6
153 ^
SEP 2 3 1976
A Social Service Field Guide to
Psychological Testing
MORTON L. ARKAVA, Ph.D.
"""" '3 0864 00025397 4
A SOCIAL SERVICE FIELD GUIDE TO
PSYCHOLOGICAL TESTING
By
Morton L. Arkava, Ph.D.
Professor and Chairman, Departnnent of Social Work
University of Montana
1974
Published by the Governor's Crime Control Commission
Departnnent of Institutions
State of Montana
(Under the Provisions of Sub-Grant #736147)
TABLE OF CONTENTS
Page
PREFACE 1
CHAPTER I - PURPOSES OF TESTING 3
What is a Test? 3
The Uses of Tests 4
Institutional decisions 5
Individual decisions 6
Misuses of tests 6
NOTES - CHAPTER I 8
CHAPTER II - CLASSIFICATION OF TESTS 9
Intelligence Tests 11
Aptitude Tests 12
Achievement Tests 13
Personality and Interest Tests 14
Personality tests 14
Interest tests 16
Specific Diagnostic Tests 18
NOTES - CHAPTER II 19
CHAPTER III - BASIC TEST CONCEPTS 20
Reliability 20
- i -
TABLE OF CONTENTS (continued)
Page
CHAPTER III (continued)
Factors afFecting reliability 21
Determining reliability 22
Validity 23
CHAPTER IV - BASIC STATISTICAL. CONCEPTS 26
Norms 26
Measures of Position 30
Measures of Central Tendency 31
Mean 31
Median 32
Mode 32
Measures of Variability 32
Range 32
The semi-interquartile range 33
Standard deviation 33
Measures of Correlation 34
TABLE I - COMPARISON OF SOME STANDARD
SCORES 35
Inferential Statistics 36
Raw and Standard Scores 37
Ratio Scores and Placement Scores 40
NOTES - CHAPTER IV 42
- II -
TABLE OF CONTENTS (continued)
Page
CHAPTER V - LIMITATIONS OF TESTS 43
Supplennentary Measures 43
Test Construction Limits 43
Effects of Culture 45
Other Limitations 47
NOTES - CHAPTER V 48
CHAPTER VI - HOW TO MAKE A TEST REFERRAL 49
Suggested Guide for Test Referrals 49
Some Hints for Dealing with Psychologists 50
CHAPTER VII - SOME COMMONLY USED TESTS 52
Differential Aptitude Test (DAT) 52
Goodenough-H arris Drawing Test (Draw-A-Man Test) 53
Other Drawing Tests 53
Minnesota Multiphasic Personality Inventory (MMPI) 58
Otis Self-Administrating Test of Mental Ability 63
General Aptitude Test Battery (GATB) 64
Strong Vocational Interest Blank (SVIB) 68
Stanford-Binet Scale 71
Vineland Social Maturity Scale 73
Thematic Apperception Test (TAT) 75
Symonds Picture Story Test (SPST) 78
Wechsler Intelligence Scale for Children (WISC) 78
- iii -
TABLE OF CONTENTS (continued)
Page
CHAPTER VII (conttnued)
Wide Range Achievennent Test 81
Bender-Gestalt 84
Rorschach 87
Wechsler Pre-School and Primary Scale of
Intelligence (WPPSI) 90
Peabody Picture Vocabulary Test 93
Wechsler Adult Intelligence Scale (WAIS) 96
Tests for Special Purposes 101
The Culture Fair Intelligence Test 101
Tests for the orthopedic handicapped 102
Tests for the hearing handicapped 1 03
Tests for the blind 1 03
CHAPTER VIII - HOW TO LEARN ABOUT SPECIFIC TESTS 104
- IV -
PREFACE
The production and consunnption of educational and psychological
tests have increased dramatically since their developnnent over fifty
years ago. Greater numbers of tests are being used to both evaluate
and guide individuals and to aid in administrative decisions. Workers
in the social services today will confront the results of various kinds
of psychological and educational tests, since most case histories in
use contain test information.
Because many persons currently practicing in the social services
lack basic educational preparation in test use, they tend to misuse and
underutilize test information. It is imperative that persons engaged
in the delivery of human services understand some sinnple test concepts
for use in effective case management.
This guide was developed to help the social service practitioner
utilize test results more rationally and consistently . It is not intended
to serve as a comprehensive textbook on psychological test administra-
tion, interpretation, or utilization, but rather to serve as a basic guide
to those persons who have little or no background in the use and
- 1 -
interpretation of psychological and educational tests. For those who have
had graduate-level coursework in psychological testing or extensive in-
service training, more advanced texts on the subject are advised.
CHAPTER I
PURPOSES OF TESTING
What is a Test?
Cronbach, a noted authorit/ on testing, has defined a test as
"... a systematic procedure for comparing the behavior of two or
more individuals."^ Others have defined tests as standardized procedures
for obtaining a sample of an individual's behavior. Psychologists and
others use tests in order to predict what a person might do or to dis-
cover what he could do. Similarly, tests may reveal why a person does
certain things. While undoubtedly the answers to these questions would
be more accurate if the individuals involved could be observed over a long
period of time, this is generally not practical in clinical or industrial
settings. Therefore, one must rely on the brief samples of behavior
provided by tests. Thus, a test involves both a sample of behavior and a
procedure for comparing that behavior with the results obtained by others .
The accuracy of the predictions, from test behavior to actual life
behavior, depends upon many factors, including the nature and con-
struction of the test — especially with respect to the concepts of validity
- 3 -
and reliability, the conditions under which the tests are given, and the
clinical and social sophistication of the examiners.
The concept of validity refers to the degree to which a test instru-
ment actually nneasures or predicts specific behavior. For exannple, if
intelligence is of spjecific interest, a desirable test instrument is one
which will give an accurate measure of the concept of intelligence as
currently defined and accepted . The concept of reliability refers to the
consistency of test results over repeated testings: how closely will an
individual's test score, or a test score on an alternate version of the
same test, approxinnate the score he obtains on earlier or later testings
with the same instrument? From a statistical point of view, reliability
is a necessary condition for validity. These concepts will be explored
later in greater detail.
The Uses of Tests
The purposes of tests are many, but generally tests are used to
provide information for decision nnaking. Cronbach has suggested that
"... the value of test information should be judged by how nnuch it
improves decisions over the best possible decisions made without the
2
test." He p>oints out that if one desires to predict school grades, the
information obtained fronn a scholastic aptitude test will not provide any
greater accuracy than previous school grades. Generally speaking, then,
tests provide information that is not easily obtained otherwise.
Test information assists in nnaking two different types of decisions:
institutional and individual . The purposes of decision making create the
major difference between these two categories.
Institutional decisions. Institutions make decisions according to
their goals in contrast to the goals and wishes of the individual. Decisions
are made from the perspective of the operation and maintenance of that
institution. Forexannple, school personnel typically make institutional
decisions concerning possible admission of students into college or special
training programs. Test results usually affect those decisions. Similarly,
the Army uses tests to assess special aptitudes and skills and to place
personnel in special assignments or training programs .
Industries, schools, and social service agencies typically make
institutional decisions in similar ways. One such agency, a parole board,
often needs to decide upon the possible release of a prisoner. To the
extent that parole boards attempt to predict the offender's behavior and to
select only those who exhibit the least potential for antisocial behavior, they
engage in institutional decision making. In a similar way, probation
officers and judges attempt to select good risks for probation. A "good
risk" is defined as someone whose potential for repeating his offense is
thought to be nninimal . In such cases, information is not always complete
enough to make an intelligent decision. A test's value lies in its potential
to provide greater accuracy in decision making over the best possible
decisions nnade without test information.
Individual decisions. Individual decisions pertain to unique and
personal conditions. They are those a person nnakes about some aspect
of his or her own life: the determination of a career, whether or not to
enter a special training program or to go to college, the selection of a
potential mate. In the social services, individual decisions may be made
from the perspective of the person involved. Under certain conditions, for
example, the social worker will make a decision for the client.
There are several ways test information can be utilized in individual
decision making. Vocational and aptitude tests are commonly used to help
people make career choices. Numbers of young people often wonder what
career best suits them or offers them the best chance of success. Fre-
quently, vocational interest and aptitude batteries allow them focus on
areas of interest with high success potential.
Misuses of tests. Many counselors tend to rely too heavily on test
information . As a result they may seriously linnit the options available
to an individual . The writer has observed a number of persons who
sought college preparation in social work simply because their high
school counselor told them they had a high score in this area on a voca-
tional interest test. One student decided to enter the field because of
her high scor^ on a vocational interest battery and social work was the
specific field nnentioned. Upon close examination, however, her interests
and aptitudes did not support her test score, personal commitments, or
her common sense. The writer has also talked with students who
decided not to pursue paKicular programs in higher education because they
had Tow test scores, despite the fact that in at least two such instances
both students had achieved very commendable previous records of
academic success .
Research has demonstrated that success in academic programs
predicts future success better than test scores. Thus, it would seem that
a major decision made on the basis of a test score alone is undesirable .
Tests do not naake decisions, they merely provide supplementary
information for those who do make the decisions. An institutional or
individual decision made on the basis of a single test score alone is a
gross misuse and a misunderstanding of the purposes and limitations of
tests . Tests merely sample behavior at any given time and place , and
as such, are subject to various errors. Consequently, in many cases
test scores and interpretations are insufficient tools, not to be
exclusively relied upon. The reader is urged to utilize all the available
information in making any kind of decision .
NOTES
CHAPTER I
1 . L.J. Cronbach, "New Light on Test Strategy from Decision Theory,"
Proceedings of 1954 Invitational Conference on Testing Problenns
(Princeton, New Jersey: Educational Testing Service, 1955), pp. 31-32.
2. Ibid.
- 8 -
CHAPTER II
CLASSIFICATION OF TESTS
Tests can be classified in a variety of ways — according to structure,
purpose, and method of administration. They may be more or less
objective or subjective, highly structured or unstructured, designed for
administration to groups or individuals. Tests employed in clinical
practice include intelligence tests, achievennent tests, aptitude tests,
interest tests, and personality tests. There are other "special diag-
nostic tests" frequently used to assess some particular limitation or
potential — such as those especially designed to measure the nature and
severity of certain types of reading or learning disabilities as well as
those disabilities imposed by organic deterioration or damage . Some
tests nneasure "talents" inherent to artistic or musical productivity.
Intelligence tests have probably the longest and most comprehensive history
in clinical, industrial, and academic settings, a fact due, perhaps, to the
belief long inherent in Western society, that achievement and productivity
highly correlate with a concept known as "intelligence." Controversies
regarding the actual nature of intelligence have raged for thousands of
- 9 -
10
years so that conflicting opinions and experinnental data occupy many
volunnes.
Since direct social service practitioners are likely to be most con-
cerned with a test's intended purpose, a classification of tests was
developed:
1 . achievement and aptitude tests, including intelligence tests;
2 . personality and interest tests; and
3 . special diagnostic tests .
A considerable amount of confusion exists concerning distinctions
between intelligence, achievement, and aptitude. The crux of the
argument is whether intelligence as a specific concept can be separated
from other factors such as previous learning, achievement, and special
kinds of aptitudes. Most theorists today would argue that intelligence
tests, aptitude tests, and achievement tests all sample and measure
various parts of the same thing. For example, Wechsler — who authored
several intelligence tests — defined intelligence as "the aggregate or
global capacity of the individual to act purposefully and think rationally
and to deal effectively with his environment."^ Others argue that intelli-
gence is a function of the total personality and cannot be separated from
other aspects of the personality. However, Wesman advocates perhaps
the most comprehensive and one of the nnost generally accepted definitions
of intelligence in the literature: "Intelligence . . . is a summation of
learning expediences." This definition recognizes that when measuring
1 1
intelligence the result of many learning experiences and diverse perfor-
mances are actually sannpled. Wesman's definition, by implication, does
away with artificial distinctions between intelligence, aptitude, and
achievement tests . He contends that all of these devices measure what
the individual has learned. The difference in labeling merely signifies
the different purposes for which the tests will be used. This can be
clarified by considering each of the three separate categories — intelli-
gence tests, aptitude tests, and achievement tests.
Intelligence Tests
Intelligence tests comprise a highly specialized field with a vast
body of literature and research surrounding their use. A tremendous
variety of intelligence tests are available and in use . Each test reflects
the specific definition of intelligence and different personality theory
connmitment of the author. Sonne tests only include verbal material,
others contain nnuch non-verbal material. Sonne stress problenn solving,
while others ennphasize nnennory. Certain intelligence tests result in a
single total score, for exannple an I.Q., whereas others yield several
scores or subscores .
Varying emphases lead to different test results. One should expect
to find differences in the intelligence test scores of the same person who
is examined with different tests. In each test, nneasures of different
kinds of abilities are obtained. Under the circumstances it would be sur-
prising if each intelligence gave us nearly the same test result.
12
For most purposes, intelligence tests are considered measures of
general learning or scholastic aptitude, most useful in predicting achieve-
ment in school, college, or training progranns.
Aptitude Tests
Aptitude tests also attempt to measure an individual's potential for
achievement. However, aptitude tests focus on more circumscribed
varieties of achievement than do intelligence tests in determining
whether an individual has the potential for achievement in a specifically
defined area. For example, an individual's artistic or nnechanical
aptitude may be measured. Although intelligence correlates to a degree
with both aptitude and achievement, studies have shown that high
intelligence does not necessarily guarantee astuteness or potential in
certain areas. Recent data shows that intelligence as generally defined
does not highly correlate with creativity, especially in the artistic sense,
as has been supposed. It is not unconnmon to observe individuals who
appear extremely intelligent in the traditional sense of the word but who
simply do not seem to possess or have developed certain aptitudes. Witness
the college professor or physician who is a whiz in the classroom or
operating roonn but who is helpless when faced with an ailing carburetor.
An aptitude test uses a sample of behavior to predict future perfor-
mance in some specific occupation or training program .
In general use are two major types of aptitude tests: (1) broad-range
aptitude test batteries to sample general aptitudes, and (2) specific
13
<
aptitude tests to sample special aptitudes such as music, mathematics,
and art .
The most widely used broad-range batteries are the Differential
Aptitude Tests (DAT), for high school students, and the General
Aptitude Test Battery (GATB) currently utilized by the United States
Employment Service. In addition, a myriad of multi-score aptitude
test batteries exist.
Often aptitude tests are employed in selecting individuals for jobs,
for admission to special training programs, or for scholarships.
Primarily, the tests predict an individual's potential for achievement
in specific occupations or endeavors .
Achievemient Tests
Achievement tests, although in many ways similar to intelligence
tests, are generally designed to determine what an individual has
actually achieved in a certain area of endeavor. They are used to
measure a person's present level of knowledge or competence in, for
example, courses like nnathennatics, science, reading, chemistry, etc.
Many achievement tests, unlike other types, are not standardized but are
produced locally. For example, teachers normally develop achievement
tests to deternnine mastery of the course material. Thus, an achieve-
ment test examines a person's success in past or present study; in
contrast, aptitude tests forecast success in some future study.
Achievement tests, most widely used in academic settings, are usually
14
reported in the form of grade levels or similar measures of compari-
son.
Personality and Interest Tests
Personality and interest tests focus on what a person typically does
or might do in a given situation. What personality and interest tests
nneasure, in contrast to intelligence or achievement tests, is far less
clearly defined . Here a trennendous number of different ternns describe
sinnilar kinds of things — ternns like adjustment, personality, tempera-
nnent, interest, preferences, values, attitudes all describe similar,
broadly defined attributes. It is difficult to say what a specific person-
ality test score means, even after having given the matter careful
consideration .
Personality tests. Clinical psychologists and others interested in
the prediction of human behavior have long favored personalib/ tests.
They realize that those patterns of behavior usually referred to as "per-
sonality" have a strong influence on what we do. Personality, for
exannple, can largely determine how people characteristically use or
direct their intelligence and special creative aptitudes. Indeed, person-
ality "deficits" or distortions lead to little constructive use of one's
talents. Thus, an assessnnent of personality is vital to those who attempt
to help a person channel his or her efforts toward constructive vocational
or social use.
15
Most personality tests rely on vaguely defined scores and scales
used inconsistently fronn one author to the next, based on an unde in-
lying rationale not always specified .
Few people agree on a standard classification of personality tests.
However, at least three different types of tests are in general use:
1 . The objective test batteries not directly subject to clinical
interpretation for initial scoring. The Minnesota Multi-
phasic Personality Inventory (MMPI) is one example of this
particular test.
2. The less commonly used situation test measures performance
in complex life-like situations or simulated situations and
tests special kinds of abilities involving overall responses
of an individual to specific situations. Industry commonly
uses it to test leadership abilities. When a person is given
a group and a specific task to accomplish, he or she is
observed in the process of completing the task,
3 . Projective tests are designed to elicit subjects' responses to
an ambiguous stimulus such as a picture or an inkblot.
The individual's response is interpreted and scored on the
assunnption that the way he organizes and responds to
unstructured or ambiguous stinnuli indicates the way he
organizes and responds to the world around him. Responses
are assunned to be projections of the subject's unconscious
16
wishes, attitudes, and values. The scoring method is similar
to the psychoanalytic method of dream interpretation.
Typical projective tests in wide use are the Rorschach (inkblot)
and the Thematic Appreception Test (TAT).
Personality tests are primarily used to predict the future behavior
of individuals in both general and specific situations . They commonly
aid in predicting post-institutional adjustment for persons released
from prisons, hospitals, and schools, or in predicting the likelihood of
marital success or job performance. Test reports typically contain
terms such as anxiety, ego, libido, cathexis, sublimation, etc. A
great deal of controversy surrounds the use of various personality tests,
especially regarding their validity and reliability. In general, projective
tests are not uniformly accurate in predicting the behavior of individuals
in either a specific or general situation, but are accurate, given extreme
individuals and extreme situations. Although it may be fascinating, a
projective test may prove disappointing if used to accurately predict
behavior in a way that might be useful to most practitioners.
Interest tests . Interest tests are specific personality tests used
mainly in vocational and educational guidance . They are difficult to
separate from aptitude tests, but come under the general category of
personality tests because they are directed toward such things as predict-
ing a person's potential satisfaction with a given type of work. The two
nnost widely u^ed interest inventories are the Strong Vocational Interest
Blank and the Kuder Preference Record (see exannple below).
17
Mr. Williams' scores on the Kuder Preference Record indicate that he
is highly interested in science, computational activities, and clerical work.
These interests are at the 95th, 91st, and 87th percentiles respectively. He
demonstrates moderate interest in art and mechanical areas also. The latter
interests are at the 75th and 70th percentiles . Training areas he may wish
to consider then are: computer programming, computer technology, x-ray
technology, laboratory technology, drafting, mechanical drawing, connputer
systems analyst, electronics technology, radar technology, chemical standards
work, industrial standards work, bookkeeping, accounting, printing, etc. As
noted earlier, his intellectual level and academic preparation are quite
sufficient for hinn to be successful in a four-year college or technical
program .
Interest tests generally rely on self-reporting techniques and are
designed to sample both leisure time and work-related activities given
specific personality aspects in the area of personal likes and dislikes.
They are used to determine the amount of preference a person displays
for one activity over another. For example, the inventories typically
sample reading interest by asking people if they would prefer to read about
adventure, business, science, or ronnance . Another example, the California
Occupational Preference Survey, samples eight interest categories: science
technical, outdoor, business, clerical, linguistic, aesthetic, and service.
Although the interest inventories are considered separately for
analysis, they are generally regarded as special personality nneasures
18
used specifically to predict occupational, vocational, and educational
adjustment. However, for purposes of classification, we may regard
them as personality measures that fall under the subcategory of objective
testing devices. Almost all of the interest batteries rely on objectively
scored testing methods based on standardized methodologies.
Specific Diagnostic Tests
A widely diverse group of tests developed for highly specific
purposes tend to defy classification. Most such tests were developed
to measure specific abilities or disabilities. Some tests diagnose
cerebral pathology such as brain lesions or other organic abnormalities .
There is disagreement as to how much these tests actually measure
underlying pathology or primary causation as contrasted to possible poor
learning conditions. For example, the Bender Visual Motor Gestalt Test,
sonnetimes regarded as a test for the diagnosis of possible brain damage,
may also be considered a straight-forward ability test. This test requires
the subject to produce various geometric designs using a pencil and paper.
The way in which they go about achieving this task is subject to various
scoring procedures. Most examiners agree that the Bender Test (ByMOT")
is basically a performance test since the examinee is affected by previous
learning. However, ther^ is indication that the Bender does some rough
screening for identifying persons with possible brain dannage.
NOTES
CHAPTER II
1 . David Wechsler, The Measurement and Appraisal of Adult
Intelligence, 4th ed. (Baltimore: Williams and Wilkins, 1958),
p. 7 .
2. Alexander G, Wesman, "Intelligence Testing," American
Psychologist 23 (1968) : 267.
- 19 -
CHAPTER III
BASIC TEST CONCEPTS
A knowledge of some basic testing concepts, including their con-
struction and utilization, is central to understanding the Unnitations of
various tests. Two concepts constitute the criteria used forjudging
a test in its totality: reliability and validity.
Reliability
Reliability refers to the consistency of nneasurement of any test.
A test cannot nneasure anything well unless that sonnething is measured
consistently. It is important to realize that although a test measures
things consistently it may not measure the desired characteristic.
In Chapter I, tests were defined as samples of behavior. Because
they are, they will show variation from sample to sample — that is, we
nnay expect differences in the behavior of an individual from one testing
situation to another. Reliability of a test is measured by the extent to
which results vary from sample to sample . It is necessary to obtain a
high degree of reliability in test results to ensure confidence in that
test and to achieve validity.
-■ - 20 -
21
Factors affecting reliability. A number of variation sources affect
the reliability of tests. They are:
1 . Test length. Assuming that fatigue does not become a major
factor, a longer test is more likely to be reliable than a
shorter test.
2 . Time between tests. The length of time between two testings
will affect reliability. The shorter the time between the two
tests the more likely it is that the re-test will be similar.
3 . Irregularity of testing conditions. Changes in conditions
from one testing situation to another will affect the test
reliability. Failure to follow specific directions for giving
the test may reveal a considerable amount of difference on
the scores obtained from the same test taken by the same
individual at different times. Extreme differences in
physical conditions — overly heated, uncomfortable test roonns,
or poor lighting conditions — will also affect reliability. Other
factors such as the examiner's responses, racial differences
between the examiner and subject, moods, illness, cheating
by the examinee, etc., may threaten test reliability.
4. Scorer error. When tests are not scored objectively, or
the details of scoring ignored, unreliability results.
Objective tests reduce the possibilities for scorer error.
Tests designed to elicit subjective responses require special
22
training for scoring. Indeed, many score errors occur as a
result of the exanniner's inexperience.
Determining reliability. Two basic procedures are used to establish
test reliability: the test/re-test procedure and the alternate form
procedure . The test/ re-test procedure involves testing and re-testing
the same individuals at different tinnes using the same instrument. Test/
re-test results are usually reported as a reliability coefficient which
represents the degree of agreement between the two nneasures.
The alternate fornn procedure involves using two different tests
designed to measure the same thing. Alternate fornn tests are used
when the examiner believes that exposure to one test will contanninate
the later responses of an individual if he or she is tested again with the
same instrument. In other words, a person may have learned what to
expect fronn a set of specific test questions, therefore, influencing
his or her second response to the same set of questions. In this case,
an alternate form of test may be developed to test similar attributes.
Alternate form reliability is also reported as a reliability coefficient
representing the degree of agreement between the two measures.
In addition to reliability deternnined by the actual construction of
the test, it is imperative that the reliability of test scoring be controlled.
Scorer reliability may be an important factor for some tests —
especially projective tests such as the Rorschach . Scorer reliability
can generally be highly developed by providing standardized training
23
for scorers and by comparing the scoring of several examiners for the
same test. It may also be reported as a numerical figure which is
usually referred to as an interscorer reliability ratio.
Validity
In order to make a statement about the general validity of a test,
it is essential that the user of any psychological test information
determine what kinds of decisions he or she is going to make regarding
the use of that test. In its broad sense, validity denotes the extent to
which a test measures or predicts that for which it was designed. In
other words, validity is the most basic and perhaps the most important
single attribute of the test for it must do what it is designed to do. If
a test is supposed to predict occupational success, the extent to which
it does so may be said to be a measure of its validity. However, it is
important to recognize that psychological tests may have a high degree
of validity for one purpose but almost no validity for other purposes .
Various measures of validity, used in psychological testing,
include face validity, content validity, predictive validity, and
construct validity . Although a great deal has been written about validity
measures, for practical purposes most social service practitioners will
prinnarily concern themselves with predictive validity.
Predictive validity, also called empirical or criteKon validity,
is established by detemnining how well a test predicts perfornnance
24
against a specific criterion. A test's validity is deternnined by opera-
tionally defining what the test should do and what outcomes it can
predict: The test's success in predicting that outcome is the extent to
which the test proves valid. Thus, if a practitioner uses an instrument
to screen people for discharge from a correctional institution, how well
that instrument predicts specific aspects of post-institutional adjustment,
such as recidivism, determines the test's validity. Again, validity is
determined by a specific definition of what the test should do . The
same test, for example, might prove ineffective in identilVing potential
salesmen for an automobile agency.
Schools establish predictive validity by using intelligence tests
to predict potential achievement. Scores obtained on specific intelli-
gence tests are compared with grades earned in school . In a similar
way the predictions of occupational preference tests are validated by
comparing ratings by individuals and employers at a later date.
A number of factors influence predictive validity.
1 . The specific criteria used to establish validation may vary
from study to study with different scores obtained from each
criteria. Therefore, it is necessary to carefully consider
which criteria is most important for the decision at hand .
2. Some tests are defined mKDre specifically than others in
terms of what they are intended to do. For example, easily
identified criteria such as school grades can validate a
25
scholastic aptitude test but it is very difficult, if not impossible,
to establish an acceptable criteria for an anxiety scale or a
value scale. Where such difficulty in defining criteria related
to the intended results of the test exists, high validity cannot
be expected.
Practitioners should keep in mind a general, useful rule of thunnb:
For a test to have any utility, it must provide accurate information that
can help to predict behavior. Thus, the less specific the objectives of
the test, the less useful it is in predicting behavior.
Predictive validity is generally reported as a numerical figure
called a validity coefficient; a nneasure of validity achieved by connputing
a coefficient of correlation between the test and a criterion.
CHAPTER IV
BASIC STATISTICAL CONCEPTS
In addition to understanding reliability and validity as essential
concepts underlying test construction, the consumer of test infornnation
should know how test results are typically reported .
Nornns
The fanniliar expression, "How are you?" and the response, "In
relation to what?", best expresses what norms are all about. That is,
they provide those standards against which a given value is compared .
Norms may be used to determine how well a person does in connparison
to other people.
Many people in the social services view test results simply in the
form of a raw score. Without more information, it is impossible to use
the test results to make a productive decision. For example, the raw
score of 65 may mean that 65 test items were answered correctly.
If there is no nornn for comparing that score with other people's responses,
one can attach no meaning to that score. A set of norms is imperative to
understand the meaning of raw test scores.
.. - 26 -
27
Cronbach has defined a test as "a systematic procedure for compai —
ing the behavior of two or more persons." In spite of the philosophical
difficulties one may have with making comparisons, psychological
testing does just this. A norm is nothing nnore than an average score
for a specified group used to make comparisons between individuals and
groups. On some tests, for example, the performance of persons in a
specific geographic location is connpared with the performance of
persons nationally.
It must be perfectly clear to the social service practitioner just how
an individual's test results compare with specific responses from other
groups. Who or what a person is compared to nnakes a great deal of
difference. Consider the following exannple. Jane Sloe, a seventeen-
year-old inmate in the state correctional institution for girls, received
a raw score of 163 on a vocabulary test designed to predict acadennic
success in collegiate programs. At this point, her probation officer
nnust decide whether to release her by September 1 so she may enter
college. Her raw score of 163 nneans that she did as well or better than:
99 percent of the residents of the state school for girls;
87 percent of the twelfth-grade students in Capital City;
83 percent of the entering freshmen at State University;
75 percent of the graduating seniors at State University;
96 percent of the custodial and treatment staff at state school for
girls;
28
15 percent of the faculty in the English Departnnent at State
University .
Although Jane's absolute performance remains unchanged, the impression
of how well she has done may differ considerably as the norm groups
change. Admittedly, this illustration is extreme, but it does point up
the importance of specifying the norm group to which one compares a
person.
Thus, to fully understand a norm group one must gather as much
infornnation describing the norm group as possible and determine how
the person tested differs fromi it. In viewing any norm group, consider
such innportant variables as age, sex, previous education, socio-economic
background, ethnic membership, and occupation. In other words, one
should use the most appropriate norm group for the individual examinee
and the situation involved .
Publishers supply norm information on most standardized tests,
especially educational achievement tests. Most test publishers routinely
report norm information and specify if they will make available both
local and national norm information. In addition, standard test
references also report nornns . Social service practitioners should use
this rule of thunnb regarding norms: the more infornnation provided
describing the norm group, the more accurately you can assess to what
extent a given individual resennbles the norm group.
29
Test manuals usually provide brcadly based or "national" norms.
When using such norms it is innportant to get more detailed information
about the groups used to establish these norms . Most of the norm groups
will not comprise an entire population; they sample what the test con-
structors think of as the relevant population. To establish norms, they
divide the relevant population into subgroups that appear in the sample in
proportion to their numbers in the population. Ideally, those individuals
who connprise the sample from each subgroup are selected randomly.
Frequently, test constructors subdivide populations according to such
characteristics as rural-urban residence, age, sex, race, socio-
economic status, religion, and geographic region. Sonnetimes they
seek to establish from a specific stratified population of people what
they consider to be normal performance ranges. In sonne cases,
however, they nnay leave out certain elements of the population in the
original norming groups, thus making the specific test irrelevant for
use on that population. For example, one of the most frequent
cirticisms of intelligence tests is that adequate sannples of Anne ri can
Indians were not included in the original norming group. Such tests,
like the Wechsler Intelligence Scales, may make a poor basis for connparing
the performance of American Indians to other segments of the population.
Social service practitioners, then, nnust focus on the detailed
description of the norm group's relevant characteristics. Furthermore,
when subgroup differences are known to be related to test performances.
30
it is important to report separate norms for the subgroups. A case in
point involves the effects of early child-rearing practices and develop-
ment in multi-lingual homes . A test that focuses on the development of
English vocabulary, standardized on a population of midwestern school
children, nnay not be a valid basis for comparing the performance of
southwestern Chicano children who come from Spanish-speaking homes.
Again, it may be desirable and even necessary to establish separate
norms for people from a similar population before meaningful comparisons
are made. One must remember that how accurately a norm group repre-
sents the population to be tested is nnore essential than the absolute size
of that norm group. True, the larger the sample the more stable the
statistics based on the sannple, but a representative norm group of modet~-
ate size is more useful than a large, poorly defined group.
Measures of Position
Numbers which tell us where a score value stands within a set of
scores are measures of position. There are two comnnonly used
nneasures of position: rank and percentile rank.
Rank is the simplest description of position. It designates the
highest, the next highest, the third highest, and on to the lowest — a simple
way of describing the position of a person or a score with respect to a
distribution of scores. However, it has a major limitation: its interpre-
tation depends on the size of the group. It is generally used in an informal
sense such as designating a person's standing in their high school
31
graduating class; but the meaning of graduating first in a class of three
in the Polaris, Montana High School is not as clear as graduating first
in a class of 5,000.
Percentile rank states a person's relative position within a defined
group. Thus, a percentile rank of 97 indicates a score as high or higher
than those made by 97 percent of the people in that particular group.
Percentile ranks are one of the nnost widely used nneasures of position
for reporting test scores, especially on scholastic achievement tests.
Although easily understood and commonly used, percentile ranks have a
major limitation; they are based on the number of people with scores
higher or lower than the specified score value. Percentile ranks tend
to obscure all information about those scores' distribution and the
absolute differences in raw scores achieved by individuals. However,
this information can be regained by focusing on other measures of
variability and central tendency.
Measures of Central Tendency
A measure of central tendency is a representative common denomina-
tor for a set of scores . Three common measures of central tendency are
in general use: the arithmetic mean, the crude mode, and the median.
Mean. The mean, or arithmetic mean, is nothing nnore than an
average. To arrive at the nnean, all the scores are added up and divided
by the number of scores .
32
Median . The median is the nnidpoint of an array of scores . It
is the point above which and below which 50 percent of the scores fall.
The median is determined by simply ordering the scores from the lowest
to the highest and selecting the middle score in the range of scores
represented. For example, if there are five scores present, as follows,
1-3-6-9-12, the middlemost score in this distribution, or the median, is 6,
The median is the score's position with respect to others and has very
little to do with the absolute value of that score .
Mode . The mode is the nnost frequent score occurring in a distribu-
tion. Thus, for the scores 1-5-5-2-5-3-5, the mode is 5. This is one of
the crudest measures of central tendency and is used only for rough
estinnates.
Measures of Variability
Measures of variability describe the extent of score dispersion in a
particular distribution and the degree to which scores vary from each
other. It is important to know about variability nneasures in order to
compare the score of a given person or group of persons with the disper-
sion that is logically or reasonably expected. Comnnon measures of
variability include range, semi-interquarterile range, and the standard
deviation .
Range . Range arrives at a rough measure of variability by
identifying the two nnost extrenne scores, i.e. , the highest and lowest
33
score. Thus, the range is simply the difference between the highest and
the lowest score .
The senni-interquartile range. The senni-interquartile range
describes the dispersion represented by the middle half of a distribution.
In other words, the semi-interquartile range represents the distance
between the twenty-fifth and the seventy-fifth percentile on a distribution.
It is used in conjunction with the median as a measure of central tendency,
especially where the test achieves atypical and highly unexpected distribu-
tions .
Standard deviation. Standard deviation is perhaps the most widely
used, most dependable measure of variability because it fits mathematically
with other statistics and thus becomes the basis for a number of other
statistical measures, including standard scores, deviation I.Q.s, T
scores, and z scores. The standard deviation is the square root of the
mean of the squared deviations from the mean (of a distribution). The
standard deviation is generally represented either by the Greek symbol a
or the letter s . Thus s = a = standard deviation .
The standard deviation is used to make interpretations of the
variability of scores in a distribution. For example, the standard devia-
tion has known characteristics. In a normal distribution of scores, about
34 percent of those scores lie between the mean and a point that is one
standard deviation on either side of the mean. Thus, 68 percent of the
scores on that distribution will be dispersed between a point lying one
standard deviation below and one standard deviation above the mean. In
34
other words, we know that in a nornnal distribution we can expect approxi-
mately 68 percent of the population to fall within one standard deviation +
of the nnean .
You will find knowledge of this dispersion useful because if you know
the standard deviation of a test and a person's score you can use this
infornnation to estimate how that person compares with others who have
taken the same test. For example, the standard deviation on the Wechsler
Adult Intelligence Scale is 15 and the mean is 100. Thus, if we find a
person who scores at 145 on the Wechsler, we can translate that information
as follows: he has scored three standard deviations above the mean or
higher than 99 percent of the people who have taken the examination. This
calculation was achieved by adding the number of standard deviations above
the mean to the percentage of cases included below the nnean (the bottom
50 percent). The standard deviations can be converted into percentages
of cases by reference to standard tables such as found in Table I.
Measures of Correlation
Measures of correlation determine how much two or more variables
relate to each other. A correlation coefficient is an index number which
expresses, in nunnerlcal values which range from .00 (no relationship)
to +1 .00 (perfect relationship) or to -1 .00 (negative relationship), the
degree of relationship between two or more variables . The higher the
number expressed as a coefficient of correlation, the more clearly
related the two variables — and one must rennember that negative correla-
tions are as important as positive correlations.
35
TABLE I
COMPARISON OF SOME STANDARD SCORES
Percent of scores in
that portion of the
distribution
Standard Deviation
Units
-4s -3s -2s -1s Mean +1s +2s +3s +4s
Approximate Cumulative
Percentages
0.1% 2% 16% 50% 84% 98% 99+%
T Scores
Deviation I .Q.S
Wechsler
Stanford-Binet
IT ED
CEEB
A OCT
20 30 40
5 10
50
15
60 70 80
55 70 85 100 115 130 145
52 68 84 100 116 132 148
20 25 30
200 300 400 500 600 700 800
J_
J L
40 60 80
100
120 140 160
36
Correlation coefficients are utilized in testing to report measures
of validity and reliability. Validity studies use them to express the
degree of relationship between the test scores and certain criterion values,
Reliability studies use them to express the degree of relationship between
scores for both test/ re-test and alternate forms of reliability.
Two of the several methods used to compute correlation coefficients
are the popular Pearson Product Moment Correlation Method and the
Spearman Rank Difference Correlation Method. Any elementary statisti-
cal textbook contains these and other methods for calculating correlation
coefficients and for determining the reliability of those coefficients.
Inferential Statistics
Inferential statistics are sonnetimes called probability statistics
because these measures tell us how much confidence nnay be placed in
descriptive statistics — those numbers used to describe the actual results
that people achieve on tests. Generally, inferential statistics are
reported as "probability values." They indicate how likely the results
obtained on a given test would occur by chance alone. Of the few inferen-
tial statistics referred to in test literature, perhaps the most important
is the standard error of measurement. The standard error is a statistic
used to estinnate how likely a specific test score will diverge from the
true score achieved by a person. In other words, the standard error indi-
cates how much a person's score would vary if he or she were exannined
37
repeatedly with the same test. This standard error of measurement is
one way of expressing a test's reliability.
Other inferential statistics commonly used include thejK^ (Chi
Square) test and the Fisher's t. For further information regarding
inferential statistics, a basic textbook on probability statistics should be
consulted .
Raw and Standard Scores
The direct numerical report of a person's test performance is
called the raw score. This score may represent the number of questions
answered, the time required to connplete the test, or any other numerical
value representing test perfornnance. Raw scores are easily misundei —
stood because they do not include a basis for connparison. Testers
generally convert them into standard scores which are then reported.
Standard scores, derived from raw scores, are used as part of a
scoring systenn which usually offers other infornnation so that all scores
provide a basis for comparison. They are used to report test results in
almost all intelligence and achievement tests . Rennemiber: (1) most
standard scores are based on the properties of the nornnal curve, and
(2) standard scores generally include such infornnation as the mean and
the standard deviation of the distribution . They are extremely useful
because they pernnit comparisons between tests that have similar types of
scoring systems. The basic standard score is the z score — a nnethod of
comparing scores on one test with scores on other tests.
38
Following is an example of the use of standard scones in comnnon
psychological tests. The I.Q., or Intelligence Quotient, used around the
turn of the century as a way of measuring the intellectual capacities of
people, was computed by determining the ratio between mental age and
chronological age . Although this particular method was used in the
original Stanford-Binet Individual Intelligence Test, it is now nearly
obsolete. Instead, the Stanford-Binet and other tests such as the Wechsler
have converted to a standard score known as a deviation I.Q.
The Wechsler Adult Intelligence Scales (WAIS) established the
deviation I.Q. by utilizing six verbal subtests and five perfornnance sub-
tests, each of which yield a raw score, establishing different norm groups
to cover different age ranges from 16 to 64 years. A raw score distribu-
tion resulted for each of the eleven subtests. Each subtest raw score was
then converted into a standard score with a mean of 10 and a standard
deviation of 3. The sunn of all the subtests provides an overall raw score
which is converted to a standard score with a standard deviation of 15 and
a mean of 100. Tables are provided for the conversion and to adapt for
different age groups. Thus, the WAIS user now only needs to consult the
appropriate table to find the I.Q. value which corresponds to the sum of
the subtest scores.
In a similar way, the 1960 revision of the Stanford-Binet was con-
verted to adopt the deviation I.Q. so that the standard deviation would be
consistent from age to age . Previously it was noted that the standard
39
deviation for different ages in the 1937 revision of the Stanford-Binet
differed by as nnuch as 8 I.Q. points. Thus, the Stanford-Binet evolved
to have a mean of 100 and a standard deviation of 16. The way in which
the raw scores are converted to scaled scores and thus the deviation I.Q.
is very sinnilar to the procedure ennployed in the Wechsler tests.
Another comnnonly encountered test which is based on a standard
score is the College Entrance Examination Board (CEEB), administered
by the Educational Testing Service. The CEEB was standardized on a
population of college applicants in 1941 . Using the scores based on the
applicants of 1941 , the testers have developed the CEEB as follows. The
mean of the test is 500, the standard deviation is 100. With this informa-
tion in mind, the test user can estimate the relative position of any
person's given score. For example, a score of 800 on the CEEB is three
standard deviations above the mean . This means that the examinee
achieved a higher score than over 99 percent of the population on which
the test was standardized.
Other connmonly used standard scores include the Stanine Score,
the T Scale Score, the C Scale Score, the Sten Score, and the Iowa Test
of Educational Development (ITED) Score. Although a number of other
types of standard scores are used. Table I will give the reader a general
idea of how the commonly used scales compare with each other.
40
Ratio Scores and Placement Scores
Because the term "I.Q." can mean several different things, it is
frequently misused and nnisunderstood. The original I.Q. concept
developed by Terman was a ratio type I.Q. found by the formula I.Q. =
100 X p^ , where MA is mental age figured fronn an intelligence test and
CA is the exanninee's chronological age at the time of testing. The
rationale of the ratio type I.Q. is widely understood but this type of score
has many limitations for it imiplies that mental age units are of equal size,
which is not verified by research evidence. Ratio I.Q.s work reasonably
well if the examinee's age is between five to fifteen. Outside of these
approximate limits, they tend toward invalidity. This is because the
differences (in nnental growth) between the ages of five and six, for
example, are nnuch greater than between the ages of fifteen and sixteen
or between twenty-five and twenty-six. Unlike chronological age units,
nnental age units are not equal. Thus, when interpreting an MA or ratio
I.Q., the social service worker nnust be cautious.
Another type of comnnonly used ratio score is the educational quotient.
It resembles the mental age concept and is used to estimate a person's
nnininnal perfornnance in comparison with other people who perform in
educational settings. Dividing an educational age score achieved on a
test (EA)by the chronological age and multiplying by 100 will result in the
educational quotient (E.Q. = 100 x ■^). Educational quotients have the
same advantages and limitations as ratio I.Q.s. However, they are only
used with school-level achievement tests .
41
The most common score used in reporting performance on standard-
ized achievement tests for school children is the grade placement score
which resembles the age scores. It is found by determining the average
score of school students at a corresponding grade placement. Grade
placement scores are usually stated in tenths of a school year. For
example, 6.2 refers to the second month of grade six. This approach
assumes that children learn relatively uniformly throughout the school
year but that no learning occurs during the summer vacation — an assumption
not necessarily true. Grade placement scores, however, are generally
used to estimate where a person should be placed in school- related work.
NOTES
CHAPTER IV
1 . Cronbach, Essentials of Psychological Tests, pp. cit. , p. 21
- 42 -
CHAPTER V
LIMITATIONS OF TESTS
Supplementary Measures
The full assessment of a person's abilities, disabilities, and various
personal qualities requires a progressive type of approach. Since a test
is simply a behavior sample to be regarded cautiously, it follows that test
scores are minimal estimates of behavior and abilities. Broad spectrum
tests like the Wechsler Intelligence Scales are initial steps to assessing
general ability. When a social service worker must make an important
decision, he or she should move through progressive stages of assess-
ment, using test information that deals with specific intellectual, perceptual,
and/or cognitive factors. Existing tests can only provide clues and rough
estinnates regarding an individual's abilities and capacities. REGARD THE
RESULTS OF ANY ONE TEST CAUTIOUSLY.
Test Construction Limits
One of the major difficulties in deciding how to best use the various
kinds of test results springs fronn the uncertainty over what tests actually
measure . Some tests are designed to measure verbal learning and
- 43 -
44
abstractions, while others assess manual skill potentials. The Wechsler
and the Stanford-Binet were originally constructed to nneasure both
performance and verbal factors. Yet many psychologists agree that the
weighting of the Stanford-Binet is more toward measuring verbal ability
than are the Wechsler tests which strike a more even balance between
verbal and performance items. However, a person's achievement on
either verbal or performance tests partially reflects previous learning
experiences. Thus a number of other factors, all related to previous
exposure to similar materials, are important determinants for assessing
test limitations .
Factors such as membership in specific geographic groups are
related to mastery of subject content. For example, urban residents
are exposed to a more diverse range of stimuli than are rural residents.
A rural resident may define an elevator as a grain storage facility, while
an urban resident nnay describe an elevator as a device used to nnove
people up and down in a building. If this were a test question, i.e. , define
an elevator, standardized on an urban population, the rural answer might
be judged unacceptable. Such a question will not fairly measure the rural
resident's potential ability to acquire knowledge. Similarly, many
achievement and ability tests do not fairly sample the potential abilities
of members of various ethnic groups.
45
Effects of Culture
For a test to be truly fair, all of the examinees should have had an
equal opportunity to acquire the needed background . There have been
many attempts to construct culture-free or culture-fair tests — those that
supposedly do not depend on previous experience — but most social
scientists believe that experience affects all behavior.
A number of cultural factors affect test performance. Previous
training experiences influence outcome. The Zuni Indians are taught
cooperation rather than competition, and so the performance of a Zuni
Indian on a competitive test may reflect this teaching and could vary
considerably fronn a person reared in a culture which stresses connpeti-
tion . Some tests may exhibit a cultural bias simply because the examinee
is not familiar with the testing materials.
Additional factors that affect test scores include the sex and race
of the examiner. Carkhuff and Pierce have reported that the race of both
the examiner and the tester appear to have a significant effect upon the
outcome of clinical interviews . "• Unless the culture and communication
patterns of the group tested are thoroughly understood, the examiner, no
matter how unprejudiced or objective, may not obtain maximum results
in the test procedure. Sensitivity to all aspects of a subject's behavior is
essential for acquiring a fair test result. This type of sensitivity does not
develop through academic efforts, but rather through prolonged, intimate
contact with the specific ethnic group.
46
Most tests used for individual mental testing do not truly represent
Sonne of the ethnic groups in the United States. For example, since
the Stanford-Binet standardized on white, nniddle-class subjects, it
primarily nneasures verbal ability and generally reflects this culture;
subjects from other ethnic groups often do poorly on this scale. The WAIS
and the WPPSI include black subjects in the norming group, but not other
ethnic groups. Furthermore, since the WISC includes only white students
in the standardization sample, the scoring criteria used in establishing
correct and incorrect responses discriminate against the responses made
by Chicanes and American Indians. Thus, when these standard instruments
for determining the ability of ethnic minority nnennbers, especially American
Indians, rural residents, Chicanos, and others not represented, are used
they should be interpreted with extreme caution and regarded as minimal
estimates of ability.
When test results are used to evaluate the potential performance of
a minority group member, the purpose those results serve should be
constantly kept in mind by the social service practitioner. Although a
test may discriminate against members of a specific minority group, this
will not necessarily diminish the test's predictive validity when used to
determine success in, say, a school program. If you want to obtain an
estimate of a minority person's chances of succeeding in a training
program, a standard intelligence test may prove a valid predicting instru-
ment.
47
Other Limitations
Keep in mind, always, that the performance of any individual on a
test is simply a sample of behavior, and as such, performance in any one
test is a minimal estimate. Naturally, the more measures of similar
variables the more reliance one can place on such estimates^ but any
contradictory scores on two or more tests that measure similar attributes
should be noted. If such a contradiction exists, seek professional inter-
pretation of the differences — at least three different variables may be
involved, including differences between the tests themselves, differences
in the individual or the group being tested, and differences in the conditions
of test administration fronn one test to another.
NOTES
CHAPTER V
1 . R.R. Carkhuff and R. Pierce, "Differential Affects of Therapist ,
Race, and Social Class upon Patient Depth of Self-Exploration in the
Initial Clinical Interview," Journal of Consulting Psychology 31 (1967)
632-35.
- 48 -
CHAPTER VI
HOW TO /W\KE A TEST REFERRAL
Most social service personnel should nnake test referrals as a
nornnal or routine part of their work. The best test referrals identify
the specific information sought .
Since testing should have as its major purpose the provision of
useful information for decision making, the psychologist who selects a
test that will provide such information needs to know what kinds of deci-
sions the social service worker intends and, in addition, will need to
know what kinds of information are already available about the subject.
Following is a suggested guide for making test referrals .
Suggested Guide for Test Referrals
1 . Reason for referral — what kind of information do you want and
what kinds of decisions will you try to make .
2. Description of the subject, including:
a. age;
b . sex;
c . education;
d . occupation and employment history;
- 49 -
50
e. ethnic membership and experience; and
f. note any special disabilities, handicaps, or physical
abnormalities .
3. The results of any previous testing, if available — including
testing dates, scores, and test names.
4. A brief history of the subject's involvement with the agency.
5 . A brief statennent of any case-management plans you have
for the subject.
Given this information, a competent examiner should be able to
select the tests that will best provide the information you need . The
examiner should also be able to interpret the tests' results in a way
that is useful in making specific case management decisions ,
Some Hints for Dealing With Psychologists
If your test report is in a form that is difficult to use in decision
making, ask the exanniner for a consultation or interpretation. But
remember, what you get out of such a meeting largely depends on the
questions asked , It is advisable to key your questions to specific decisions
about the subject. For example, will Johnny get through college at
State University? Or, is there a possibility Johnny may commit suicide?
Although no examiner can answer either question with certainty,
he/she can provide some information about the probability of either event
occurring . Ethical examiners will also provide the necessary explanations
about the limitation of the instrunnent .
51
Some test reports may be confusing because the examiner used
special tests which the test results do not explain. It is always appropriate
to ask the examiner to explain the purpose of all tests given.
You should also prepare people referred for psychological tests for
the actual examination procedure . Such preparation should include an
explanation of why the referral was made and a description of what they
should expect in the testing situation — how long it will take, where it will
be done, etc. You can obtain this information from the examiner on request.
It is not generally appropriate to request the examiner to administer
a specific test. Leave test selection up to the examiner unless you can
make a special case to justify an exception.
Many test reports contain much technical jargon. You can always
ask examiners to explain all terms you do not understand in a test
report. Technical jargon is meant for other psychologists and does not
generally convey a great deal of meaning to many test information users.
Psychological examiners have no magical powers. The same infor-
mation coming out of a psychological examination might well come from
others in everyday situations. People who have known the subject over
a period of time and have observed him/her in different situations can often
tell you more about the subject than most examiners. Remember that
tests only sample behavior; what happens in real life also fully indicates
what to expect from that person.
CHAPTER VII
SOME COMMONLY USED TESTS
Differential Aptitude Test (DAT)
The DAT battery, originally published in 1947, is currently avail-
able in two forms. High schools use it for counseling students in grades
eight through twelve. The eight tests measure aptitudes which previous
research had found important in guidance.
The DAT prinnarily provides a standardized procedure for mea-
suring boys' and girls' nnultiple aptitudes for educational and vocational
guidance . It yields separate scores from eight subtests plus a score
resulting fronn a combination of two of the eight subtests . The eight tests
are: Verbal Reasoning, Numerical Ability, Abstract Reasoning, Clerical
Speed and Accuracy, Mechanical Reasoning, Space Relations, Language
Usage (which includes Spelling and Language Usage) which deals with
grammar. The tests require six to thirty minutes of working time, plus
additional time for directions. Thus, each of three sessions need eighty
nninutes. Except for the clerical test, tests are not timed.
Both test/ re-test and alternate forms of reliability determination
have achieved highly acceptable reliability figures.
- 52 -
53
Validity has primarily been established by attempting to match test
performance with later course grades. In this respect, predictive
validity has been high enough to demonstrate correlations ranging from
.70 to .80. In general, however, the predictive validity for the subscales
instrument is not very high, usually around the area of .50. The best
overall predictor of grades is the combination score reported for verbal
reasoning and numerical ability.
The sampling of over 50,000 students from 195 different schools
in 43 states, representing all major geographic areas in the United
States, established DAT norms. (There are separate norms for boys
and girls and also for Fall and Spring Semester testing.) These norms
are expressed as percentile ranks and stanines . Remennber: Although
the DAT predicts success in coursework and grades reasonably well, it
does not adequately predict vocational success. Therefore, use it with
considerable caution when advising students on career selection. The
DAT should be used with other instruments for best results .
Goodenough-Harris Drawing Test (Draw-A-Man Test)
This test was designed for children five to fifteen years of age to
evaluate intelligence by analysis of the child's drawings of a man and a
woman. It can be used as an initial screening test, a rapid way of gaining
an impression of a child's general ability levels and as a means of esti-
mating the mental ability of children for whom the usual verbal tests of
ability are inappropriate .
54
The test booklets provide three spaces for the child to produce
drawings: one for the drawing of a nnan, one for a woman, and one for
a self-portrait. The examiner asks the child to draw the very best
picture possible of a man, a woman, and himself or herself. The child
is cautioned to make a whole person, not simply a head and shoulders
view. Although the test is not time limited, the child usually completes
it in ten to fifteen minutes. Tests may be administered either to groups
or to individual children.
The test contains fairly explicit scoring directions. It has, according
to research, relatively high coefficients of interscorer reliability (approxi-
mately .90), but tests for test/ re-test reliability only range from a test/
re-test reliability of .94 for a one-day interval between testing to .65
for a three-year interval between testings. Most test/ re-test reliability
coefficients for other tests range between .60 and .70.
The Draw-A-Man Test's validity has been prinnarily demonstrated
by correlations with scores on other tests. Correlations with the earlier
fornns of the Stanford-Binet range fronn .30 to .74, and range similarly
in other tests: reports show correlations of this test and the WISC
between .40 and .50.
Samples of 300 children at each age level fromi 5 to 15 years,
selected as representative of the population of the United States accord-
ing to father's occupation and geographic region, established norms for
these scoring scales . (The test manual reports standard score norms
which have a mean of 100 and' a standard deviation of 15.)
55
Limitations. Because, on this test, the standardized sample of
population held only a few critical variables constant (such as father's
occupation and geographic region) the nornns may not be useful for special
ethnic groups such as American Indians or residents of extremely rural
areas. Also, this test's validity has been established prinnarily by
comparison with scores on other tests ,
Thus, the Draw-A-Man Test may only generally indicate the
likelihood of a child scoring well on another test, and nnay be invalid as
a predictor of potential performiance in training programs. In sumnnary,
the social service practitioner should regard this test with caution for
it only roughly estinnates intellectual ability. Other nneasures of general
intellectual ability should supplement it.
Other Drawing Tests
Although almost every art media, technique, and type of subject
matter has been investigated in the search for significant diagnostic
clues, special attention centers on drawings of the human figure. The
Machover Draw-A- Person Test is a well-known example. In this test,
the examiner provides the subject with a letter-size sheet of paper and a
medium-soft pencil and tells him/her to simply "draw a person," or — to
young children — "draw somebody" or "draw a boy or girl." Upon comple-
tion of the first drawing, the examiner asks the subject to draw a person
of the opposite sex from the first figure. While the subject draws, the
examiner notes comments, the sequence in which parts are drawn, and
56
other procedural details. An inquiry nnay follow this drawing in which
the subject is asked to make up a story about each person drawn "as if
he were a character in a play or novel." During the inquiry a series of
questions elicits specific information about age, schooling, occupation,
family, and other facts associated with the characters portrayed.
Qualitative judgements, involving the preparation of a composite
personality description from the analysis of the many features of the
drawing, and considering the absolute and relative male and fennale figures'
size, their position on the page, quality of lines, sequence of parts drawn,
stance, front or profile view, position of arms, depiction of clothing,
and background and grounding effects — all of these make up the scoring
of this test. Omission of bodily parts, disproportions, shading, amount
and distribution of details, erasures, symmetry and other stylistic
features produce special interpretations. Each major body part,
such as head, individual facial features, hair, neck, shoulders, breast,
trunk, hips, and extremities, is regarded as significant.
The interpretive guide to the Draw-A- Person Test contains sweeping
generalizations, such as "disproportionately large heads will often be
given by individuals suffering from organic brain disease" or "the sex
given the proportionately larger head is the sex that is accorded more
intellectual and social authority." But no evidence supports these
statements. The guide also refers to a "file of thousands of drawings"
examined in clinical context and a few selected cases are cited for
57
illustrative purposes, but no systematic presentation of data acconnpanies
the published test reports .
Validation studies of this test by other investigators have yielded
conflicting results. Attempts to develop semi-objective scoring procedures
which utilize rating scales or checklists have met with little success. The
test may succeed nnore with children and other relatively naive subjects
than with sophisticated adult groups. Although it appears to differentiate
between seriously disturbed persons and normals, its discriminative
value within relatively nornnal groups is questionable . Research on the
P raw- A- Person Test has been inadequate largely because of failure to
cross-validate .
The House-Tree- Person Projective Technique (H-T-P) devised by
Buck, has aroused considerable interest as witnessed by the number of
relevant research publications. In this test, the subject is told to draw
as good a picture of a house as possible, then the sanne for a "tree" and
a "person." Meanwhile, the exanniner takes copious notes on tinne,
sequence of parts drawn, spontaneous comments by the subject, and
expressions of ennotion. Oral inquiry, including a long set of standardized
questions, follows completion of the drawings. The examiner analyzes the
drawings both quantitively and qualitatively, chiefly on the basis of their
formal or stylistic characteKstics.
In discussing the rationale underlying the choice of objects to draw.
Buck maintains that "house" should arouse association concerning the
58
subject's home and those lived with; "tree" should evoke associations
pertaining to life goals and ability to derive satisfaction from the environ-
nnent in general; and "person" should call up associations dealing with
interpersonal relationships. Some clinicians may find helpful leads in such
drawings when considered jointly with other information about the individual
case. The elaborate, lengthy adnninistrative and scoring procedures
described by Buck appear unwarranted in light of the highly inadequate
nature of the supporting data.
Minnesota Multiphasic Personality Inventory (MMPI)
The design of the MMPI provides an objective assessment of sonne
of the major personality characteristics that affect personal and social
adjustment. The scales provide a measurement for the personality status
of literate adolescents and adults together with a basis for evaluating the
acceptability and dependability of each test record . Nine scales were
originally developed for the test's clinical use and were named for the
abnornnal conditions on which their construction was based. Since they
have proved meaningful within the normal range of behavior, these scales
are now referred to by their abbreviations — Hs (hypochondrasis),
D (depression), Hy (hysteria), Pd (psychopathic deviate), Mf (masculinity-
feminity). Pa (paranoia), Pt (psychoasthenia), Sc (schizophrenia), and Ma
(hypomania) — to avoid possible misleading connotations . Development of
these test items has produced a number of other scales: Si (social
59
introversion) is connnnonly scored, as well as three validating scales:
L (lie), F (validity), and K (correction). A wide variety of untrained
personnel can administer this inventory, however, a thoroughly trained
clinical or educational psychologist should interpret the results .
One can expect test subjects sixteen years of age or older with at
least six years of successful schooling to complete the MM PI without
difficulty. When an individual is specifically referred for testing, one
can generally ascertain beforehand whether the MM PI is appropriate for
use and thus avoid the embarrassment that would arise from failure during
the actual administration. The full-scale edition of the MM PI requires the
subject to give a true or false response to 566 separate questions (see
Table I). The raw scores thus obtained are converted to a kind of
standardized score called a T score on which the MM PI profile and code
are based . The test itenns are presented either in a card form for
individual use or in a booklet with a separate answer sheet for individual
ex amination or large-scale group testing programs. Such a profile
provides a scale for clinical comparison of the relative "strength" of various
personality trends.
Clinicians who use the MM PI usually tend to emphasize one
particular scale of the nine. The MMPI should not be evaluated on the
basis of one scale alone but rather on the pattern of scores for the
entire nine scales including the validity indicators. The test affords an
infinitely large number of patterns. Thus, although scorers nnay often
60
feel that they have seen some given pattern a number of times before,
alnnost no exact duplicates exist.
Although originally thought of as an aid to psychiatric diagnosis and
evaluation, the MMPI has been used in many different settings and
validated against hundreds of different criteria. The rapid rise of
these tests' non-psychiatric application has stimulated a substantial
growth in new scales and scoring procedures.
Reliability and validity research on the MMPI are not entirely
convincing. Validity studies do not show high correlations between
MMPI profiles scores and actual psychiatric diagnoses, although the
instrument was initially developed for this purpose .
Indeed, the available categories of psychiatric diagnoses are subject
to criticism since it is questionable whether or not the MMPI actually
achieves its intended objectives when used strictly clinically. But
where the MMPI is used to screen large populations, such as military
recruits, college students, or business executives, it serves as a
reasonably reliable, general screening device. It is most useful in
identifying those persons who achieve extreme scores on the subscales —
thus identifying those who require further study.
The use of the MMPI requires professionally trained, experienced,
and sophisticated practitioners, because of the complexity of the
personality characteristics of the inventory, the meanings of the scales,
and the way ip which the scales relate to each other in predicting behavior.
61
The original MM PI was standardized on a sample of about 700 nornnal
visitors at the University of Minnesota Hospital (ranging in age from 16 to
55 and representing a cross section of the Minnesota population) in
contrast to sonne 800 clinical cases (from the Neuro- Psychiatric Division
of the University of Minnesota Hospital).
The test/ re-test method determines the reliability for the MM PI .
Reliability results show that the coefficients of correlation vary consider-
ably with different subscales . The test/ re-test reliability coefficients
range fronn .46 to .93 with the majority lying between .70 and .90 —
a fairly high degree of reliability.
The predictive nnethod, which compares the scores obtained on
special scales with clinical diagnoses for newly admitted psychiatric
patients determined validity for the MM PI. In approximately 60 per-
cent of the cases this method predicted the corresponding clinical
diagnosis.
Sample test reports . The client's responses to the MM PI indicate
a dependent, immature, impulsive, demanding woman who attempts to
exploit and control others. She seems able to maintain relationships only
with those she can keep in subservient positions. Probably her fear
of abandonment creates this fear. Unfortunately, she seems not to
recognize the alienating effect of her domineering tactics. Her imperious
manner and her repeated demands will drive any away from her except
62
those even more emotionally disturbed than she. She is a very angry
woman who seenns especially resentful toward men . While she pretends
to heterosexuality, she may spend a great deal of her tinne trying to prove
this through sexual acting out, repeated love affairs, etc. , probably
because she has an annorphous sexual identity. While not psychotic,
apparently she is poorly controlled, disorganization-prone, moody, and
hypertensive. Her obesity is probably a function of anxiety. She eats to
ward off the loneliness and to control the gnawing ennptiness of feared
abandonment. She has a personality disorder, perhaps a passive-
aggressive personality of the aggressive type . She needs individual
psychotherapy and will probably not lose weight nor be able to stabilize
vocationally without this. She will benefit best fronn a reality-oriented,
problem-solving approach although she nnight make use of "insight."
She will probably have a stormy relationship with any therapist.
Minnesota Counseling Inventory
An effort to adapt the previously discussed Minnesota Multiphasic
Personality Inventory for use with normal high school students and college
freshmen led to the development of the Minnesota Counseling Inventory.
Many of the 355 true/false itenns of the latter inventory came from the MM PI,
and several other scales have a close resemblance to the MM PI scales.
With norms based on over 20,000 high school students tested in ten states,
this test provides scores in seven areas designated as: Family
63
Relationships, Social Relationships, Emotional Stability, Conformity,
Adjustment to Reality, Mood, and Leadership. The "Confornnity" scale
has a strong resemblance to the MM PI Pd scale and "Adjustnnent to
Reality" sinnilarly resennbles the Sc scale. Also, two vertification scores
exhibit similar traits to the MM PI validity scales. The comparison of
random samples of students with groups nominated by teachers as out-
standing examples of the quality tests, validated the total scores on the
different scales . Test reliability established by split-half and re-test
procedures is at an acceptable level. But the seven area scores are not
as distinct as their titles imply. Only counselors familiar enough with
its construction to evaluate its complex scores should use this inventory.
Otis Self-Administrating Test of Mental Ability
An early test that has been widely used in personnel screening on
a group basis is the Otis Self -Administrating Test of Mental Ability.
This test also helped to develop the basis for the highest level norms for
Otis Quick-Scoring Mental Ability Tests used as an academic screening
device from the early grades through high school level . Industry uses the
Otis Self-Administrating Test of Mental Ability for screening applicants
for such varied jobs as clerks, calculating machine operators, assembly
line workers, and foremen and other supervisory personnel. A number of
validation studies have checked the Otis against an industrial criterian,
most of which have demonstrated that the scores of the applicants compare
with actual job performance creating significant validity coefficients. In
64
semi-skilled jobs, the Otis Test correlates moderately well with success
in learning the job and ease of initial adaptation. It does not, however,
correlate highly with subsequent job achievement. This would be expected
for routine jobs, but also holds true for high-level, professional jobs
since it discrinninates poorly at these upper levels.
General Aptitude Test Battery (GATB)
The U.S. Ennployment Service produced this battery. Throughout
the country it helps to guide people seeking work . State employment
services give these tests as well as other non-profit agencies whose
personnel have been trained in the use of the test by the Employment
Service. High school juniors and seniors often take thenn through a
cooperative plan which makes the results available to both the high school
counselor and the employment service . Versions of the tests have been
prepared for a nunnber of foreign countries.
The Employment Service constructed the test to help guide persons
into suitable work. Each of the thousands of jobs in the modern industrial
world has its own aptitude requirements. When an employer asks for
referrals of potential employees, he wants applicants likely to succeed.
The U.S. Employment Service working with state agencies, therefore,
conducts studies of the psychological characteristics of particular jobs
and accumulates information on the meanings of a test score . The
following illustrates the small sample of the occupations studied:
assembler of dry cell batteries, aircraft electrician, teacher, x-ray
65
technician, nurses' aide, sheet metal worker, baker, cook, spot welder,
comptonnetor operator, corn husking nnachine operator, knitting-machine
fixer, food packer.
Predictions for such jobs takes us far beyond the academic ability
and reasoning ability which predominate nnost aptitude tests. The diversity
of occupations rules out the possibility of devising a separate aptitude test
for each job. For guidance, a limited number of diversified tests are
needed which everyone can take and which can be linked together in various
combinations to predict success in different situations. With this end in
view, the current GATB measures nine distinctive factors:
G - General reasoning ability (a composite of tests entitled
Vocabulary, Three-Dimensional Space, and Arithmetic
Reasoning)
V - Verbal aptitude (Vocabulary)
N - Numerical aptitude (Computation, Arithmetic Reasoning)
S - Spatial aptitude (Three-Dimensional Space)
P - Form perception (Tool Matching, Form Matching)
Q - Clerical perception (Name Comparison)
K - Motor coordination (Mark Making)
F - Finger dexterity (Assemble, Disassemble)
M - Manual dexterity (Place, Turn)
66
No other similar test exceeds the efficiency of the OATR . Each of
its paper-pencil tests takes about six minutes . The psychomotor tests
require even less working time but several minutes are used for demon-
stration practice. The entire battery can be given in two and one-quarter
hours. The simple procedures allow trustworthy administration of the
tests by relatively untrained testers to subjects who have limited education
or poor command of English. The psychomotor tests are designed so that
each subject leaves all the materials as they were found — ready for the
next subject.
This marked speeding of nearly all the GATB subtests may reduce
their validity for many purposes, especially if the person has some reading
deficit, is upset by tests, or has taken few tests. But since the U.S ,
Employment Service had access to workers in all areas of the country,
all types of industry and agriculture, and most occupational levels, it
could obtain a highly representative normal sample. It drew 4,000 cases
from the records on hand to form a group which properly represented
all occupational, sex, and age groups in proportion to census data.
Test results are reported as standard scores with a mean of 100 and
a standard deviation of 20. Extensive research has demonstrated good
reliability and validity. Validity does vary between and among different
occupations. The social service practitioner should use the GATB in
conjunction with the U.S. Employment Service's Dictionary of Occupational
Titles .
67
An example of a HATR test result is as follows: Mr. Smith's
scores on the Intelligence, Verbal Aptitude, and Numerical Aptitude
sections of the General Aptitude Test Battery indicate that his achieve-
ment is far above average in each of these categories. His Intelligence
score is at the 99th percentile as compared to the general working popula-
tion, while his Verbal and Numerical scores compared to the same population
are both at the 96th percentile. All of his scores on the remaining
aptitudes of the battery are at the 75th percentile or above including his
scores on the Manual and Finger Dexterity Form Board. It is apparent,
then, that Mr. Smith is intellectually capable of undertaking technical or
college training in any of the occupational aptitude patterns covered by the
GATB . That is, he has the intellect and dexterity necessary to handle
any of the many occupational categories listed fronn Occupational Aptitude
Pattern 1 through Occupational Aptitude Pattern 35 inclusive . His interest
profile from the other tests in the battery suggests that he may wish to
consider any of the following occupations listed in Occupational Aptitude
Pattern 1: physician, civil engineer, highway engineer, etc. Under
Occupational Aptitude Pattern 2, he may wish to consider training as a
pharmacist, cost accountant, tax accountant, or statistician. Appropriate
occupations from Occupational Aptitude Pattern 3 are teacher, survey
worker, group worker, or caseworker.
68
Strong Vocational Interest Blank (SVIB)
One of the most widely used interest tests is the Strong Vocational
Interest Blank, first published in 1927. The Strong contains questions on
hundreds of activities both vocational and avocational. Most of the 400
items require a "like-indifferent-dislike" response to activities or topics:
biology, fishing, being an aviator, planning a sales cannpaign, etc.
Because research has demonstrated that the majority of nnen in a particular
occupation have roughly similar interests, the Strong assumes that a
person having a typical occupational group pattern will find satisfaction in
that field .
The Strong determined the interest pattern for a profession by giving
the questionnaire to successful members of that particular profession and
by comparing the responses of the group with those of men of similar age
selected randomly from the whole range of occupations ordinarily entered
by college men. A weighted scoring key assesses how closely the subject's
interests correspond to those of the professional group. On each item, the
percentage of nnen-in-general for each answer was compared to the pei —
centage of nnen-in-the-occupation giving the answer. Engineers dislike
"actor" more commonly than other men; therefore, response D or dislike
is assigned a positive weight in the engineers' scale. "Liking to be the
author of a technical book" (a significant indicator of engineering interests)
is especially common among engineers, thus acquiring a weight of plus
three. In contr^ist, 40 percent of artists respond "like" to "actor." So
69
the artist's scale weights "actor" at plus two for like, zero for indifferent,
and minus 1 for dislike.
Occupational scores convert into letter grades ranging from A to C .
Seventy percent of successful men in the occupation fall into the A group on
that scale. Interests of a person who falls below B plus are quite different
from those of the bulk of the occupational group. Only 2 percent of the men
in the occupation fall as low as C.
The test has available a great number of keys for male occupations
and a woman's blank which can be scored for a number of occupations
typically entered by women . The Strong contains items varied enough
to predict almost anything, and a new key can be made for any vocational
or specialized interest group. Strong keys can score not only vocational
interests, but also it can score answers which men give more frequently
than women, for example, and create a "masculinity-feminity key."
Extensive research has demonstrated considerable predictive
validity for this instrunnent. Strong demonstrated that interest scores
obtained by college undergraduates predicted their occupations of eighteen
years later. His interest scales successfully differentiate nnembers of an
occupation from the population in general and the occupations from each
other. Given the amount of research on both the reliability and validity of
the Strong, it is reasonably assumed to be one of the best occupational
predictive instruments available .
70
However, caution seems indicated in interpreting the results for
both the Strong and the Kuderfor a number of studies have demonstrated
that examinees can fake these inventories: Examinees, told to attennpt
responding in a way that they thought life insurance salesnnen would,
generally succeeded in making themselves appear like life insurance
salesmen. In other words, if a person suspects what characteristics are
being screened, this person can fake a response. However, evidence does
not suggest that people in general fake their responses but that most
people are genuinely interested in their test outcomes. Below is an
example of a Strong test .
The results of the Strong Vocational Interest Blank indicate a
client highly interested in religious activities, social service, and music,
as well as public speaking, business management, art, teaching, mathema-
tics, technical supervision. His general interests show a similarity to
those men successful as music teachers and music performers, but also
sinnilar to those of credit managers, chamber of commerce executives,
business education teachers, social workers, YMCA staff members,
rehabilitation counselors, public administrators, physical therapists, and
librarians. Surprisingly, in view of his stated vocational aspirations, his
interests do not parallel those of computer programmers. With this in
mind, he must revise his planning. While connputer science is not an
altogether inappropriate career choice, the client would probably be
happier in a career more oriented toward administration and dealing with
the public. Connputer science may provide an opportunity for this.
71
especially if supplemented by general business and/or nnanagennent courses.
He nnay also wish to consider a business curriculunn, college, or trade
school . Counseling also appeals to hinn so various types of social work nnay
be feasible. But since he shows ardent interests in music, he should explore
the possibility of becoming a music teacher or performer.
Stanford-Binet Scale
The Stanford-Binet scales for measuring intelligence (since 1937
known as the Stanford-Binet Scale) has gone through several revisions, all
of them using a common principle: the average capacities of children of
various ages represent differences in degrees of brightness along with
differences in levels of development. Thus, knowledge of intellectual
performance levels of typical children of a given age facilitates comparison
with any specific child by comparing his/her score with the average. The
principal criterion employed by Binet and Simon in the standardization and
age-placement of tests was: any item successfully completed by two-thirds
to three-fourths of a representative age group of children of a given age was
designated as "average" performance for that age group, and their ideal
standard placed the test at a year-level passed by 75 percent in that age
group.
The following procedure describes the method of scoring the Stanford-
Binet Scale . The examiner selects a starting point in a range of tasks
where the subject can pass all items. This is called the "basal year."
The examiner then proceeds upward in the scale until the subject fails all
72
items, a level called the "terminal year." Each item carries specified
credit in terms of nnonths contributing to the mental age score . These
credits, added to the age value of the basal year, total the mental age.
For example, assume a basal year of six; then, three test items passed at
the seven-year level give additional credit of six nnonths, two passed at
the eight-year level give further credit of four months, but all failed at
the nine-year level. Thus, the subject's mental age is six years, ten
months .
The 1937 revision of the 1916 scale differs in many details from its
predecessor (unsatisfactory items were eliminated and new ones added),
but it shares the essential and basic conception , It has two equivalent
fornns, L and N, each of which contain 129 test items as compared with
the 90 items in the first Stanford-Binet . The 1937 scale extends down-
ward to the level of age two and upward through three levels of "superior
adult" (known as superior adult I, II, and III) thus increasing its usefulness.
From the ages of two through five, this scale provides groups of
test items at half-year intervals and thus obtains more accurate and highly
differentiating test results. The half-yearly intervals are possible because
the mental growth rate proceeds most rapidly in the earlier years creating
more rapid periodic increments susceptible to testing.
Although the 1937 scale like that of 1916 relies predominantly on its
verbal character, it does provide performance and other non-verbal materials,
especially through the age of four years. Performance materials demand
73
the subject to do something — build a pattern or make a design with blocks
or fill in a form built with variously shaped blocks. Other non-verbal
materials include such activities as copying a geometric figure, completing
the picture of a man, discriminating between forms, etc. In all these, the
child must use verbal ability inasmuch as he/she must understand verbal
directions. In these tests, verbal ability can also be a factor if the child
knows the names of the objects or geometric figures and this knowledge
helps the manipulation or classification of them.
Since the 1937 scale was standardized on only Amierican-born,
white, primarily urban subjects, it is also extremely verbal and thus
additionally culturally loaded. Though this test is still used with children,
the Wechsler Intelligence Scales have largely replaced it,
Vineland Social Maturity Scale
This scale, designed for use with individuals from infancy to the age
of thirty years, models itself on the construction and standardization of
the Stanford-Binet scale .
Unlike nnany other scales, this one is based upon a well-defined
rationale and systennatic construction. It groups behavior items at
age levels as in the Stanford-Binet; these items represent progressive
maturation and adjustment to the environment in the following categories:
Self-help - reaches for nearby objects (age 0-1)
Self-direction - buys own clothing (age 15-18)
Locomotion - walks about room unattended (age 1-2)
74
Occupation - helps at little household tasks (age 3-4)
systematizes own work (age 25 plus)
Connmunication - makes telephone calls (age 10-11)
Socialization - demands personal attention (age 0-1)
advances general welfare (age 25 plus)
Examiners score items after interviewing someone well-acquainted
with the subject or the subject himself. Then, a social age is obtained
which is divided by chronological age, yielding a social quotient (S.Q.).
Although this social maturity scale highly correlates with intelli-
gence test results (about .80), the author maintains that its content and
rated function are distinct enough for use in the study of an individual's
general behavioral development, since social age provides a procedural
basis to guide the care and training of an individual .
While the scale aids in diagnosing the normal population as well as
the mentally deficient, it was first conceived to facilitate the diagnosis of
mental retardation. Primarily it differentiates between mentally retarded
individuals who can conduct their personal and social life with greater
independence and the mentally retarded who are socially inadequate .
Clinics widely use the Vineland Scale with children and adolescents.
And, in addition, it is a valuable device for interviewing and counseling
both parents and children.
75
Thematic Apperception Test (TAT)
Commonly referred to as the TAT, this projective personality
test consists of thirty picture cards plus one blank card. An examiner
uses the cards in various combinations depending upon sex and age; some
are used with all subjects and others with only one sex or age group.
The examiner uses only twenty total pictures with any subject which are
usually administered in two test sessions, ten pictures at a tinne.
Examinees are told that the TAT tests imagination. They are to
make up stories to suit themselves and are assured no right or wrong
responses exist. The examiner shows pictures one at a time, giving
simple instructions and asking the subject:
1 . to tell what he/she thinks led up to the depicted scene, how
it came about;
2 . to give an account of what is happening and the feelings of
the characters in the picture; and
3 . to tell what the outcome will be .
The test has no time limits and an examiner encourages the subject to
continue for as long as five minutes on a picture . Sometimes the
examiner uses an interview to learn the origins of the stories, especially
associations to places, names of persons, dates, specific and unusual
information are sought. This is an important aspect of the process because
it enables the examiner to clarify stories' meanings. For instance, a
boy ten years of age made up a surprisingly large number of stories
7G
dealing with death. The interview revealed these as nornnal responses:
his father was an undertaker and they lived above the funeral parlor.
Although the TAT uses pictures more structured than an inkblot,
they possess enough ambiguity to allow wide latitude for individual
differences in responses. The TAT is, like the Rorschach, a projective
method. Murray designed the TAT to elicit "drives, sentiments, and
conflicts" by analysis of the story produced by the subject. He based the
test upon the principal that when interpreting an ambiguous social situation,
one is apt to reveal aspects of one's own personality that would not or
could not be admitted otherwise because they are unconscious . The
subject, while absorbed in the picture and attempting to construct an
appropriate account of it, is off guard and becomes less aware or quite
unaware of himself/herself in the situation. In creating stories based
upon somewhat vague pictures, the subject utilizes and organizes content
of unique personal experiences. The examiner regards everything the
subject says as having meaning. From these stories, the skilled
examiner/interpretor draws inferences regarding the subject's person-
ality traits and their organization. The limitations of other projective devices
also limit the TAT . A number of different elaborate and special-purpose
schemes allow scoring of the TAT, but they show little uniformity in pro-
cedure for analysis of test results and few clinicians report the specific
system in use. Thus, comparisons between examiners are often impossible.
77
Unless one of two objective, specific scoring systems is used along with
a specially trained scorer, reliability for the TAT is generally low.
Validity research has not demonstrated the TAT's practical use.
It helps little in predicting behavior and thus is of little value in decision
nnaking. However, it has been useful for research in achievement motiva-
tion . Below is an example of the TAT .
The client's responses to the TAT indicate a chronically anxious,
impulsive person who becomes flighty, disorganized, and hypernnanic
under stress. He avoids close relationships because he can relate only
in a superficial, exploitive way. He wishes those stronger than himself
would take care of him and thus he may go to rather great lengths to nnake
people he sees as superior notice him. He has a negative, poorly defined
identity. He feels alone, helpless, and unable to function without high
anxiety unless involved in a constant frenzy of activity. His feminine
interests equip him little to compete with more aggressive peers. While
he is not necessarily an overt honnosexual, he may be prinnarily homo-
erotic in his sexual responses. He fears exploitation and attack. He is
afraid of failure and so may not see tasks through to their conclusion. He
has many personality deficits and functions in a way which will interfere
with constructive achievennent in a vocational training program. In fact,
his enrollment in a training program should probably be made contingent
upon regular psychological treatment. He will respond best to supportive,
problenn-solving approach and behavior modification techniques ennphasizing
reward for constructive efforts.
78
Symonds Picture Story Test (SPST)
The Symonds Picture Story Test is a projective technique designed
for the study of the personality of adolescent boys and girls. The SPST
is identical to the Thematic Apperception Test except that it uses a
different set of pictures especially designed for the study of adolescent
fantasy. But, the SPST similarly uses twenty pictures divided into two
sets. If both sets are used, the examiner should use one set at a first
setting and the second set at a second setting at least a day later (usage
has demonstrated Set B the more effective of the two). The examiner
individually administers the test in an interview situation which requires
about an hour to run through the ten pictures . The author recommends
interpreting the results of the test within the context of the subject's life
history nnaterial secured by casework with psychoanalytic study.
Many of the limitations inherent in the use of any projective device
affect this test, and the connments about the use of the Thematic Apper-
ception Test apply fully to the SPST — with the additional observation that
the SPST has not been subject to as nnuch research as the TAT. Normative
data based on forty cases are available in the manual .
Wechsler Intelligence Scale for Children (WISC)
Examiners frequently use the WISC — an individually adnninistered
general intelligence test — to predict academic success or discover
intellectual or academic deficiencies which may be interfering with school
achievement. Like the Wechsler Adult Intelligence Scale, the WISC
79
obtains I.Q.s by comparing each subject's test penfornnance with the scores
earned by individuals in his or her age group. I.Q.s obtained by successive
yvISC re-tests always compare the subjects to their own age group at each
time of testing. Each person tested is assigned an I.Q. which represents
the intelligence rating relative to his or her age. The WISC uses a mean
of 100 and a standard deviation of 15, and places I.Q.s from 90 to 110 in
the average range. In terms of percentile limits, the highest 1 percent
would have I.Q.s of 135 and above, and the lowest 1 percent I.Q.s to 65
and below. The middle 50 percent of children in each age will have I.Q.s
ranging fronn 90 to 110.
The WISC consists of 12 subtests which, like the adult scales, divide
into two subgroups identified as verbal and performance. The verbal sub-
tests are: Information, Comprehension, Arithmetic, Similarities,
Vocabulary, Digit Span; the performance subtests are: Picture Completion,
Picture Arrangement, Block Design, Object Assembly, Coding, Mazes.
In the standardization of the WISC, every subject took all twelve
tests, but to shorten the time required for examination the scale has been
reduced to ten tests. (The Digit Span in the verbal, and Mazes in the
performance part were omitted primarily on the basis of their relatively
low correlation with the other tests on the scale, and, in the case of Mazes,
the time factor.) One can use all subtests but in this case all twelve tests
must be prorated before computing the I.Q.s. Usually a trained clinical or
school psychologist administers the test.
80
An original norming group of only white urban children sampling
1 , 100 girls and 1 , 100 boys in 11 age groups standardized the WISC. Its
present limitations as a diagnostic instrument for intelligence probably
lie in these inadequate sampling procedures — how can the WISC validly
test those who radically depart from those in the original sampling? The
onnission of rural American Indian children in the original sampling
procedures sharply limits this test's usefulness with them. Shifts in
population distribution, general levels of education, and increased use of
information dispersennent by nnass media means may further invalidate
WISC results on present day populations.
Though I.Q. tests can and will be adnninistered to so-called disad-
vantaged groups, the social service practitioner must remember to interpret
the results with great caution, as most items on these tests are culturally
biased in one way or another. Rennember: These tests mininnally estimate
"intellectual ability" and the results should be supplemented by intelligence
tests when the nneaning of the I.Q. test is vague. Use special care in
making future predictions on the basis of I.Q. tests alone — especially with
young children. Research suggests that the I.Q. does change under certain
conditions. Below is an exannple of a WISC test.
Jimmy Jones is functioning in the bright normal range of intelligence .
On the WISC, he achieved a Verbal I.Q. of 112, a Performance I.Q. of 115,
and a Full Scale I.Q. of 114. Although Jimmy's relatively high scores on
General Comprehension, Similarities, and Picture Arrangement indicate
81
that he has considerable abstracting ability and that his intellectual potential
is quite high, the discrepantly low scores on Arithmetic, General Infornnation,
and Vocabulary indicate that he has not been able to nnake the nnost of his
intellectual potential. Judging from his history and present living circum-
stances, the discrepancy is probably due to the effects of severe cultural
deprivation. The relatively low score on Digit Span also suggests a signifi-
cant level of anxiety which may be indigenous to test situations. This often
appears in children from culturally deprived environments and, of course,
adds to the type of school under-achievennent that nnay be reflected in his
low scores on the Verbal subtests correlated with such achievennent .
Certainly, his high scores on Picture Completion, Block Design, and Object
Assembly indicate extremely good perceptual-nnotor functioning. This lends
strength to the impression of intellectual functioning that is substantially
higher when measured by his overall performance on the Verbal subtests.
Probably, his potential lies well within the superior range of intelligence
(I.Q. = 120-130).
Wide Range Achievennent Test
The Wide Range Achievement Test first standardized in 1936 and revised
most recently in 1965, consists of three subtests, each divided into two
levels — Level I designed for children between the ages of five years and
eleven years and eleven months, and Level II designed for persons from
twelve years to adulthood. The three subtests at both levels are:
82
1 . Reading - recognizing and nanning letters and pronouncing words.
2. Spelling - copying marks resembling letters, writing the name,
and writing single words to dictation.
3. Arithnnetic - counting, reading nunnber symbols, solving oral
problems, and performing written connputations .
Untrained school personnel can administer this test to large groups.
The Wide Range Achievement Test has proved valuable in a number of areas:
1 . the accurate diagnosis of reading, spelling, and arithmetic
disabilities in persons of all ages;
2. the deternnination of instructional levels in school children;
3 . the assignment of children to instructional groups progressing
at similar rates and their transfer to faster or slower rates
in keeping with individual learning rates;
4. the establishment of degrees of literacy and arithmetic pro-
ficiency of mentally retarded persons;
5. the checking of school achievement of adults referred for
vocational rehabilitation and job placement;
6. the selection of personnel at various occupational levels for
pronnotion in business, industry, and the National Services; and
7. the selection of students for professionalized technical schools.
Test scores are reported as grade rorms and standard scores.
Originally the actual mean grade levels of the children in each age group
tested established the grade norms. Such an arbitrary score as grade
83
rating may vary with promotion practices and socioeconomic levels. For
example, in 1936, the average person in the nornn group obtained a 9.1
grade rating at age 17, but in 1963, the average person obtained a grade
rating of 10.8 at the age of 17. This may mean that more people stay in
school longer, or that more persons obtain higher grade ratings but not
necessarily higher achievennent than they did 25 years earlier. Despite
these variations, the grade ratings tend to be a rather stable score. The
comparability of the old and new grade ratings are striking through nearly
all educational levels except the upper ratings. The grade ratings above
age 14 are more arbitrary than those below 14.
The standard score in the Wide Range Achievement Test connpares
to the I.Q. of standard tests. Persons of different ages may receive
identical grade scores. For exannple, a 5.5 grade stands for superior
achievement if obtained by a 7-year-old, but represents defective achieve-
ment if obtained by a 25-year-old person. The standard score shows whether
the grade rating lies above average or below average for any particular age
level . The standard scores used are based on the distribution of the grade
ratings for each group.
While the Wide Range Achievement Test provides a useful measure of
actual achievement, the social service practitioner should not use it alone
for it contains multiple and varying reasons for underachievement . For
this reason, as with any other test, a lone achievennent test score may mis-
lead especially if interpreted by those unfanniliar with the complexities of
achievement testing.
84
Bender-Gestalt
The Bender-Gestalt was designed to test visual nnotor performance
skills. It is used as an aid in assessing perceptual-motor coordination
disorders which are often related to organic brain dysfunction. The test
has proven somewhat useful in diagnosing various types of retardation, and
personality patterns or trends .
The Bender-Gestalt consists of nine different geometric figures,
printed on cards, which the subject is asked to reproduce. This basic
procedure, called the "copy phase," is in some adaptations of the Bender-
Gestalt . For example, the examiner may ask subjects to recall the figures
they drew, elaborate upon or change the figures they reproduced.
While an extremely experienced clinician may profitably use this test,
it does not lend itself well for "cook book" interpretation or for use by
unexperienced examiners, nor should it be used in making final determina-
tions regarding organic brain dysfunction, perceptual -motor deficits, or
personality nnalfunction . A visual-motor task, in testing personality
reaction, provides a sannple of behavior involving connplex functions.
Like other so-called "projective" procedures, such complex behavior
examples are best interpreted from a consistent theoretical frame of
reference. The Bender-Gestalt results only hint at possible brain disorders
or personality malfunctions. The theory underlying personality testing
through a visual-motor task is that such testing has some special
characteristics and possible advantages. The theory notes that probably
85
styles of perceiving and reproducing figures which are relatively neutral,
i.e. , have few associations with one's past, tap sonne personality facets
which conscious attempts cannot disguise . A few highly skilled clinicians
piece together some good hunches about a subject from the drawings, but
since they rely on hunches the Bender-Gestalt remains an experimental
instrument, and validation studies prove disappointing despite extensive use.
Most social service practitioners would not find Bendei — Gestalt test results
useful in decision making. After all, how does the social worker use a
a test report indicating "suspicion of organicity"? Thus, regard this tests'
results with caution. Below is an example of a Bender-Gestalt test result.
This is a record of a 15-year-old girl who is quite inhibited and
generally fearful in her behavior. There is evidence of some mild, diffuse
organic damage probably occuring in early childhood, perhaps between 4 and
7 years of age, and possibly due to an encephalitic condition. Although she has
partially compensated for the intracranial damage, the organic factor still
exerts something of a handicap to her adjustment. However, at present,
her central problem appears to be neurotic inhibition accompanied by some
depression and apathy. The organic factor certainly contributes to her
present adjustment difficulties but is insufficient to explain them. At present,
her prime means of defense are withdrawal, denial, and isolation. She is
quite fearful of rejection in general and is especially fearful of rejection
at the hands of those she perceives as authority or parental figures. While
she usually tries to conform on a conscious level, she shows fairly
86
pronounced passive, oppositional tendencies. Under stress, there nnay be
some regression to narcissism and orally dependent behavior. Despite
this, she shows some progression and has established some behavioral
configurations characteristic of both anal and oedipal periods of adjustment.
She desperately needs closeness with people but is fearful of interpersonal
relationships and has not developed skills for encouraging these. Rather,
she remains in superficial, rather distant relationships, while embellishing
these with fantasy. Presently she is moderately depressed, partly because
she does not get the attention she needs and partly because of guilt over
impulses which she ordinarily inhibits. At this point, she seems to be
especially fearful of heterosexual relationships and may suffer fronn
unresolved sexual feelings for her father. Although her major identification
is female, her self-image is that of an inadequate person. Characteristically,
she remains withdrawn, aloof, and rather "retarded" in her behavior. An
estimate of her intellectual abilities nneasured by her present Bender
performance would yield an I.Q. of approximately 75. This rough estimate
probably characterizes her present school performance but is not a good
reflection of her potential. Despite her attempt to cooperate on the test,
there is considerable evidence of marked impairment of intellectual
functioning on the basis of neurotic problems . If her neurotic difficulties
were resolved, and her severe inhibitions rennoved, she should probably
function within or near the average range of intelligence .
87
Rorschach
The Rorschach test, or "Inkblot," originally developed in 1921
by Hermann Rorschach, has been considerably researched to expand and
improve upon its diagnostic virtues and uses. It is used primarily as a
personality test based on the "projective" method. The test consists of
ten inkblots presented one at a time to the subject. The first seven blots
are essentially black and white although blots two and three have smaller
red blotches. The last three blots are multi-colored. Typically, the test
is individually administered in three phases. During the first, the subject
gives spontaneous responses to the inkblots. During the second, or "inquiry"
phase, the examiner asks questions to determine how "the characteristics"
of the inkblots triggered the subject's response, such as if color, shape,
or shading helped the person see what he or she saw. In a later phase
(sometimes used), called "testing the linnits," the examiner attempts to
get additional scoring material, especially if the subject has given extremely
unusual responses or has not "seen" the concepts commonly "seen."
The Rorschach scoring systenn allots five scores to a response. The
scoring is deternnined according to the "area chosen," the content chosen,
the form level of the response (this refers to how accurately or arbitrarily,
or how definitely or vaguely the response fornn is conceived), and the
"popularity" of the response (whether or not the response is found often or
considered extremely rare).
88
In addition to the Rorschach's complex scoring, a qualitative
approach can also add further data. The way a subject approaches the
card, the pauses, the difficulties, the apparently extraneous comments
can all add further data when interpreted by a skilled clinician.
The theory implies that persons' reactions to these abstract blots
will give clues to their reactions to life — one person organizes the blot in
minute detail while another may give it a slap dash once over. Is the
subject interested in unusual details or in the more common ones? Are
colors perceived and reported in the blot description? And so forth.
Although the Rorschach has been used for over fifty years and has
been extensively researched to establish its predictive validity, the
results have proven somewhat disappointing and uneven. In research
experiments, clinicians asked to make a diagnosis based on Rorschach
responses alone, without any other data available, could not accurately
predict behavior nor diagnose psychiatric disorders. The research
indicates that the Rorschach is fickle — sometimes it works while other
tinnes it does not. In general, most clinicians agree that the Rorschach
has some predictive validity which does better than chance alone. But
experienced exanniners do equally well by asking subjects direct questions.
However, apparently no empirical evidence demonstrates that the Rorschach
or any other projective instrument will reliably predict behavior in the day-
to-day world. Thus, as a tool in making decisions on practical problems,
the instrument \s limited.
89
A number of split-half and test/re-test reliabiUty studies are
available on Rorschach protocals. Reported values differ considerably
fronn study to study and for different types of subscores. However, in
general, the use of a specific scoring system produces unifornnly positive
and fairly high correlations in nnany cases.
The Rorschach, as other projective tests, is a clinical instrument
that should give reliable, valid results only when used by persons having
both the special technical training and an advanced sophistication in the
understanding and application of a specific personality theory. The tests
are generally time consuming, both to give and to score, and they are
sometimes hard to justify by the results obtained. It is the author's
impression that a highly experienced clinician, willing to engage the
subject in in-depth interviews, can obtain similar results and perhaps
make nnore meaningful inferences regarding behavior prediction than the
Rorschach. Below is a sample test report on the Rorschach .
The Rorschach results indicate that the client is a relatively well
adjusted man whose personality conflicts are not sharp enough to cause
hinn chronic or severe anxiety. Indeed, the anxiety he experienced in
the present testing situation is probably germane to those circunnstances
in which he feels he is being evaluated or judged. He fears being found
in need — not an unrealistic fear in our society. He is overly concerned
about his physical functioning at present, probably due to his history of
physical disorder. A sensitive young man responsive to the needs of
90
others, he is motivated to improve himself and has the impulse control
necessary for ennotional and intellectual growth. He seems to get along
well with others and will probably work hard to do a good job in anything
he undertakes. He likes to make friends and seenns something of an extro-
vert. He will probably be popular with his peers, co-workers, etc.
Wechsler Pre-School and Prinnary Scale of Intelligence (WPPSI)
Published in 1967, this scale is designed to test the intelligence of
children fronn ages four to six and one-half years . The scale includes
eleven subtests, ten of which determine the I.Q. score. Eight of the
subtests are downward extensions and adaptations of the Wechsler
Intelligence Scale for Children (WISC). The other newly constructed
three replace WISC subtests that, for a variety of reasons, proved
unsuitable . As in the WISC and the WAIS , the subtests group into Verbal
and Performance scales from which Verbal, Perfornnance, and Full Scale
I.Q.s are found. However, in order to enhance variety and to help main-
tain the young child's interest and cooperation, the adnninistration of
Verbal and Performance subtests are alternated in the WPPSI . Total
testing time ranges from 50 to 75 nninutes in one or two testing sessions.
Abbreviated scales or short forms of the scale are not recommended . The
subtests include Information, Vocabulary, Arithmetic, Similarities,
Comprehension, Sentences, Animal House, Picture Completion, Mazes,
Geometric Design, and Block Design. "Sentences" is a memory test
substituted for the WISC Digit Span. The child repeats each sentence
91
immediately after the exanniner orally presents it. This test can be
alternatively used for one of the other Verbal tests; or it can be admin-
istered as an additional test to seek further information about the child
and so it is not included in the total score for calculating the I.Q.
"Animal House" is basically similar to the WATS Digit Symbol and the
Wise Coding test. A key at the top of the board has pictures of a dog,
chicken, fish, and cat, each with a differently colored cylinder (its
"house") under it. The child should insert the correctly colored cylinder
in the hole beneath each animal on the board. Time, errors, and omissions
determine the score. "Geometric Design" requires the copying often
simple designs with a colored pencil .
The WPPSI was standardized on a national sample of 1 ,200 boys
and girls in each of six and one-half year age groups from four to six
and one-half, where each child was tested within six weeks of the required
birthday or mid-year date. The sample was stratified against 1960 census
data with reference to geographical region, urban-world residence, pro-
portion of whites and non-whites, and father's occupational level. Raw
scores on each subtest are converted to normalized standard scores with
a mean of ten and a standard deviation of three within each one-fourth
year group. The sum of the scaled scores on the Verbal, Perfornnance,
and Full Scale are then converted to deviation I.Q.s with a mean of 100
and a standard deviation of 15. Although Wechsler advises against using
nnental age scores because of their possible misinterpretation, the
92
manual provides a table for converting subtest raw scores to "test ages"
in one-fourth year units .
Reliability coefficients for the Full Scale I.Q. are acceptably high
and consistent \Aath the other Wechsler scales. The manual also provides
tables for evaluating the significance of score differences. This data
suggests that a difference of fifteen points or more between the Verbal
and Performance I.Q. is significant enough to be investigated. Stability
over time was also checked in a group of fifty kindergarten children
re-tested after an average interval of eleven weeks. Under these condi-
tions, the reliability coefficients for the Full Scale I.Q. , the Verbal I.Q. ,
and the Performance I.Q. were satisfactorily high.
The manual reports comparisons with the Stanford-Binet and the
Wise. Along with the WISC. the Stanford-Binet correlates higher with
the Verbal I.Q. (.76) than with the Performance I.Q. (.66). This group,
which was sonnewhat below average in ability, had approximately the
same mean on Stanford-Binet and the WPPSI (91 .3 versus 89.6), Similar
comparisons at different ability levels are needed ,
Owing to its recent publication, little can be concluded at this time
about WPPSI's validity and practical use. The procedures followed in
standardizing the scale and estimating reliability and validity are of
uniformly high technical quality. Both the size and composition of the
norm and sample are considerably advanced over the pre-school tests
previously available. But observe caution when using any test score
93
involving young children, for many variables, difficult to control, affect
the uncertain procedure of assessing young children.
Peabody Picture Vocabulary Test
The Peabody Picture Vocabulary Test, designed to provide a verbal
intelligence estimate through nneasuring hearing vocabulary, is effective
with average subjects, and has special value with certain other groups.
Reading is not required, so the scale is especially fair for non-readers,
and since responses are non-oral , the test can be used for the speech
impaired (expressly the aphasic and the stutterer). It is also used with
certain autistic, withdrawn, and psychotic persons. Since neither point-
ing nor oral responses are required the test can be used with orthopedically
handicapped and cerebral palsied persons, and also with some visually
handicapped and perceptually impaired persons. Thus, the scale allows
for any English-speaking resident of the United States between two years,
six months and eighteen years who can hear words, see the drawings,
and indicate "yes" or "no" in a manner which connmunicates . The Peabody
Picture Vocabulary Test has a number of advantages:
1 . The test has high interest value and thus establishes good
rapport.
2. It needs no extensive, specialized preparation for its administra-
tion.
3. It is quickly given in ten to fifteen minutes.
94
4. Scoring is completely objective and quickly acconnplished in
one to two minutes.
5. It is completely untimed and thus is an ability rather than a
speed test.
6. No oral response is required.
7. Alternate forms of the test are provided to facilitate repeated
measures.
8. The test covers a wide age range.
The administration of the Peabody Picture Vocabulary Test requires
no special preparation other than complete fanniliarity with the test
materials which include giving the test prior to its use as a standardized
measure. The exanniner must know the correct pronunciation of each of
the test words as given in Webster's New Collegiate Dictionary. If all the
instructions are completely observed, psychologists, teachers, speech
therapists, physicians, counselors, and social workers should be able to
give the scale accurately.
Only ten to fifteen minutes are usually required for this untimed test.
The scale is adnninistered only over the critical range of items for a
particular subject. The starting point, basal, and ceiling vary from
testee to testee . The examiner presents a series of pictures to each
subject. There are four pictures to a page and each is numbered. The
examiner says a word describing one of these four pictures and asks the
subject to point to or tell the number of the picture which the word
95
describes. Subjects are encouraged to "guess" if they do not know which
picture best conforms to the nneaning of the word presented. The exanniner
starts subjects at different "picture levels" according to the age ranges
specified in the manual, and proceeds forward from the starting point until
the subject makes the first error. If the subject does not make eight
consecutive correct responses prior to this first error, the examiner
returns immediately to the starting point and works backwards (through
the next lowest age range) until a total of eight consecutive correct answers
are made by the subject. Responses above the starting point — as well as
below — are counted in order to establish the basal of eight consecutive
correct answers. The exanniner then continues testing forward from the
point of the first error until the subject makes six errors in any eight
consecutive presentations, counting the last itenn presented as the subject's
ceiling. The total score is the number of correct responses. All items
below the basal point are assumed correct; all itenns above the ceiling
item are assumed incorrect. To get the total raw score, the examiner
subtracts the errors from the number of the last item presented, or
ceiling item. By using special tables, the raw score can be converted to
three types of derived scores:
1 . an age equivalent (nnental age);
2 . standard score equivalent (intelligence quotient); and
3. a percentile equivalent.
96
The age norms for converting raw scores on the Peabody Picture
Vocabulary Test to nnental age scores are given in the manual . Age
equivalents supposedly provide an index of the level of a given subject's
development. For example, 75 is the mean raw score on Form A for
children who have a chronological age of 10.0. Therefore, regardless
of subjects' chronological ages, if they obtain a raw score of 75 on a
Peabody Picture Vocabulary Test, they supposedly possess a mental
age of ten years since their ability to score on this test is like the average
10-year-old's. Approximate grade equivalents derive from age equivalents
by the rule of five . Thus , a child with a mental age of 11.0 would have a
grade equivalent of six (subtract five from the mental age) indicating an
accumulative capacity to achieve at the beginning grade six level . Age
norms have a number of advantages. They provide an easily understood
index of the subject's developmental level. They are useful in comparing
mental age with chronological age, achievement age, social age, and so
on. In addition to the age norms, they provide standard score norms which
may provide an "index of brightness" for a given child in comparison with
other children of the same age. The Peabody was standardized with a mean
of 100 and a standard deviation of 15.
Wechsler Adult Intelligence Scale (WAIS)
The WAIS , the adult form of the Wechsler Intelligence Test, is used
to assess general and specific intellectual ability for persons sixteen
97
years and above. The WAIS consists of eleven subtests grouped into a
verbal scale and a perfornnance scale.
Verbal Scale
1 . Infornnation: Twenty-nine questions covering a wide variety
of information that adults presunnably should acquire in our
culture . An effort was made to avoid specialized or academic
knowledge .
2. Comprehension: Fourteen items in each of which the subject
explains what should be done under certain circunnstances,
why certain practices are followed, the meaning of proverbs,
etc. These are designed to measure practical judgment and
connmon sense . This test resennbles the Stanford-Binet
comprehension items but its specific content was chosen to
be nnore consonant with the interests and activities of adults.
3. Arithmetic: Fourteen problems similar to those encountered
in elementary school arithmetic. Each problem, orally
presented, is to be solved without the use of paper and pencil.
4. Sinnilarities: Fifteen items requiring the subject to say how
two things are alike .
5. Digit Span: Orally presented lists of three to nine digits to be
orally reproduced. In the second part, the subject must
reproduce backwards lists of two to eight digits ,
98
6. Vocabulary: Forty words of increasing difficulty presented both
orally and visually. The subject is asked what each word nneans ,
Performance Scale
7. Digit Synnbol: This is a version of a fanniliar code-substitution
test which dates back to the early Woodworth-Wells Association
Test and has often been included in non-language intelligence
scales. The key contains nine symbols paired with nine digits.
The subject's score is the number of symbols correctly
written within one and a half minutes.
8 . Picture Completion: Twenty-one cards , each containing a
picture with sonne part missing . The subject must tell what is
missing from each picture .
9 . Block Design: This test is reproduced in designs increasing
in complexity requiring from four to nine cubes. The cubes
or blocks have only red, white, and red-and-white sides.
10. Picture Arrangement: Each item consists of a set of cards
containing pictures to be rearranged in proper sequence so as
to tell a story.
1 1 . Object Assembly: This test includes a number of pieces to
be assembled very much in the manner of a jigsaw puzzle.
The subtest includes four pictures to be reproduced including
mannequin, hand, profile of a face, and side view of an
elephant.
99
Both speed and accuracy of performance are taken into account in
scoring A rithnnetic. Digit Synnbol , Block Design, Picture Arrangement,
and Object Assembly.
The WAIS standardization sample was carefully chosen to ensure
its representativeness. The principal normative sample consisted of
1 ,700 cases including an equal nunnber of men and women distributed over
7 age levels between 16 and 64 years. Subjects were selected to match
as closely as possible the proportions of the 1950 U.S. Census with
regard to geographic residence, urban-rural residence, race, white
versus non-white, occupational level, and education. At each age level,
one nnan and woman from an institution for mental defectives was included .
Supplementary norms for older persons were established by testing an
"old-age sannple" of 475 persons aged 60 years and over in a typical
mid-western city.
Raw scores on each WAIS subtest are converted into standard scores
with a mean of 10 and a standard deviation of 3. These scaled scores were
derived fronn a reference group of 500 cases which included all persons
between the ages of 20 and 34 in the standardization sample . All subtest
scores are thus expressed in comparable units. Verbal, Performance, and
Full Scale scores are found by adding the scaled scores on the six verbal
subtests, the five performance subtests, and all eleven subtests respectively.
The manual provides tables which convert these three scores to deviation
I.Q.s with a mean of 100 and a standard deviation of 15. However, such
100
I.Q.s are found according to the specific age group. Thus, they show an
individual's standing in comparison with persons of his or her own age
level. Deriving I.Q.s separately for each age level compares the indivi-
duals with the declining norm beyond the peak age. The age decrement
is greater in performance than verbal scores and also varies from one
subtest to another. Thus, Digit Symbol, with its heavy dependence on
speed and visual perception, shows the maximunn age decline. However,
on the other performance subtests speed may be an unimportant factor in
the observed decline. In a special study on this point, subjects in the old-
age sample were given those tests under both timed and untinned conditions.
Not only were the score differences under the two conditions slight but
the decrements from the 60-64 to the 70-74 age group were virtually the
same under timed and untimed conditions.
The WAIS has demonstrated consistently high reliability coefficients
through the split-half reliability technique. Validity has primarily been
established through demonstrating correlations between test scores and
scholastic achievement. The WAIS has also been compared to other
instruments for similarity in scores achieved by the same subjects. In
all respects the WAIS has demonstrated relatively high correlations . In
summary, the WAIS is perhaps the best general adult intelligence test
currently available . Following is a sample test report on the WAIS .
The client is functioning within the normal range of intelligence.
On the WAIS, she achieved a Verbal I.Q. of 92, a Performance I.Q. of
101
95, and a Full Scale I.Q. of 92 . The client's vocabulary Is snnaller
than average, she thinks in a slightly "scattered" way and has trouble
completing tasks that require concentration or the systennmatic organiza-
tion of intellectual nnaterial. There is great variability in her intellectual
perfornnance, in fact, and this is typical of the intellectual functioning of
those who experience severe anxiety. In this client's case, the results
are blocking, inattention to detail, nnild confusion, and diminished ability
to maintain cognitive set. She works better at structured and unambiguous
problems than she does at those requiring her to be organized or to work
out novel solutions. The degree of variability in her perfornnance suggests
that she would be functioning near the bright normal range had she had
better learning opportunities and were she not handicapped by chronic
anxiety and emotional difficulties.
Tests for Special Purposes
A variety of tests have been developed for a number of specialized
purposes . The following are examples of special purpose tests with
references to further information for the consumer.
The Culture Fair Intelligence Test. This is a paper and pencil test
developed by Cattell and Cattell, published by the Institute for Personality
and Ability Testing. The test is available for three different age levels,
ranging from children to adults. The test's purpose is to provide a
measure of ability directed at separating the evaluation of natural
intelligence from that contaminated or obscured by education . The
102
Culture Fair Intelligence Test used both the classical I.Q. with a mean
of 100 and a standard deviation of 24 and a standard score I.Q. with a
mean of 100 and a standard deviation of 16. The best research available
on the test indicates that when used in industrial countries sinnilar to
the United States the results have been consistent from country to
country. In very dissimilar countries, however, the results are signi-
ficantly different from those obtained with the standardization sample.
Extrenne caution is urged in interpreting the results of this test for
people who come from markedly different cultures. The Institute for
Personality and Ability Testing, Channpagne, Illinois, offers a manual
providing more infornnation about this test.
Tests for the orthopedic handicapped. The Pictorial Test of Intelli-
gence, available through Houghton-Mifflin Company Publishers, requires
neither nnanipulative nor speaking responses. It was designed to assess
the general intellectual ability of children between the ages of three and
eight and can also be used to test those children who are orthopedically
handicapped and cannot respond orally or in writing. The manual provides
information with regard to deviation I.Q. norms and mental age norms
and percentile norms. Thus, scores may be reported in all three forms.
Other tests which have been used with orthopedically handicapped include
the Progressive Matrices Test, the Peabody Picture Vocabulary Test,
and the Columbia Mental Maturity Scale .
103
Tests for the hearing handicapped . Several tests have been used to
assess the nnental ability of people who are hearing handicapped: the
Nebraska Test of Learning Aptitude, the Pintner- Peterson Performance
Scale, the Arthur Point Scale, and the Point Scale of Performance.
Tests for the deaf include the Point Scale of Performance Tests
available in two forms from C.H . Stoelting Co. and from the Psycho-
logical Corporation. Both are designed to test persons from five years
of age to adulthood. The purpose of the scale is to provide a measurement
of the intellectual ability of deaf children, children suffering fronn reading
handicaps, and non-English-speaking children. The test was standardized
on about 1 , 100 public school children from middle-class American homes.
Scores are reported in the form of mental age norms and a ratio I.Q.
Tests for the blind . Several standard tests have been adapted for
use with blind populations, including the Stanford-Binet and Wechsler
scales . The Interim Hayes-Binet Scale is composed of items in forms
L and M of the Stanford-Binet which do not require vision. Currently,
a special adaptation of the Wechsler scale is widely used for testing the
blind . The major adaptation of the Wechsler omitts the performance
subtests. The Haptic Intelligence Scale for Adult Blind is also available
which was designed to test blind adults aged sixteen and above. This test's
results are reported in the form of deviation I.Q.s with a nnean of 100 and
a standard deviation of 15, The test manual published by Psychology Research
in Chicago authored by Shurrager and Shurrager contains further infornnation .
The test is also described in Buros' Mental Measurements Yearbook.
CHAPTER VIII
HOW TO LEARN ABOUT SPECIFIC TESTS
Although detailed information is provided here about a number of
different tests, this is not intended to be a comprehensive reference guide
describing all of the large number of tests available. Following are some
standard references which social service practitioners might use to obtain
more infornnation about specific tests.
Mental Measurements Yearbook, Oscar K . Buros, editor, Island
Park, New Jersey: Gryphon Press, 7th edition, 2 volumes, 1972. Also
by the same author Tests in Print, Island Park, New Jersey: Gryphon
Press . Currently there are seven editions of the Mental Measurements
Yearbook, the latest published in 1972. The Mental Measurennents
Yearbook lists most of the published standardized tests in print as of
the year the book was printed. Those tests not reviewed in the earlier
editions are described and criticized by various authorities . The Tests
in Print book is a comprehensive test bibliography and index and provides
the following information about available tests: the name of the test, the
levels for which it is used, the publication date, specialized comments
- 1 04 -
105
about the test by various authorities, the number and types of scores pro-
vided, the authors, the publisher, and the reference to test reviews in
Mental Measurennent Yearbook.
Other good books on testing include:
Cronbach, Lee J . Essentials of Psychological Testing . New York:
Harper and Row, 1970.
Thorndike, Robert L. and Elizabeth Hogen. Measurement and
Evaluation in Psychology and Education, 3rd edition. New York: John
Wiley and Sons, 1969.
Robb, George, L. C. Bernardoni, and R. W. Johnson. Assessment
of Individual Mental Ability. San Francisco: Intext Ed. Publishers, 1972.
Berdie, Ralph, et . al . Testing in Guidance and Counseling. New
York: McGraw Hill Book Co. , 1963.
Other sources of infornnation are the test reviews and research in
professional periodicals. Journals such as Educational and Psychological
Measurement, The Journal of Educational Measurement, The Journal of
Counseling Psychology, and the Personnel and Guidance Journal typically
carry reviews of some of the more recent published or revised tests .
ft
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11