Bias and Ethical Reporting of ResultsIn the debate on the costs and benefits of using test scores to render decisions, opponents of tests have pointed out the intrinsic and extrinsic biases inherent i

Intrinsic and Extrinsic Bias PSY3700 Multimedia Assessment and Psychometrics ©20 16 South University 2 Intrinsic and Extrinsic Bias Test Item Bias Intrinsic test bias exists when significant mean differences between groups are not related to differences in the construct or trait under investigation but due to a test's characteristics. Known as measurement invariance, this essential property of a test is missi ng when group differences are due to test bias or when a test is deemed to be more valid for one group than for another (Rust & Golombok, 2009). Differential content validity and predictive validity for groups can contribute to the overall intrinsic bias of a test. An example of differential content validity is that test items reflect the values of mainstream society and so are not good indicators of the intelligence or personalities of members of minority groups.

In this way, cultural bias may contribute to the differential content validity of a test, which is a form of intrinsic bias. Likewise, differential predictive validity can be reflected in test scores that are more valid for one group than for another. An example of this is the predictive validity of the SAT. Although the SAT scores on the math and verbal subtests have respectable predictive validity for Caucasian students, only the verbal subtest score is predictive of the future college performance of African -American students (Groth -Marnat, 2003) . To reduce intrinsic bias, the use of several statistical models shows promising results. One model uses adjustment for intrinsic bias by using separate regression equations and different cutoff scores for groups. Another is the use of quotas. Although b oth have been used to make selection decisions, neither model has been able to overcome the criticism that group equity is achieved at the expense of the individual. That is, many individuals falling between the cracks found in the range between the cutoff scores of groups are compromised for the sake of balancing inequities in the process of selection (Rust & Golombok, 2009). When decisions based on test results lead to inequities though the test does not contain bias, there is extrinsic bias. An example of extrinsic test bias is the effect of inequities in opportunities afforded by social environments that discriminate against members of minority groups (Rust & Golombok, 2009). As such, future efforts should focus more on changing societal inequities that deny equal access to education and career opportunities to minority individuals (Groth -Marnat, 2003). Without changing the underlying social fabric, the use of test scores to draw inferences about one's intelligence and personality will lead to social dep rivation (Rust & Golombok, 2009) and lost opportunity for many individuals. Because of the ill effects of extrinsic bias, legislation has handed down the ruling that a test developer or user has the burden of proof to support the validity of the measure t o render decisions. Further, in many US states, the use of Intelligence quotient (IQ) test scores to select school children for remedial classes has been outlawed (Rust & Golombok, 2009). On the other hand, in many states, the use of IQ test scores to qual ify students to receive special education services is still commonplace. Consequently, the debate over the costs versus benefits of using tests to make decisions in academic and occupational settings is likely to continue. The social context that affects t he usefulness of testing is marked by inherent inequities that are not easily resolved. PSY3700 Multimedia Assessment and Psychometrics ©20 16 South University 3 Intrinsic and Extrinsic Bias Test Item Bias References Groth -Marnat, G. (2003). Handbook of psychological assessment (4th ed.). Hoboken, NJ: John Wiley & Sons. Rust, J., & Golombok, S. (2009). Modern psychometrics: The science of psychological assessment (3rd ed.). New York, NY: Taylor & Francis. © 201 6 South University