IQ Testing: Reliability, Validity, and Scores

An IQ test, as most of you know, is an intelligence test. IQ stands for Intelligence Quotient which is the calculation of one’s score on an IQ test as compared with other test takers in the same age-group.

There are several different types of intelligence tests, two of the most popular being the Stanford-Binet intelligence test, and the Wescheler intelligence test. In fact, if you’ve ever encountered an “official” IQ test, chances are that you took one of these two tests.

These sorts of tests are used in a wide range of areas including colleges, high schools, the military, and the government. Each of these assessments is meant to either measure intelligence, or provide predictive information on future performance in order to best place the individual. In order for an intelligence test to be used in an official manner, meaning in a school, government office, or likewise, the tests have to meet both reliability and validity requirements.

The reliability requirement refers to the consistency of a person’s score (Romero, Kemp, 2007). For example, you wouldn’t think very highly of a test that you scored a 90 on the first time, a 113 the next time, and 50 on another separate occasion. A test that yielded these sorts of results from the same person on different takes would be considered unreliable because there is obviously something wrong with the test, and not the person taking it. For a test to be considered reliable it must past a reliability test. At the moment there are several different reliability tests that can be conducted in order to assess an IQ test’s reliability. A test creator can assess the IQ test’s reliability through a test-rest method in which an identical test is administered on two separate occasions. The second method is called the split-half test in which the test is split in half and each half is issued on a separate occasion. The third form of reliability testing is the alternate-form which requires test writers to create two different, yet similar versions of the test. Questions in this form of reliability assessment are based on the same concepts, yet worded in a manner that would make the question seem new or unrecognizable to a test taker. As with the other forms of reliability testing the two forms of the test are given on different occasions, yet administered to the same person in order to best assess the consistency of the scores obtained.

The second requirement for an IQ test is validity. Validity is the measure of whether a test actually tests what it is supposed to. A reliable test doesn’t mean it is a valid one. It doesn’t mean that the test is actually testing the knowledge or skills necessary to rank highly on an IQ test. For this reason it is also important that a test undergo validity testing. As with reliability testing, there are several testing methods involved in validity testing. Depending on the type of test, test makers and/or psychologists will make the decision as to which validity tests are required to prove the test valid and eligible for administration on a public scale. These three validity tests are criterion validity, content validity, and predictive validity. The first of the three, criterion validity is where the test is measured against some yardstick (Romero, Kemp, 2007). For example, you can measure some tests against similar tests that have passed the validity assessment. Criterion validity plays a role in the assessment of tests that are looking to test specific traits or abilities. Content validity is concerned with the assessment of a test’s material; of whether or not the test covers all the material it is supposed to. Lastly, predictive validity deals with the “test’s ability to predict future performance” (Romero, Kemp, 2007). Such tests as those we take before entering high school, or college are predictive tests, they are meant to assess our future performance.

So now that you know what is required before a test even makes it onto the examination table of a school, college, or government office, how exactly do IQ tests measure up to this? At this point it can be said that no psychological test is perfect. No test can score a 100 on both reliability and validity. However, IQ tests that have made it through testing have a high rating for both reliability and validity. I previously mentioned the Wescheler test otherwise known as WAIS. The WAIS scored a .95+ on split half reliability, a very strong rating. Additionally, the University of Canberra notes that the WAIS is comparable to in validity to the Stanford-Binet, another prominent IQ test (2004). The University of Canberra also notes that evidence supports that the WAIS is a valid measure of global intelligence (2004). For both the WAIS and the Stanford-Binet to which is it compared, this is good news and evidence of good tests.

As mentioned earlier, you’ve probably taken some form of IQ test before, however, the sorts of questions that are asked on an IQ test will be reviewed nonetheless. Typically, an IQ test consists of multiple choice questions so as to eliminate bias among the answers, meaning there is no debate over the right or wrong answer. Additionally, most IQ tests consist of a number of different testing areas. There are sections for verbal knowledge, perception, and spatial knowledge. If the IQ test is given in these sorts of subsets, an individual score will be given for each section, along with the overall total score of the test. At this point, IQ tests are graded with a mean score of 100. This puts all tests on the same level. This means that the average person has an IQ of 100. Obviously those below are below average, while those above are above average.

Scores are further broken down in terms of understanding the meaning of the scores by standard deviation of 15. This 15 standard deviation puts 50% of people who take the test at a score of 90-110, and 70% of people between 85-115. Scoring between 90 and 109 is considered normal intelligence, while a score of 70 is bordering on mental deficiency and 120-140 is considered highly intelligent.

This should give you a better understanding of IQ tests, their purpose, their reliability and validity, and what IQ scores actually mean.

University of Canberra. (2004). Intelligence and Psychological Testing. Retrieved June 2, 2007, from University of Canberra Website: http://www.wilderdom.com/personality/intelligence.html

Romero, A, & Kemp, S (2007). PsychologyDemystified.New York City: McGraw Hill.