ADHERING TO FORMULA. It cannot be too strongly emphasized that unless we follow a standardized procedure the tests lose their significance. The danger is chiefly that of unintentionally and unconsciously introducing variations which will affect the meaning of the test. One who has not had a thorough training in the methods of mental testing cannot appreciate how numerous are the opportunities for the unconscious transformation of a test. Many of these are pointed out in the description of the individual tests, but it would be folly to undertake to warn the experimenter against every possible error of this kind.
Sometimes the omission or the addition of a single phrase in giving the test will alter materially the significance of the response.
Only the trained psychologist can vary the formula without risk of invalidating the result, and even he must be on his guard. All sorts of misunderstandings regarding the correct placing of tests and regarding their accuracy or inaccuracy have come about through the failure of different investigators to follow the same procedure.
One who would use the tests for any serious purpose, therefore, must study the procedure for each and every test until he knows it thoroughly. After that a considerable amount of practice is necessary before one learns to avoid slips. During the early stages of practice it is necessary to refer to the printed instructions frequently in order to check up errors before they have become habitual.
The instructions. .h.i.therto available are at fault in not defining the procedure with sufficient definiteness, and it is the purpose of this volume to make good this deficiency as far as possible.
It is too much, however, to suppose that the instructions can be made "fool-proof." With whatever definiteness they may be set forth, situations are sure to arise which the examiner cannot be formally prepared for. There is no limit to the mult.i.tude of misunderstandings possible. After testing hundreds of children one still finds new examples of misapprehension. In a few such cases the instruction may be repeated, if there is reason to think the child"s hearing was at fault or if some extraordinary distraction has occurred. But unless otherwise stated in the directions, the repet.i.tion of a question is ordinarily to be avoided. Supplementary explanations are hardly ever permissible.
In short, numberless situations may arise in the use of a test which may injure the validity of the response, events which cannot always be dealt with by preconceived rule. Accordingly, although we must urge unceasingly the importance of following the standard procedure, it is not to be supposed that formulas are an adequate subst.i.tute either for scientific judgment or for common sense.
SCORING. The exact method of scoring the individual tests is set forth in the following chapters. Reference to the record booklet for use in testing will show that the records are to be kept in detail. Each subdivision of a test should be scored separately, in order that the clinical picture may be as complete as possible. This helps in the final evaluation of the results. It makes much difference, for example, whether success in repeating six digits is earned by repeating all three correctly or only one; or whether the child"s lack of success with the absurdities is due to failure on two, three, four, or all of them. Time should be recorded whenever called for in the record blanks.
RECORDING RESPONSES. Plus and minus signs alone are not usually sufficient. Whenever possible the entire response should be recorded. If the test results are to be used by any other person than the examiner, this is absolutely essential. Any other standard of completeness opens the door to carelessness and inaccuracy. In nearly all the tests, except that of naming sixty words, the examiner will find it possible by the liberal use of abbreviations to record practically the entire response _verbatim_. In doing so, however, one must be careful to avoid keeping the child waiting. Occasionally it is necessary to leave off recording altogether because of the embarra.s.sment sometimes aroused in the child by seeing his answer written down. The writer has met the latter difficulty several times. When for any reason it is not feasible to record anything more than score marks, success may be indicated by the sign +, failure by -, and half credit by . An exceptionally good response may be indicated by ++ and an exceptionally poor response by --.
If there is a slight doubt about a success or failure the sign? may be added to the + or -. In general, however, score the response either + or -, avoiding half credit as far as it is possible to do so.
If the entire response is not recorded it is necessary to record at least the score mark for each test _when the test is given_. It must be borne in mind that the scoring is not a purely mechanical affair.
Instead, the judgment of the examiner must come into play with every record made. If the scoring is delayed, there is not only the danger of forgetting a response, but the judgment is likely to be influenced by the subject"s responses to succeeding questions. Our special record booklet contains wide margins, so that extended notes and observations regarding the child"s responses and behavior can be recorded as the test proceeds.
SCATTERING OF SUCCESSES. It is sometimes a source of concern to the untrained examiner that the successes and failures should be scattered over quite an extensive range of years. Why, it may be asked, should not a child who has 10-year intelligence answer correctly all the tests up to and including group X, and fail on all the tests beyond? There are two reasons why such is almost never the case. In the first place, the intelligence of an individual is ordinarily not even. There are many different kinds of intelligence, and in some of these the subject is better endowed than in others. A second reason lies in the fact that no test can be purely and simply a test of native intelligence. Given a certain degree of intelligence, accidents of experience and training bring it about that this intelligence will work more successfully with some kinds of material than with others. For both of these reasons there results a scattering of successes and failures over three or four years.
The subject fails first in one or two tests of a group, then in two or three tests of the following group, the number of failures increasing until there are no successes at all. Success "tapers off" from 100 per cent to 0. Once in a great while a child fails on several of the tests of a given year and succeeds with a majority of those in the next higher year. This is only an extreme instance of uneven intelligence or of specialized experience, and does not necessarily reflect upon the reliability of the tests for children in general. The method of calculation given above strikes a kind of average and gives the general level of intelligence, which is essentially the thing we want to know.
SUPPLEMENTARY CONSIDERATIONS. It would be a mistake to suppose that any set of mental tests could be devised which would give us complete information about a child"s native intelligence. There are no tests which are absolutely pure tests of intelligence. All are influenced to a greater or less degree also by training and by social environment. For this reason, all the ascertainable facts bearing on such influences should be added to the record of the mental examination, and should be given due weight in reaching a final conclusion as to the level of intelligence.
The following supplementary information should be gathered, when possible:--
1. Social status (very superior, superior, average, inferior, or very inferior).
2. The teacher"s estimate of the child"s intelligence (very superior, superior, average, inferior, or very inferior).
3. School opportunities, including years of attendance, regularity, r.e.t.a.r.dation or acceleration, etc.
4. Quality of school work (very superior, superior, average, inferior, or very inferior).
5. Physical handicaps, if any (adenoids, diseased tonsils, partial deafness, imperfect vision, malnutrition, etc.).
In addition, the examiner will need to take account of the general att.i.tude of the child during the examination. This is provided for in the record blanks under the heading "comments." The comments should describe as fully as possible the conduct and att.i.tude of the child during the examination, with emphasis upon such disturbing factors as fear, timidity, unwillingness to answer, overconfidence, carelessness, lack of attention, etc. Sometimes, also, it is desirable to verify the child"s age and to make record of the verification.
Once more let it be urged that no degree of mechanical perfection of the tests can ever take the place of good judgment and psychological insight. Intelligence is too complicated to be weighed, like a bag of grain, by any one who can read figures.
ALTERNATIVE TESTS. The tests designated as "alternative tests" are not intended for regular use. Inasmuch as they have been standardized and belong in the year group where they are placed, they may be used as subst.i.tute tests on certain occasions. Sometimes one of the regular tests is spoiled in giving it, or the requisite material for it may not be at hand. Sometimes there may be reason to suspect that the subject has become acquainted with some of the tests. In such cases it is a great convenience to have a few subst.i.tutes available.
It is necessary, however, to warn against a possible misuse of alternative tests. _It is not permissible to count success in an alternative test as offsetting failure in a regular test._ This would give the subject too much leeway of failure. There are very exceptional cases, however, when it is legitimate to break this rule; namely, when one of the regular tests would be obviously unfair to the subject being tested. In year X, for example, one of the three alternative tests should be subst.i.tuted for the reading test (X, 4) in case we are testing a subject who has not had the equivalent of at least two years of school work. In year VIII, it would be permissible to subst.i.tute the alternative test of naming six coins, instead of the vocabulary test, in the case of a subject who came from a home where English was not spoken.
In VII, it would perhaps not be unfair to subst.i.tute the alternative test, in place of the test of copying a diamond, in the case of a subject who, because of timidity or embarra.s.sment, refused to attempt the diamond. But it would be going entirely too far to subst.i.tute an alternative test in the place of every regular test which the subject responded to by silence. In the large majority of cases persistent silence deserves to be scored failure.
Certain tests have been made alternatives because of their inferior value, some because the presence of other tests of similar nature in the same year rendered them less necessary.
FINDING MENTAL AGE. As there are six tests in each age group from III to X, each test in this part of the scale counts 2 months toward mental age. There are eight tests in group XII, which, because of the omission of the 11-year group, have a combined value of 24 months, or 3 months each. Similarly, each of the six tests in XIV has a value of 4 months (24 6 = 4). The tests of the "average adult" group are given a value of 5 months each, and those of the "superior adult" group a value of 6 months each. These values are in a sense arbitrary, but they are justified in the fact that they are such as to cause ordinary adults to test at the "average adult" level.
The calculation of mental age is therefore simplicity itself. The rule is: (1) Credit the subject with all the tests below the point where the examination begins (remembering that the examination goes back until a year group has been found in which all the tests are pa.s.sed); and (2) add to this basal credit 2 months for each test pa.s.sed successfully up to and including year X, 3 months for each test pa.s.sed in XII, 4 months for each test pa.s.sed in XIV, 5 months for each success in "average adult," and 6 months for each success in "superior adult."
For example, let us suppose that a child pa.s.ses all the tests in VI, five of the six tests in VII, three in VIII, two in IX, and one in X.
The total credit earned is as follows:--
_Years__Months_ Credit presupposed, years I to V 5 Credit earned in VI, 6 tests pa.s.sed, 2 months each 1 Credit earned in VII, 5 tests pa.s.sed, 2 months each 10 Credit earned in VIII, 3 tests pa.s.sed, 2 months each 6 Credit earned in IX, 2 tests pa.s.sed, 2 months each 4 Credit earned in X, 1 test pa.s.sed, 2 months 2 ---- ---- Total credit 7 10
Taking a subject who tests higher, let us suppose the following tests are pa.s.sed: All in X, six of the eight in XII, two of the six in XIV, and one of the six in "average adult." The total credit is as follows:--
_Years__Months_ Credit presupposed, years I to IX 9 Credit earned in X, 6 tests pa.s.sed, 2 months each 1 Credit earned in XII, 6 tests pa.s.sed, 3 months each 1 6 Credit earned in XIV, 2 tests pa.s.sed, 4 months each 0 8 Credit earned in "average adult," 1 success, 5 months 5 ---- ---- Total credit 12 7
One other point: If one or more tests of a year group have been omitted, as sometimes happens either from oversight or lack of time, the question arises how the tests which were given in such a year group should be evaluated. Suppose, for example, a subject has been given only four of the six tests in a given year, and that he pa.s.ses two, or half of those given. In such a case the probability would be that had all six tests been given, three would have been pa.s.sed; that is, one half of all.
It is evident, therefore, that when a test has been omitted, a proportionately larger value should be a.s.signed to each of those given.
If all six tests are given in any year group below XII, each has a value of 2 months. If only four are given, each has a value of 3 months (12 4 = 3). If five tests only are given, each has a value of 2.4 months (12 5 = 2.4). If in year group XII only six of the eight tests are given, each has a value of 4 months (24 6 = 4). If in the "average adult" group only five of the six tests are given, each has a value of 6 months instead of the usual 5 months. In this connection it will need to be remembered that the six "average adult" tests have a combined value of 30 months (6 tests, 5 months each); also that the combined value of the six "superior adult" tests is 36 months (6 6 = 36). Accordingly, if only five of the six "superior adult"
tests are given, the value of each is 36 5 = 7.2 months.
For example, let us suppose that a subject has been tested as follows: All the six tests in X were given and all were pa.s.sed; only six of the eight in XII were given and five were pa.s.sed; five of the six in XIV were given and three were pa.s.sed; five of the six in "average adult"
were given and one was pa.s.sed; five were given in "superior adult" and no credit earned. The result would be as follows:--
_Years__Months_ Credit presupposed, years I to IX 9 Credit earned in X, 6 given, 6 successes 1 Credit earned in XII, 6 given, 5 pa.s.sed. Unit value of each test given is 24 6 = 4. Total value of the 5 tests pa.s.sed is 5 4 or 1 8 Credit earned in XIV, 5 tests given, 3 pa.s.sed. Unit value of each of the 5 given is 24 5 = 4.8.
Value of the 3 pa.s.sed is 3 4.8, or 0 14+ Credit earned in "average adult," 5 tests given, 1 pa.s.sed. Unit value of the 5 tests given is 30 5 = 6. Value of the 1 success 0 6 Credit earned in "superior adult" 0 0 ---- ---- Total credit 13 4+
The calculation of mental age is really simpler than our verbal ill.u.s.trations make it appear. After the operation has been performed twenty or thirty times, it can be done in less than a half-minute without danger of error.
THE USE OF THE INTELLIGENCE QUOTIENT. As elsewhere explained, the mental age alone does not tell us what we want to know about a child"s intelligence status. The significance of a given number of years of r.e.t.a.r.dation or acceleration depends upon the age of the child. A 3-year-old child who is r.e.t.a.r.ded one year is ordinarily feeble-minded; a 10-year-old r.e.t.a.r.ded one year is only a little below normal. The child who at 3 years of age is r.e.t.a.r.ded one year will probably be r.e.t.a.r.ded two years at the age of 6, three years at the age of 9, and four years at the age of 12.
What we want to know, therefore, is the ratio existing between mental age and real age. This is the intelligence quotient, or I Q. To find it we simply divide mental age (expressed in years and months) by real age (also expressed in years and months). The process is easier if we express each age in terms of months alone before dividing. The division can, of course, be performed almost instantaneously and with much less danger of error by the use of a slide rule or a division table. One who has to calculate many intelligence quotients should by all means use some kind of mechanical help.
HOW TO FIND THE I Q OF ADULT SUBJECTS. Native intelligence, in so far as it can be measured by tests now available, appears to improve but little after the age of 15 or 16 years. It follows that in calculating the I Q of an adult subject, it will be necessary to disregard the years he has lived beyond the point where intelligence attains its final development.
Although the location of this point is not exactly known, it will be sufficiently accurate for our purpose to a.s.sume its location at 16 years. Accordingly, any person over 16 years of age, however old, is for purposes of calculating I Q considered to be just 16 years old. If a youth of 18 and a man of 60 years both have a mental age of 12 years, the I Q in each case is 12 16, or .75.
The significance of various values of the I Q is set forth elsewhere.[44] Here it need only be repeated that 100 I Q means exactly average intelligence; that nearly all who are below 70 or 75 I Q are feeble-minded; and that the child of 125 I Q is about as much above the average as the high-grade feeble-minded individual is below the average.
For ordinary purposes all who fall between 95 and 105 I Q may be considered as average in intelligence.
[44] See Chapter VI.
MATERIAL FOR USE IN TESTING. It is strongly recommended that in testing by the Stanford revision the regular Stanford record booklets be used. These are so arranged as to make testing accurate, rapid, and convenient. They contain square, diamond, round field, vocabulary list, fables, sentences, digits, and selections for memory tests, the reading selection barred for scoring, the dissected sentences, arithmetical problems, etc. One is required for each child tested.[45]
[45] Houghton Mifflin Company will supply all the printed material needed in the tests, including the lines for the forms for VI, 2, the four pictures for "enumeration," "description," and "interpretation,"
the pictures for V, 3 and VI, 2, the colors, designs for X, 3, the code for Average Adult 6, and score cards for square, diamond, designs, and ball-and-field.
This is all the material required for the use of the Stanford revision, except the five weights for IX, 2, and V, 1, and the Healy-Fernald Construction Puzzle for X. These may be purchased of C. H. Stoelting & Co., 3037 Carroll Avenue, Chicago. It is not necessary, however, to have the weights and the Construction Puzzle, as the presence of one or more alternative tests in each year makes it possible to subst.i.tute other tests instead of those requiring these materials. This saves considerable expense. Apart from these, which may either be made at home (see pages 278, 279) or dispensed with, the only necessary equipment for using the Stanford revision is a copy of this book with the accompanying set of printed matter, and the record booklets. The record booklets are supplied only in packages of 25.
CHAPTER IX