The Factor of Experience In Intelligence Testing
- Author:
Rudolph Pintner and Donald G. Paterson.
Ohio State University, Columbus, Ohio.
This paper will endeavor to show that the factor of experience in intelligence testing depends upon the type of test used. In a previous article1 the writers showed that those defectives over fifteen years of age, as compared with defectives below fifteen years of age of the same mental age, greatly excelled in two tests of the BinetSimon scale, i. e. naming the days of the week and the months of the year. Results from a similar comparison with a different type of test show that when real intelligence tests are used children greatly excel adults of the same mental age computed by the present BinetSimon scale. A discussion of the results of our previous paper was made in the February (1915) number of The Psychological Clinic by Dr J. E. Wallace Wallin,2 who restated a previous study of his concerning this factor which we call experience. He showed that adults excel in reading tests as compared with children, concluding that here the practice or experience factor was the cause. In the form board test however the children excelled, due according to Wallin’s opinion to a more extensive practice with such devices in the school room. In the color-naming test the children also excelled in the time taken to re-act to the situation. The children excel in uncontrolled free association as they do in immediate memory also, i. e. recalling the content of a reading selection. In all these cases Wallin’s contention is that this superiority on the part of the children is due to schoolroom practice. We do not agree with this position but hold rather that the children excel because these tests are more adequate measures of intelligence than are such tests as the weekdays, months, and reading tests. If we accept Stern’s3 definition of general intelligence as general mental adaptability to new problems and conditions of life, then those tests that call for adaptations on the part of the re-agent to relatively new situations will best discriminate between degrees of 1 Experience and the Binet-Simon tests. The Psychological Clinic, Vol. VIII, No. 7, December 15, 1914.
2 J. E. W. Wallin. The Binet-Simon tests in relation to the factors of experience and maturity. The Psychological Clinic, Vol. VIII, No. 9, February 15, 1915. ‘Stern, William. Psychological methods of testing intelligence. Tr. by G. M. Whipple. Baltimore: Warwick & York, 1914, p. 3.
intelligence. Empirical evidence shows this to be the case. Based upon an analysis of the Binet-Simon tests given to feebleminded children, Chotzen4 found that among children of the same mental level, some tests show a decided increase in capacity with increase in chronological age, others no alterations, while there was an actual decrease in others. Stern5 quotes Chotzen thus, “The tests accompanied by strong increase with age relate, then, almost exclusively to matters of information… . No increase at all is present with tests that demand ability to judge and to combine or with such as put severe demands upon apprehension?comparison, problemquestions, noting omissions, repeating five digits.” It is thus seen that tests in which the experience factor is ruled out to a relatively greater extent correlate most highly with general intelligence. TABLE I.?SHOWING RESULTS OF KNOX CUBE TEST. M.A. 21 or above. No. Ave. No. Lines Passed. 20 or below. No. Ave. No. Lines Passed. 15 or above. No. Ave. No. Lines Passed. 14 or below. No. Ave. No Lines Passed. 15 to 21. No. Ave. No, Lines Passed. 1.14 21 2.04 14 1.64 14 2.35 1.42 11 1.18 37 2.13 31 1.58 17 2.52 20 1.80 15 1.80 36 2.80 32 2.03 19 3.31 17 2.23 11 2.45 39 3.92 36 3.16 14 4.00 25 3.i 24 2.79 39 4.20 51 3.54 12 4.16 27 4.29 10 21 4.27 31 4.54 55 4.38 4.80 26 4.50 11 13 ! 3.76 15 5.46 26 4.73 4.00 13 5.69 Total . . 102 Grand Total. .320 218
Keeping these points in mind we will proceed to consider the results obtained in testing 320 feebleminded inmates of the Ohio Institution for Feebleminded Youth6 with the Knox Cube Test7 revised and standardized by Pintner.8 The tests were not made tIbid., p. 89. 6 Ibid., p. 89.
6 The writers wish to acknowledge the courtesy of the Superintendent, Dr E. J. Emerick. 7 Howard A. Knox. A Scale based on the work at Ellis Island, for estimating mental defect. J. of the Am. Med. Assoc., March 7, 1914, Vol. LXII, pp. 741-747. 8 The standardization of this extended Cube Test to be published shortly.
with the view of showing such facts as we here present, but for the purpose of comparison with normal children of corresponding ages. Bias in giving the tests to the older and younger feebleminded individuals is thus ruled out. The test as here given consisted of ten different combinations of tapping four Binet cubes with another cube. Each combination is called a “line”. Table I shows the average number of lines passed by feebleminded children of different mental ages (determined by the Binet-Simon Scale, Goddard Revision). The cases are classified in five groups based upon chronological age. Under the first heading is given the mental age, under the second, the group above 21 years chronologically; the third, those below 21; the fourth, those above 15; the fifth, below 15 and the last those between 15 and 21 years. Under each group is given the number tested and the average number of lines passed. This chronological grouping is quite arbitrary and was decided upon merely to see what changes take place in the relative efficiency of three stages in chronological development, i. e. children below 15, between 15 and 21, and over 21. Wallin holds that the difference between those over 21 and below, of the same mental age, is due to physical growth, to maturity. As we conceive it, that question can not enter into such a consideration as we have before us, for this reason, that if their superiority in any test is held to be due to the factor of growth alone, then they should excel in all the tests of that age and would consequently test higher. The comparison would then break down. The factor that causes this discrepancy between the two groups is one of experience, i. e. they will excel in those tests which depend upon daily experience. The difference is due to the kind of test and not to any maturity difference or any intellectual superiority.
Although maturity and experience are two factors which are no doubt closely inter-related, yet they are different, and it is well to avoid confusion between them. The distinction is not made clear by Wallin, and it seems to us that he at times regards them as one factor only. We doubt whether maturity, “the processes of physical maturation” as Wallin calls it, enters very much, if at all, into the tests that have been discussed in these articles. We can well imagine that very young children may not be able to do as well on the form board because of their inability to co-ordinate the movements involved in grasping and placing the blocks. This would be due to a maturity difference and would apply to young children only, younger than any that have been considered by Wallin or by us. Experience, on the other hand, refers to the number of times an individual has been confronted with the same situation, and should be kept distinct from maturity. We can not conceive of maturity differences entering into any of the tests under discussion with the individuals tested, but experience differences obviously do, and yet only in some specific tests. We have mentioned the days and months test as cases in point. Wallin suggests the reading test as well, and doubtless rapidity of reading depends to some extent upon experience. The children are superior to the adults on the form board. We agree with Wallin that the form board “represents a novel situation,” and we feel that he contradicts himself when he says further on that the children do better because of their greater experience with situations of this kind. In considering the results in table I it will be seen at once that on the cube test in each group there is a constant increase in efficiency for each mental age. However between the five groups in any mental age (age XI excepted) we find differences of efficiency that are significantly constant. Take age V for example, we find those below 15 doing the best with an average of 2.35; those aged 20 or below come second with an average of 2.04; those aged 15 or above, third with 1.64; those aged 15 to 21, fourth with 1.42, and last those over 21 passing only 1.14 lines. The same tendency is seen for each mental age. The curve shows graphically the average efficiency per mental age for those over 21 (solid line) and for those below 15 (dotted line).
This curve enables us to see at a glance that the children are superior to adults throughout all mental ages. This is in direct contrast to the results found in our paper on the days and months test.
Now, we could take the average number of lines passed for all ages for each group, but this would not be accurate for we have not tested the same number in each group in each mental age. It gives us an average which is not a proper basis for comparison. By inspection of table I we see that the younger the child chronologically, the more lines he passes. This holds for all mental ages up to X. Now, the Binet-Simon scale is more adequate at the upper end even although it may not be accurately standardized in ages XI and XII. We mean by this, that ages X, XI, and XII are composed for the most part of tests that present new situations to the re-agent. We never put dissected sentences together, we do not draw designs from memory and so on. But we do make change and count pennies, we often have to find the date and think of the day or month and sometimes buy stamps and count them. These are things an individual, even an inmate of an institution, is doing more or less constantly. So from our point of view and in line with Chotzen’s results, we would not expect to find any age increase in efficiency when subjects mentally arrested above a mental age of IX are considered. Thus there would be no great difference between the five groups in mental ages X and XI. That is exactly what we find. Those over 21 make a great jump in efficiency from IX to X, and although the tendency for the younger children to excel is still present, it is not nearly so conspicuous. We can not explain the apparently chance dropping off in age XI of the adult group, but it is to be noted that only 13 eleven year olds were tested as compared with 24 and 21 in ages IX and X. Our interpretation of these results is that up to age X the Binet scale contains some tests which depend very largely upon daily experience and hence the older the child tested the more liberal is the estimation of his mentality as computed by the present Binet scale. That is, given two individuals, one twelve and the other twenty-one, both having a mental age of VIII, we believe that the one twenty-one years old is not really VIII but has a mentality below VIII. But if both cases had been measured by a scale that has eliminated the experience factor as far as possible, their difference in mentality would have been shown. Now any two individuals, no matter what their chronological age, testing the same by such a scale, would show about the same average performance in any other test involving practically no experience. We will next take up the consideration of a different test, i. e. Knox’s Feature Profile. This test is placed before the child with the following verbal direction, “Put this together.” It pre-supposes on the part of the subject an ability to read a diagram, to perceive the seven pieces of the profile, as a head, to synthesize in other words the various parts into a whole. The results are given in table II.
TABLE II.?SHOWING RESULTS OF FEATURE PROFILE TEST. Chron. Age Group. Mental Age. P F P F P F % F % 10 % 11 % 21 or above.. 0 0 25 18 36 17 10 63 11 91 20 or below. 0 1 12 16 17 65 22 85 15 or above.. 0 1 14 3 14 17 11 15 16 16 50 36 13 73 15 15 to 21. 7 12 0 4 10 55 19 87 80 14 or below ..070 Total over 21?86 Total below 21 = 93 Grand Total.…179 33 71 100 100
Under each mental age are given three columns, one showing the number of passes (i. e. when subject puts the parts together correctly within five minutes), one the number of failures, and the third the percentage of passes. The first column gives the chronological group considered and by reading across the table is found the increase in efficiency from mental age to mental age, by reading up and down is found the difference in efficiency for each group in any mental age. There are 179 cases recorded, 86 above twenty-one chronologically and 93 below twenty-one. This table is not so satisfactory as table I in showing these group differences because of the relatively few cases tested below 15 years.
Below age IX there is no basis of comparison because of the lack of subjects. In age IX however we find that those below fifteen have a percentage of passes of 71, those below twenty-one have 65 per cent, those fifteen to twenty-one have 55 per cent, those above fifteen have 50 per cent, while those above twenty-one trail along behind with only 36 per cent of passes. It is significant that even with the small number of cases we find exactly the same tendency that manifested itself in table I. In age X we find the same tendency as in age IX but to a lesser extent, as in table I. In the Cube Test it is obvious that the test does not depend upon experience in the limited sense of the term as we use it, but depends upon the ability to imitate, to attend to an externally enforced rhythm. It may be said that the Feature Profile Test depends upon school training in that children in school play with blocks and puzzles, but even so the test is so simple that an adult (who has had years of school training also) were he really of the same mental age as the child would be as efficient in meeting the new situation. We hold then that it is not a question of physical growth but rather of a too liberal estimation of an adult’s intelligence by the Binet-Simon scale.
In conclusion we may ask, what does all this show? We believe that first, it is of value in calling attention to the fact that the interpreter of Binet test results must take into account this factor of experience when summing up his case, and secondly, that tests dependent upon experience (when proven so by sufficient evidence) should be eliminated from scales designed to test adults over fifteen years of age, and thirdly that it is of the utmost value as a method of determining the relative value of any additional tests that the psychologist may wish to standardize. With reference to this last, point we mean that by this method we have empirical pi oof that the test in question is or is not a real test of general intelligence. Using this method then as a criterion, we can say that naming the days of the week and the months of the year (and possibly other Binet tests such as counting thirteen pennies, copying a square and a diamond, naming colors, telling forenoon from afternoon, and defining in terms of use) lack the essentials of true intelligence tests while other tests such as the Knox Cube Test and Feature Profile Test possess those characteristics which meet the requirements laid down by Stern for tests of general intelligence. ñ
Disclaimer
The historical material in this project falls into one of three categories for clearances and permissions:
Material currently under copyright, made available with a Creative Commons license chosen by the publisher.
Material that is in the public domain
Material identified by the Welcome Trust as an Orphan Work, made available with a Creative Commons Attribution-NonCommercial 4.0 International License.
While we are in the process of adding metadata to the articles, please check the article at its original source for specific copyrights.