The field of psychological assessment is called psychometrics, which is the scientific approach to the measurement of psychological characteristics. Psychometrics makes use of standardized tests, which are a set of tasks administered under controlled conditions and are used to assess someone's knowledge, skill, or other psychological characteristic. The field of psychometrics started in the 1800s as there became a need to identify the mentally ill or retarded so as to properly treat them. Psychological testing is now one of the largest aspects of applied psychology. Psychological tests now exist for almost any screening process. Psychologists attempt to make as pure a measurement of characteristics as possible, but some of these tests have become controversial because people suspect that they are biased in one way or another.
In general, psychological tests are used to meausure differences among individuals. Sir Francis Galton (1822-1911), an English scientist and mathematician, formulated much of the logic of measuring these differences. He believed that the differences between people were inherited, and that they could be represented by a normal distribution curve. The normal distribution curve is the bell-shaped curve you are probably used to seeing, and it is symmetric with respect to the mean score. Once the mean score was was found for each characteristic, it could be determined through psychological testing where someone stood relative to everyone else. Not all human characteristics follow this normal distribution, but psychological tests do all attempt to measure differences within a population.
The first step in creating a test is to write a detailed definition of what the test is attempting to measure. The authors must then come up with a large number of items which lend themselves to determining whatever the test is designed to measure. The items must then be tested on a trial group of people with similar characteristics to those who will eventually be taking the test. Through these pilot sessions, items are either dropped or kept on the test based on how well they discriminated between individuals. Norms , or standards for achievement on the test, are established based on how a large standardization sample does on the test. The standardization sample must be carefully selected to make sure that they are representative of the group that will eventually take the test.
No test will be absolutely consistent. Even physical measurements such as height and weight will vary from measurement to measurement, and psychological measurements are bound to be even less consistent. Nonetheless, it is necessary to make psychological tests as reliable as possible.
There are several ways to measure the reliability of a test. The first method is called test-retest reliability, and it involves giving the same test to the same people at different times. The closer the scores are for the individual at the different times, the more reliable the test is. The problem with test-retest reliability is that people may remember the test from taking it earlier and therefore give the same answers. This may inflate the reliability scores of the test.
Another method of testing reliability is called split-half reliability. The measurement of reliability compares the scores on two halves of the same test to measure reliability.
Yet another method of measuring the reliability of a test is called equivalent form reliability. In this method, two different forms of a test are administered to the same people, and the correlation between the scores on the two tests are calculated.
Still confused about reliability? Maybe this will help.
Validity refers to whether or not a test is measuring the real-world characteristic that it is supposed to be measuring. It is the true measure of how useful a test is.
One way of measuring the validity of a test is called criterion-related validity. This measurement compares a test's results to the individual's behavior in the real world. There are two types of criterion related validity. Predictive validity refers to a test's ability to predict an individual's future performance in a given area. Concurrent validity, on the other hand, refers to how well a test matches up with an individual's current performance in an area of interest.
Another type of validity, content-related validity , refers to how well a test covers the material that it is supposed to cover. For example, the content-related validity would tell you whether a final exam asks questions about the whole year, or if it just concentrates on a few specific items from the course.
The face validity of a test is perhaps the simplest measure of a test's validity. It refers to whether or not the test's material has any relevance to what the test is supposed to measure.
The true measure of a test's validity, however, is the test's construct validity. This type of validity refers to how well a test's results correspond to the individual's actual characteristics in the area being measured. The construct validity of a test is difficult to determine, but it can be done by comparing a test's results with that of other tests, or by comparing an individual's scores before and after some sort of training designed to alter the characteristic being tested.
Still confused about validity? Try this site.
One simple way of increasing a test's reliability and validity is simply by increasing the number of items on the test. This method only works so well, however, so other methods must be tried. One of these other methods is item analysis. In item analysis, the scores for each item on the test are correlated with the scores for the test as a whole. If the correlation is low for a specific item, it may be dropped from the test.
Most people can agree that intelligence varies between individuals, and that there exists a continuum of intellectual abilities that can be roughly ordered. Determining this continuum, however, is extremely difficult. In fact, simply defining intelligence is hard to do. In 1923, psychologist E.G. Boring defined intelligence to be "what the tests test." This definition was influential for a time, but we will define intelligence to be the capacity to acquire and use knowledge. Because intelligence is a capacity, it cannot be measured directly. It can, however, be measured indirectly through testing what knowledge an individual has already acquired, as well as by testing how well an individual can use that knowledge. Of course, different people have had different opportunities to acquire knowledge, which is where much of the controversy in this field comes in.
The earliest intelligence theories were predominantly ability theories. This view is often referred to as the psychometric view, because it is heavily based on measurement issues. These abilities theories have been devided on whether intelligence is one general capacity, or a number of seperate capacities.
Charles Spearman was a British psychologist who argued that intelligence is based on one factor known as general intelligence, or simply g. There are, however, specific abilities, or s, which each require a certain amount of the g factor. According to Spearman's theory, intelligence tests should measure intelligence without becoming clouded by specific abilities. The best such test is Raven's Progressive Matrices, which consists of a series of abstract patterns with one piece missing. The subject must pick out the missing piece in a multiple choice format.
L.L. Thurstone, among others, argues that intelligence does not depend just on one factor, such as Spearman's g, but rather on a number of seperate factors which make up intelligence as a whole. Thurstone thought that intelligence depended on six or seven primary mental abilities. One person may be exceptional in one of these areas while being quite poor in another.
Like Thurstone, Howard Gardner argues for several autonomous factors which make up human intelligence. His theory of multiple intelligences argues for between six and twelve seperate intelligences.
J.P. Guilford's model of intelligence has a very large number of intellectual abilities. His model depends on three factors: operations, which refer to what an individual does; contents, which refer to the material on which the individual performs these operations; and products, which are the basic forms in which the information can be fit. His model of intelligence is shaped like a cube, with these three factors making up the cube's height, width, and depth. The cube contains 150 smaller cubes (5 operations X 6 products X 5 contents), each of which refer to one type of intelligence. According to Guilford's model, each of these 150 intelligences can and should be measured individually to determine a person's intelligence.
Many of the ability models of intelligence rely on a process called factor analysis, in which many different tasks are intercorrelated. Patterns of items that tend to go together can be isolated as factors.
Still confused about ability theories of intelligence?
As cognitive psychology gains more and more influence, so too soes the idea of cognitive competencies. Theorists have looked to elementary cognitive processes such as short-term memory scanning, long-term memory retrieval, or the speed with which information is scanned and encoded as the key to intelligence. One of the most heavily studied aspects of this theory the relation between reaction time and intelligence. There seems to be a fairly strong correlation between these two measurements.
Sternberg's triarchic theory of intelligence includes three parts. The contextual dimension relates intelligence to the external world in which an individual lives. For example, intelligence involving working with computers would not have been a factor several years ago, but it is now a very important aspect of intelligence. The second dimension of Sternberg's theory is the experiential dimension. This dimension of intelligence deals with the ability to use past experiences to gain insight into new situations. The third aspect is the componential dimension, which is fairly similar to the cognitive theories of intelligence described above. It refers to the the set of mental mechanisms which dictate intelligent behavior.
Still confused about Sternberg's theories of intelligence?
Originally, all intelligence tests were individual tests; meaning that they were given in a one-to-one situation. During World War I, however, group tests started to appear as a mass testing program was developed. Currently, individual tests are administered when evaluating individuals who are suspected of being either gifted or retarded. Group tests are used in other situations such as in the education system and in military programs. Because of the close contact with the examiner involved in individual tests, they may be more accurate because the examiner is more likely to be able to determine if the subject is having a bad day. The two most popular individual tests are the Stanford-Binet and the Wechsler scales.
At the beginning of this century, Alfred Binet was the Minister of Public Instruction in Paris, France, and he and his colleagues began designing an intelligence test in an effort to study procedures for the education of mentally retarded children. The test emphasized verbal skills and was arranged in order of difficulty. Items were grouped by level so that 80-90% of any given age could pass that age's level. The person's score on the test could therefore be expressed as a mental age, a score which corresponds to the average age of normal children who perform as well as the individual being scored. The basis for Binet's evaluation was the assumption that people who performed below their age level were retarded, thpse who performed at it were normal, and those who performed above it were gifted. In 1916, Lewis Terman revised Binet's test for use in the United States while at Stanford University. The new test became known as the Stanford-Binet Intelligence Scale. It was an individually administered test, and the first to use the idea of an intelligence quotient, or IQ, which could be derived from a ratio between a person's mental age (MA) and chronological age (CA). The Stanford Binet now computes a score known as the standard age score, hich uses Galton's ideas about the normal distribution of human characteristics which can be seen in further detail here.
The Stanford Binet Intelligence Scale has many benefits, but also some problems. It is good because it has been around so long that it has become very well standardized. It also allows the administrator of the test to have close, personal contact with the subject, which allows the examiner to makes inferences about the subject not related to intelligence. The test is also one of the very few which can be administered to someone several times at a variety of ages. The problem with the test is that it may try to do too much. It is extremely difficult to come up with a test that works for so many ages, and it may not adequately measure intelligence at certain ages. It also has not worked well for identifying people who are gifted when its results are compared to those of other tests.
The other major individual tests are called the deviation IQ, which is based on the score in relation to the normal distribution curve of intelligence. It has some advantages over the Stanford Binet Test because it can target specific age ranges, and contains subtests which can be analyzed for patterns in a person's scores.
Group tests are much simpler to administer than group tests and can therefore be given by people who are not highly trained. They can also be given to as many people as can fit in a room, so they can test large numbers of people more efficiently. They are also able to be scored very quickly because they are usually multiple choice.
Aptitude tests are tests which are designed to test how much someone would benefit from being admitted into an educational program or course of study. Therefore, they are generally concerned with the subject's ability rather than the subject's prior knowledge. Achievement tests, on the other hand, are tests which evaluate how well someone has learned the material from a given course.
The SAT's are one of the most widely used and known aptitude tests today. Millions have participated since the test was developed in the 1920's. The purpose is not only to measure intelligence but to measure aptitude for college level work. The test is broken down into two sections, verbal and mathematics, each of which have a mean score of 500 and a standard deviation of 100. The reference sample for this standardization was a group of 10,000 students in 1941, meaning from year to year the actual mean and standard deviation may vary. Lower scores in the early 1990's (verbal mean 422 and mathematics mean 474 in 1991) resulted from a broadening of the population of student taking the test and changes in the home and school environment. Many specifically blame the switch from casual reading to television viewing for the verbal drop.
Many criticize the multiple choice format of the SAT. This takes the focus off critical analysis and true aptitude and back to just test taking skill. Test show multiple choice is not a true indicator of skill in the area being tested. Students taking the reading comprehension without doing the selected readings scored higher than they would simply by guessing, although lower than the sample who read the material, in a 1990 experiment
In 1994, partly as a result of these criticisms, the college board made several changes. The test was split into SAT I: reasoning tests and SAT II: subject tests. The SAT I now contains some short answer in addition to multiple choice, and the SAT II focuses on different areas of aptitude with more comprehensive testing.
Sometimes it is important to measure how fast an individual can perform a certain task. For this purpose speed tests are used. Speed tests consist of problems designed to be very easy for the test taker, but with far too little time to complete all of them. Power tests, on the other hand, are used to determine mastery of the material. Questions range from very easy to very difficult with plenty of time to focus on every problem. If any test taker can finish all the problems in either test, the examiner would not know how much more they could have accomplished had the test been beyond their limits.
Most standardized tests include a sampling of both types of testing. Many tests have both varying levels of difficulty and a time limit. Often the time limit is not actually a factor as all students finish well within the time limits. Choosing a test depend heavily on the purpose of the examination. If testing for the best prospects for college admission, one would choose a standardized group aptitude test, probably with both speed and power elements. To try to pick out gifted students one would prefer a group achievement test. If speed of performance was not relevant it would be a pure power test.
One of the major rationale for the development of intelligence tests was to assess which students needed special educational assistance, such as those with mental retardation. In 1978 Public Law 94-142 went into effect in the United States that altered the way public schools treat children with mental retardation.
The American Association on Mental Retardation defines mental retardation in terms of limits in adaptation skills as well as in terms of tested intelligence. An individual is usually classified as mentally retarded if he or she scores below a 70 to 75 on either the Stanford Binet or Wechsler Series tests. People with moderate to severe mental retardation usually also have other disabilities, which makes this segment of the population a particular challenge for the education system. Before Public Law (PL) 94-142, very few children with mental retardation received an education. PL 94-142 now mandates that all children, regardless of handicapping condition, must be included in the education system. It also says that children cannot be classified as mentally retarded based on one condition, such as an IQ score, and that assessments of mental retardation must be made by a licensed proffesional. PL 94-142 has made a tremendous difference in the lives of the mentally retarded.
Self-report inventories are the most frequently used type of personality inventory. In self-report inventories, individuals answer a series of questions about themselves. One of the reasons that this assessment technique is so popular is because of the belief that people know themselves better than anyone else.
The Minnesota Multiphasic Personality Inventory (MMPI-2) is
the most widely used psychological test. It consists of more than 500 questions about the
individual which must be answered with true, false, or cannot say. It is an Try here for more information on the MMPI-2.
There are a number of other self-report inventories which are used for a variety of purposes.
These other tests include personality tests for normal individuals, such as the California
Psychological Inventory; sex-role inventories, such as the Bem Sex-Role Inventory; value scales,
such as the Allport-Vernon-Lindzey Study of Values; and many more.
One of the other self-report inventories that is gaining popularity is the NEO-PI (neuroticism,
extraversion, and openness personality inventory). The authors of this test take a strong
stand on the universal nature of personality. It is becoming a popular personality test for
normal populations, and it can now be taken in self-report form or in third-person form, where
people can attempt to evaluate others' personalities.
Self-report inventories are very easy to give, and provide a quick evaluation of personality.
It is so easy to use, in fact, that it has become somewhat overused. More and more employers are
using personality profiles in the hiring process, which is not what the tests are meant to do.
In self-report inventories, it is often obvious which answers are more socially desirable than
others. Because of this, self-reports may become self-presentations, which is not what is meant
to be assessed. Some tests such as the MMPI-2, therefore, contain lie scales which are
meant to correct for people trying to make themselves look either good or bad instead of answering
truthfully. The lie scale consists of questions which are designed to be answered in only one way
if a person is telling the truth. Such a question might be "Have you ever told a lie?" to which
everyone should answer "yes." From the lie scale, a correction scale can then be
determined, which will help to correct for people who are attempting to lie on their personality
profile.
It is questionable whether or not this type of correction factor actually works, but
their is some evidence which suggests that it does. In an experiment done in a medium-security
prison, prisoners were told to attempt to fake insanity on a personality profile, and individuals
from a mental hospital deemed insane took the same test. 80% of the fakers were identified, while
none of the truly insane patients were misidentified. It is therefore clear that while it is
possible to fake insanity on this profile, the odds are against a person succeeding if a lie scale
is used.
Thematic Apperception Test (TAT)
Other Self-Report Inventories
Strengths and Weaknesses of Self-Report Inventories
Faking a Personality Profile
Projective Techniques
Inkblot Techniques
The Thematic Apperception Test
Strengths and Weaknesses of Projective Techniques
Behavior Assessment
Social Issues in Psychological Assessment
Examiner Familiarity
Characteristics of the Examinee
Test Anxiety
Age
Sex
A Nature-Nurture Issue
Research on Test Bias
Psychological Assessment in Different Cultures
