Some Performance Test Norms for Children
- Author:
Clairette P. Armstrong
New YorTc City Problem This study is an investigation of the reactions of normal public school children from eight to fifteen years of age on two Army Performance scales standardized for the most part on adult males during the war.
Scales of real diagnostic value for school children, that gauge the specialized intelligence involved in non-language tests, are sought. How accurately do these two scales evaluate the ability of the child to manipulate concrete objects, to respond to actual objective situations?
Reasons for Research As Wells points out, “Most groups coming to clinical attention make a readier and more interested response to non-language tests. This is not conclusive evidence that non-language tests are more serviceable from the standpoint of diagnosis but it is a factor tending in this direction.”1 Many clinics have a preponderating number of children of foreign-born parents with a foreign-home environment. Often English is not spoken in the home and there results a language difficulty for the child, tantamount to a verbal deficiency. Scales are necessary in clinical practice, convenient to administer and score, that test different abilities of a psycho-motor nature. There is room for many tests of capacity to react to concrete situations. Method and Procedure The short scale of the Army Individual Performance Test2 consisted of designs, manikin and feature profile, mazes, cube imitation and ship tests. The Army Beta 3 is composed of seven items: iF. L. Wells, Mental Tests in Clinical Practice, p. 117. 2 Yoakum & Yerkes, Army Mental Tests, p. 126. 3 Ibid., p. 79. mazes, cube analysis, X-0 series, digit-symbol, number checking, pictorial completion, geometrical construction, all of -which afford opportunity for a wide variety of skills. The Army Beta Test as an individual test was administered to 436 children, whole grades, 4th to 8th inclusive, in public schools in New York City and in towns in northern Westchester County. The test was administered exactly according to Army directions, in pantomine. The demonstration tests were mimeographed instead of being shown on a blackboard. The experimenter performed the first item of each practice series on paper just as in the army the experimenter did on the blackboard; then instead of having an assistant demonstrate the next item of the series, the child was told to do it. Generally the stimulus was successfully reacted to immediately, but if the subject failed to comprehend what was expected the examiner performed the next item for him without comment. Occasionally the examiner was required to demonstrate all items of one practice sheet and even then the subject might not comprehend, but never was this the case for all tests of the series; indeed it rarely occurred with any one child for more than one item.
The Army Individual Performance Test, short scale, was also administered to the same children and a few more, numbering 461. In scoring this test, the digit symbol test is included although administered only with the Beta Test first. The procedure of administering may be the same in the two scales. It was included in the scoring because it is a valuable adjunct to the short Individual Performance scale and exercises a different function from the other items of the scale.
For comparison, results are recorded of tests administered to 417 of these children on the Otis Group Test for grades 4 to 8. Raw scores of the Army Individual Performance Test4 were transmuted into weighted scores and mental ages5 found from the tables, as with the Beta.6 For statistical purposes the scales offer much, since I.Q.s readily computed facilitate statistical comparisons. An I.Q. has many advantages and probably some disadvantages inherited from the Stanford-Binet situation. If it is generally remembered that the I.Q. is merely a ratio between a mental age 4 Ibid., p. 125. 5 Ibid., p. 127. 6 Ibid., p. 133.
and a chronological age, a useful symbol in various situations, the “dogma” may tend to disappear from the I.Q. The I.Q. may be regarded in its true light, not as a mystic label of a cryptic faculty called intelligence which because of its esoteric character defies definition, but which none the less may be irremediably and irrevocably attached to an individual. If I.Q.s of concrete abilities are a commonplace as well as I.Q.s of verbal abilities, the usefulness of the I.Q. may be restored and a more accurate and enlightened interpretation of it may prevail. Results
Table I shows the number of children from 9 to 14 years of age (omitting the few of 8 and 15 years), the median scores and the test norms, all at six month intervals, on the two Army Performance scales and the verbal test. There is a consistent superiority of these children on all three tests up to 13 years, 6 months and a general inferiority to the norms from then on. On the Individual Performance Test most of the 452 children exceeded the Army norms.
At 12 years they were inferior merely by a fraction and continued so to a varying degree at 13^2 and 14^/2 again negligibly at 14. A few children fell below the army norms on the Beta at 131/2 and 14i/o years. The children were ahead of the Otis norms at all ages up to 13^ where they dropped slightly behind and remained so. Table I also shows by interpolations from the tables of norms the number of months’ difference per six month interval, the superiority of the New York boys and girls indicated by plus, inferiority by minus. The greatest superiority over the norms is on the Beta Test. The test is evidently far too easy by from 12 to 28 months, except at 13y2 and 14y2 years. The Individual Performance scale is too easy by from 4 to 22 months except at 12 years where the scores are practically identical, and at 13^> and 141/2 where the children fall well below. There is a like superiority on the Otis.
“Were yearly intervals considered instead of semi-yearly the differences would be somewhat less. A scrutiny of the scores shows that the tests distinguish between children of different ages for there is a regular augmentation of the score year by year.
Table I Comparison Median Scores op Children with Norms Army Performance Test Army Age I Norm 9 67 9.5 86 10 108 10.5 127 11 144 11.5 158 12 169 12.5 179 13 188 13.5 197 14 205 14.5 214 No. Children 16 44 37 39 54 41 41 36 41 47 37 19 452 Median Score 139.5 127.5 143.5 159 157.5 171.66 168.75 186 194.16 178.75 203 185.83 Diff. (mos.) +22 + 12 +12 +12 + 6 + 7 -.25 + 5 + 4 -12 - 1 -19 Army Norm 17 24 30 37 42 47 53 58 63 67 71 75 Beta Test No. Children 11 41 36 39 49 38 36 35 41 48 37 19 430 Median Score 44.33 50.25 56.5 54.8 62 64 63.5 70.5 72.75 64.5 74 72.5 Diff. (mos.) +26 +27 +28 + 18 +23 + 18 + 12 + 18 +15 - 5 + 5 - 3 Otis Test Otis Norm 15 19 23 27 31 35 38 41 44 46 49 51 No. Children 11 39 35 37 44 38 34 33 41 44 36 19 411 Median Diff. Score (mos.) 29.5 +21 23 +6 28.5 + 8 32.5 + 7 33.5 + 3 41.5 +12 43.75 +11 47.25 +14 49.5 +13 45.5 - 1 47.5 - 2 49.5 - 3
As is predictable from the first table, Table II demonstrates a consistent lowering of median I.Q.s as age advances for the 452 children between 9 and 14 years on the Performance, the 430 on the Beta and the 411 on the Otis tests.
The I.Q. gradually diminishing yearly is 17 points less for 14 year olds than for 9 on the Individual Performance Test and 21 points less on the Beta. There is a similar tendency with the Otis but the decrease of I.Q. with augmenting age is not consistent. However the I.Q. of the 14 year olds is 11 points below the 9 year olds. The I.Q.s were computed from Binet mental age equivalents for the sake of uniformity with the Army tests.
Table II Median I.Q.s per Age Level Age 9 10 11 12 13 14 Teat Performance 111.0 107.36 104.33 101.0 95.5 93.85 Beta 118.0 115.28 111.4 105.46 100.7 97.25 Otis 105.5 103.2 105.5 104.2 99.5 94.57
The frequency distributions of I.Q.s on the three tests for 461 children on the Performance Test, 436 on the Beta and 417 on the Otis, ages 8 to 15 inclusive, are given in Table III. Histograms illustrate, each accompanied by the frequency polygon on the same coordinate axis. The frequency curves are generally symmetrical in all three cases and show a fairly normal distribution of I.Q.s. There is only a slight skewness for each test.
The average I.Q.s are also given in Table III, about 103 for the Individual Performance Test, 109 for the Beta and 103 for the Otis. The middle fifty per cent on the Performance Test fall between 92 and 113, on the Beta between 99 and 117 and on the Otis between 90 and 115. Terman’s 7 middle fifty per cent fell between 93 and 108. The Performance Test spreads somewhat above this at the upper end and is negligibly different at the lower; on the Beta the middle fifty per cent are consistently higher; on the Otis they cover a wider area, somewhat lower and distinctly higher. ” Terman, The Measurement of Intelligence, p. 67.
Army Individual Performance Test / -X65- 75- 85- 95- 105- 115- 125- 13574 84 94 104 114 124 134 144 % 1.74 8.68 20.83 24.29 25.81 12.36 4.98 1.3 (Terman) 2.3 8.6 20.1 33.9 23.1 9.0 2.3 .55 Army Individual Performance Test Army Beta 65^. 75?- 85~ 95~ 105_ 115~ 125- 135- 145- iv74 84 94 104 114 124 134 144 ir.4 1fU .23 3.67 12.84 20.87 33.03 16.74 11.01 1 37 01 Otia Test 55- 65- 75- 85- 95- 105- 115- 125- 135- 145- 155- 16564 74 84 94 104 114 124 134 144 154 164 174 % .72 3.11 10.31 21.58 21.58 17.03 12.95 7.44 4.08 .72 .24 .24
Table III I.Q. Distributions (Ages 8-15) 165-174. 155?164. 145-154 135-144 125-134 115-124 105-114 95-104 85?94. 75-84. 65-74. 55-64. Av. I.Q…. Sigma dist. Middle 50 per cent. Performance No. 6 23 57 119 112 96 40 8 Per cent 1.3 4.98 12.36 25.81 24.29 20.83 8.68 1.74 461 99.99 102.9 14.45 92-112.6 Beta No. 1 6 48 73 144 91 56 16 1 Per cent .23 1.37 11.01 16.74 33.03 20.87 12.84 3.67 .23 436 99.99 108.55 13.45 Otis No. 1 1 3 17 31 54 71 90 90 43 13 3 Per cent .24 .24 .72 4.08 7.44 12.95 17.03 21.58 21.58 10.31 3.11 .72 99-117.14 417 100.00 103.28 18.02 90.22-115.38
Terman’s percentages of children per I.Q. interval are also given under the Individual Performance Test column diagram, though his step intervals are slightly different, 66-75 instead of 65-74 as with the New York children. The Performance Test shows a fairly similar spread but at a slightly higher level throughout except that Terman’s with 33.9 per cent are somewhat more massed between 96-105; between 95-104 are 24.29 per cent of the New York group with the largest percentage 25.8 per cent at the 105-114 interval in contrast with Terman’s 23 per cent.
Table IV Middle 50 Per Cent (I.Q.s) QhQ3 & Q Yr. 9 10 11 12 13 14 Beta Test Q1-Q3 110.10-127.00 108.50-124.50 105.84-110.08 99.13-112.77 90.90-110.90 90.56-110.33 Q 8.4 8.0 6.62 6.82 10.00 9.88 Performance Test Qi-Qj 104.13-118.04 99.40-116.35 94.63-115.50 91.87-111.27 86.83-106.59 87.54-105.70 Q 6.95 8.47 10.43 9.70 9.88 9.08 Otis Test Q1-Q1 99.7-120.4 91.8-108.7 91.1-126.2 89.2-116.3 87.7-111.6 84.9-105.5 Q 10.4 8.4 17.5 13.5 11.9 10.3
Table IV gives the I.Q. ranges of the middle 50 per cent per age level as well as Q, on the two non-language scales and the verbal. On the former the interquartile ranges are not absolutely equivalent for the different age levels but they are less variable than on the Otis test, as Q easily shows. The middle 50 per cent of the scores on the non-language tests tend to fall between from nearly seven to slightly over ten points above and below the norm for a given age. On the Stanford-Binet the range at five years was 97-111 and at eleven 92-108, seven and eight points above and below the norms.
The wider range of the New York group may be partly due to the fact that the children were unselected and not within two months of a birthday as with Terman’s group which also was larger than the New York group. The spread on the Otis and marked variability may be attributable to the presence of children of foreign-born parents, and a possible language or verbal inequality.
Table V Correlations (Scores) (417 boys and girls) r P.E.j. Otis & Performance 629 ?.0199 Otis & Beta 633 ?.0197 Beta & Performance 801 ?.0188 Table V shows the correlations by product moment formula, between the test scores of 417 children having all three tests, to be positive and fairly high. The Pearson r .629 between the Army Individual Performance Test and the Otis, and .633 between the latter and the Beta, support the concept of g. The correlation .801 between the two non-language tests may be emphasized by one identical item in the scale of six and in the scale of seven items.
Discussion
The appearance of greater ease on all these tests for younger than for older children may be partly explained by a perhaps irremediable characteristic of tests. There is greater opportunity for younger children to score high than for older who are checked by the finite quality of tests?the more and more limited spread with advancing age. Again, the selective effect of the public school, since only grades 4 to 8 were tested, and dull pupils pile up in the higher grammar grades, plays its role. The fact that 16 years was used as the age of mental maturity instead of 14 would be another factor in a somewhat lower I.Q. for those 14 and 15 of age. D. W. Oates suggests “That efficiency of reaction as measured by ‘ability to speed’ reaches maturity early, while efficiency of meaning matures later.”8 Ability to speed may be a factor in the extremely high I.Q. of younger children on these tests for it is an important item in the scales.
The following criticism may be summarized concerning the Beta:9 That the test shows a massing of scores at the upper part of the scale with a large number making a perfect score on the separate tests and that for the total English-speaking group of 653 recruits in the army the mode falls near the upper limits of the scale with no trailing off; that it does not differentiate well, is not satisfactory for the high grade and is distinctly easier than Alpha or Binet. This may be due to the excessive amount of time allowed for each item of the Beta. Children in the New York group frequently finished far ahead of the allotted time. Occupational level may be an important factor in the high I.Q.s in the New York Group where there were no children of unskilled laborers. It is generally conceded on the basis of considerable research that children in inferior occupational classes have lower test scores than children in higher occupational classes.
This is true of ability to react to objective as well as to verbal situations.10 Since the schools were all in better residential districts there were fewer children of foreign-parentage than are found in public sc ools in other localities. This is another factor leading to the ig ler I.Q. First generation American children have been found in erior on non-language as well as verbal tests by many investigators among others Young11 and Kirkpatrick.12 ^ie biological Significance of Intelligence Tests. J. Educ. Psychol., 1S30, 21, 422-450. Memoirs National Academy of Sciences, V. 15, 1921, p. 394. 10 p P A __ * * Armstrong, A Study of the Intelligence of Eural and Urban Children. J. Ed. Sociology, 1931, 5, p. 311. 1 ? Young, Mental Differences in Certain Immigrant Groups. Univ. Ore.PuU.,l922, V. 15, pp. 417_434. . Kirkpatrick, Intelligence and Immigration. Ment. Measure. Monog., I, 1926, N. Summary and Conclusions 1. The Army Individual Performance Test, short scale, is somewhat easy for New York public school children from 9 to 11 years who at 12 approach the norm but at 13 and 14 it becomes somewhat difficult. 2. The Army Beta is far too easy save at 13 and 14 years. 3. These two non-language tests discriminate well and consistently between age levels. 4. They follow a fairly normal frequency curve. 5. The middle 50 per cent per age level on the non-language tests do not vary greatly as to range. 6. The higher than average I.Q.s of the New York group may be attributed to some selection as to occupational level and nationality. 7. The two non-language tests are testing the same thing as the verbal test to a marked degree. 8. The two non-language tests for the most part are testing manipulative ability fairly coincidently. Therefore it may be concluded: 1. As conspicuous success for older as for younger children on these tests is militated against by earlier maturation in capacity to speed, by the intrinsic characteristic of tests, their finite quality and by the selective factor of public school grading. 2. The Army Individual Performance Test, short scale, is expedient for clinical practice if test results are interpreted in the light of facts brought out. 3. The Beta should be restandardized, probably on a much stricter time basis.
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.