The article intends to review the merits of the cultural - test –bias hypothesis. The hypothesis claims that the differences observed in mental tests are not to be taken with confidence. The hypothesis states that the racial or ethnic differences are a result of biases integrated in the mental tests. These biases have developed over the years through wrong and irresponsible psychometric methodology. If at all they exist and the hypothesis is proven true, the implications would be huge. Practically, there would the unjust denial of placement in educational opportunities. There would also be the unfair denial of employment opportunities. On the scientific front, the repercussions would vibrate through the very foundation of the discipline. It would render 100 years of psychological research useless and rudimentary. It would through applied psychology into a quagmire.
In resolving this issue of bias, the legislatures and courts have taken center stage in recent years. The state of New York has brought into law “truth in testing” laws and is considering complimentary bills. The courts have also been involved in answering the question “Are intelligence tests used to diagnose mental retardation biased against cultural and ethnic minorities?”. California in 1979 said “yes” while just one year later, the state of Illinois said “no”. These conflicting judgments by two courts of equivalent standing are confusing. It only adds to the chaos and confusion in determining the legality and existence of bias.
The criticisms directed to aptitude and intelligence tests are well documented. Points of criticism include the tests inability to predict the future behaviors of minority children. Another point of criticism is the tests evaluate different attributes when applied to children who are not from the white, middle – class culture. Due to this inadequacy of the tests, race and ethnicity are used as moderators. Researchers revert to evaluating the test performance of minority children on their race or ethnic background. An example where this is used is in the case of Larry.P and PASE. The plaintiff witnesses intimated that it is not likely that a white middle – class child would respond to a smaller child who starts a fight by fighting back. The white middle – class child would respond by looking for other solutions. The witnesses claimed that a black child must respond through violence. The cause of this is that other options like mediation are not applicable in the ghetto setting. This leads researchers in interpreting differently the behavior and test results of the two groups. The results of a test should not be determined by external characteristics of the test subjects. The results must be dictated wholly by empirical data.
Bias in Construction Validity of Intelligence Tests
The validity of amplitude and intelligence tests can be accurately determined through several methods. The first step in this determination is defining bias in construct validity. Bias exists where one test is shown to evaluate different hypothetical traits for one group than for a separate group. One of the most popular methods of determining validity is called factor analysis. This involves comparing the factor – analytic results of aptitude and intelligence tests across different races. One of the more vocal proponents of this empirical approach is Hillard. He states that the IQ test can be used to establish construct validity of the test results. The factors that come from a given test should be similar and the same from one population or race or ethnic group to another. Many studies that have been done indicate a consistency of these factors among different races and ethnic groups. No concrete evidence has been tabled on the bias in construct validity.
Bias in Criterion – Related Validity of Intelligence Tests
Under this empirical approach, the definition of bias is different and involves predictive validity. Bias is considered to be present when the conclusion drawn from a test is not made with the smallest random error. Bias is also observed if there is a constant error in a conclusion in relation to a particular group. In this approach, the regression line is the guiding factor. Where the line differs significantly between groups, there is definitely bias. In a situation where there is no bias, a condition known as homogeneity of regression is evident across groups and races. There is however no concrete evidence to support a consistent pattern of bias in predictive validity. Bias happens with no particular pattern and in an unpredictable manner.
Bias in Personality Scales
There is not much research done in this area of bias. The author has taken the initiative himself and a few friends to investigate this aspect of bias. The outcome of their research fails to provide strong evidence to indicate bias under personality scale. It is pretty much similar to the findings in the constructive and predictive validity. The data is promising and enticing. However, as it is still in the preliminary stages, the current data takes precedence.
In conclusion, the article reiterates that up to now there lack conclusive evidence to suggest the presence of bias in the results of intelligence and aptitude tests. The article cautions that this position may change soon. With the development of new techniques and better methodology, they will shed new light into the subject. Relevant actors in the discipline are encouraged to keep abreast with the emerging issues in the subject. The article recommends that attention should be given to the realities on the ground and the potential consequences of bias in the test. The author advices us to be skeptical while approaching the available data and information on the topic.