Originally a standardized test was simply a standard test – of academic achievement or of knowledge in a specific academic or vocational domain. It has since acquired the meaning of a written test whose scores are interpreted by reference to the scores of a norm group which has taken the test and which is usually considered to be representative of the population which takes the test. For example, standardized tests of academic achievement provide conversion tables showing the percentile ranks in the norm group of all possible raw scores. Some standardized tests are now analyzed with item response theory.
Criticisms of standardized tests
Standardized tests are widely used in education, placement, and certification. Their validity has been criticized on several grounds.
Some of the criticisms are standard psychometric ones. For example, scores on tests of achievement in mathematics problem-solving are often correlated with scores on tests of language ability; this suggests that the mathematics test is actually measuring the linguistic ability required to understand the presentation of the problems rather than the mathematical ability required to solve them. Educational tests also tend to become outdated as curriculum changes.
Standardized tests are also widely criticized as culturally inappropriate for many groups, both in content and in process. Criticism of content usually centers on the differing relevance of the content to people from different cultures – for example, newly arrived immigrants can be expected to have greater difficulty with an intelligence test which asks them to name past leaders of the country to which they have recently immigrated.
Attempts have been made to develop culture-free and culture-fair (culture-neutral) tests of intelligence, but on the whole these attempts have not been successful. Conceptions of intelligence vary widely from culture to culture, and abstracting the few common elements, or what appear to be the few common elements, cannot be depended on to produce a reliable guide to intelligence.
A common criticism of standardized testing programs in schools is that they encourage teachers to "teach to the test." That is, teachers concentrate on the parts of the curriculum they know will be covered on the test and neglect those that will not. This criticism is certainly worth considering if teachers have foreknowledge of the test and the test is not comprehensive. However, if enough alternative forms of the test are provided, if teachers do not know which form will be used, and if the forms provide a comprehensive sampling of the curriculum, this danger would probably be avoided. Despite the obvious danger of teaching to the test in certain circumstances, though, little research has investigated the prevalence of the phenomenon, or its effects. Furthermore, any form of testing will promote teaching to the test if the consequences of testing are serious.
A related criticism is that students whose teachers train them in test-taking skills unrelated to content will perform better than equally accomplished students whose teachers do not. Some simple test-taking skills can improve scores on multiple-choice standardized tests, so this criticism points to a real danger, especially if standardized tests are used (incorrectly) as the sole measures of achievement or skill. However, little research has investigated the prevalence or effects of this training.
Standardized tests are also criticized for emphasizing recall and recognition rather than higher-order cognitive skills. However, this criticism is not generally valid. While many standardized tests do emphasize recall and recognition, many others assess analytical skills.
Another criticism is that standardized tests assess inadequate samples of skills. Again, however, this criticism cannot validly be made of all standardized tests, although it can be made about the majority of tests of any type.
Large-scale attempts have been made to substitute performance assessment or "authentic" testing for standardized academic testing. Performance tests require actual performance of a skill; for example, instead of answering questions about a science experiment a student would be required to perform it. However, performance tests have poor reliability simply because they accumulate so little data. Standardized tests have been found to predict scores on performance tests better than other performance tests do.
Much of the opposition to standardized tests has centred on the incorrect use of these tests. In particular, the use of standardized tests of academic achievement to assess individual students is questionable, given the tests' reliability – they are simply not accurate enough to provide adequate assessments of individual students by themselves.
Perhaps the most important criticism of standardized testing is that many standardized tests fail to meet the standards of their own field. For example, tests of adult literacy are widely used although there is little evidence that they assess literacy accurately.
Advantages of standardized tests
Perhaps the simplest advantage of standardized tests is that they are standardized. While some people may systematically score lower on certain tests, these differences will be systematic. On the opposite end of the spectrum, scores on subjective tests change significantly according to whoever is grading them. In the case of college admissions, for example, interviews with prospective students has been repeatedly shown to predict later college performance no better than chance, while statistical measures such as prior GPA or SAT scores are much more accurate.
One of the main advantages of standardized testing is that it is able to provide assessments that are psychometrically valid and reliable, and well as results which are generalizable and replicable.
Another advantage is aggregation. A well designed standardized test provides an assessment of an individual's mastery of a domain of knowledge or skill which at some level of aggregation will provide useful information. That is, while individual assessments may not be accurate enough for practical purposes, the mean scores of classes, schools, branches of a company, or other groups may well provide useful information because of the reduction of error accomplished by increasing the sample size.
While standardized tests are often criticized as unfair, the psychometric standards applied in the development of standardized tests would produce fairer testing if applied in other types of testing. In particular, the effectiveness of each test item in accomplishing the goal of the test would have to be demonstrated.