective citations from the literature, and they discuss Vasquez and Jones' evidence for both predictive validity and predictive bias in educational and employment settings. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Multiple sources of test bias on the WISC-R and Bender-Gestalt Test.

Test bias on the WISC-R and Bender-Gestalt Test was assessed through multiple statistical procedures: internal consistency, item difficulty, correlations of item–total correlations, concurrent validity, and construct validity. These procedures were applied to WISC-R and Bender data of 436 7–14 yr olds from 3 racial–ethnic groups (Anglo-American, Black, and Mexican American) and 2 social classes (low and middle). The statistical procedures investigated bias as a function of children's sex, race, socioeconomic status, age, birth order, health, family size, family structure, and urban acculturation. On the Bender, evidence of bias was infrequent and irregular. On the WISC-R, group differences were most discernible for age (as expected), sex, family structure, and race. Although these differences are noteworthy, consistent patterns of bias were not apparent among the 9 comparison groups. (10 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The WISC-R and internal evidence of test bias with minority groups.

Judgment of test bias can be made on the basis of a number of criteria. This study examines the evidence on the cultural bias of the WISC-R for Anglo-American, Black, and Mexican-American children. The performance of 1,050 children aged 5–11 yrs on the internal criteria of reliability and a number of order-of-item-difficulty measures was assessed. Minority group Ss responded to the test in the same general way as did Anglo-American Ss, and there was no clear pattern to the items on the test that were more difficult for minority Ss. Thus, the WISC-R appears to be nonbiased for minority group children. Explanations other than specific item content should be sought for the observed discrepancy between majority group and minority group IQ test performance. (27 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The real test bias: Influences of testing on teaching and learning.

Notes that there is evidence that tests influence teacher and student performance and that multiple-choice tests tend not to measure the more complex cognitive abilities. The more economical multiple-choice tests have nearly driven out other testing procedures that might be used in school evaluation. It is suggested that the greater cost of tests in other formats might be justified by their value for instruction (i.e., to encourage the teaching of higher level cognitive skills and to provide practice with feedback). (56 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Methodological considerations in validating computer-based test interpretations: Controlling for response bias.

Although computer-based test interpretation (CBTI) systems have been operational for nearly 25 years, their availability and adoption in routine clinical practice have grown exponentially in recent years. This article addresses methodological considerations in CBTI validation studies, emphasizing those design issues relevant to customer satisfaction studies. Specifically, issues of response bias are addressed as they relate to selection of raters and test respondents, use of random reports as a "control" for spurious ratings of report validity, and both the format and content of ratings. Deficiencies of various studies from the research literature are reviewed, and advantages and limitations of design alternatives are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

On reducing experimenter bias: The effects of forewarning.

Tested the hypothesis, with 67 male and 70 female undergraduates, that forewarning would reduce the effects of E bias. Es were given different expectations as to how Ss would perform on a person perception task. Each E then ran a number of Ss. 1/2 of each E's Ss were forewarned of the possibility of E bias and 1/2 were not forewarned. The hypothesis is supported among females (p     

Sex and test verification bias. Impact on the diagnostic value of exercise echocardiography

This trial represents the Egyptian experience in cyanoacrylate injection for hemostasis of bleeding gastric varices. One hundred patients with portal hypertension due to schistosomal hepatic fibrosis and/or posthepatitic liver cirrhosis were included. All patients presented with bleeding from gastric varices either fundal (80 patients) or inferior extension of esophageal varices (20 patients) were enrolled. Injection therapy was administered as the first active measure. No tamponade or drugs were used. Cyanoacrylate was mixed with lipid and injected through a hand-made probe. A mean of 3 (range 1-9) ampoules of cyanoacrylate were used per injection session. Bleeding stopped at the end of all sessions. Ten patients (12.5%) with fundal varices had rebleeding during the first 24 hours. Reinjection could control bleeding in 6 patients with a total success rate of 95%. Four patients were managed surgically. Fatal pulmonary embolism developed in one patient (1.25%) with fundal varix. Five more patients (6.25%) died from bleeding-related liver failure. In conclusion, injection of cyanoacrylate is highly satisfactory in controlling bleeding from both types of gastric varices.     

An analysis and evaluation of test and item bias in the prediction context.

Presents a perspective on test and item bias in predicting the performance of individuals in majority and minority groups. It is suggested that a difference between groups in the slope of the regression of the criterion on a test used for prediction represents bias that can frequently be corrected. Given a criterion with adequate measurement properties, the selection of items with essentially equivalent discrimination indices is expected to produce approximately parallel slopes. Accordingly, items having different discrimination indices are themselves biased. Intercept bias, on the other hand, is essentially not under the control of the test constructor. It is argued that the goal of producing a test without intercept bias is an unreasonable one. These problems are illustrated by factor models of intercept bias. It is concluded that relative differences in item-difficulty levels from group to group, given items with adequate discrimination indices, should not be labeled as bias. Such items are not automatic candidates for discard. (13 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Critical analysis of the statistical and ethical implications of various definitions of test bias.

Three mutually incompatible ethical positions in regard to the fair and unbiased use of psychological tests for different groups such as Blacks and Whites are defined, including unqualified and qualified individualism and quotas. Five statistical definitions of "test bias" are also reviewed and are related to the ethical judgments. Each definition is critically examined for its weaknesses on either technical or social grounds. It is argued that a statistical attempt to define "fair use" without recourse to substantive and causal analysis is doomed to failure. (PsycINFO Database Record (c) 2010 APA, all rights reserved)