Children Tested by the Point Scale and the Performance Scale
- Author:
Rudolph Pintner, Ph.D., and Jeannette C, Reamer.,
Ohio State University, Columbus, Ohio.
The present study is a comparison between ratings of the same children on the Yerkes-Bridges Point Scale1 and the PintnerPaterson Performance Scale.3 Since the authors of the latter scale suggest four possible ways of computing mental age, a comparison of the ratings on the Performance Scale by these four methods is made.
This study therefore raises two distinct problems. The one, rather technioal in its nature, is as to the best method, out of four possible methods, of computing mental age on the Pintner-Paterson Performance Scale. The other, of very wide interest, is as to the best method of measuring general intelligence, whether by a scale involving no language responses or by a mixed scale of language and performance tests. The latter question is merely raised in this study and we cannot at the present time expect a conclusive answer, which would necessitate exact knowledge of the r61e of language in the development of intelligence. Does the development of language indicate the development of intelligence? Are language stimulus and language response necessary to develop intelligence? Are individuals who are limited in the reception of such stimuli and in their responses to them handicapped to that extent in the development of their intelligence? Is the measurement of the ability to do things as adequate a measurement of the general intelligence of an individual as is a measurement based upon language responses? Are the language responses of the individual correlated well or not so well with his performances? Can there be development of ability to make language responses out of proportion to ability to perform, and if so is the ability to make the language response or the nonlanguage response the best measurement of the intelligence of the individual? More specifically, will the mentality of the individual be best measured by a scale of performance tests, or by a mixed scale of performance and language tests, or by a scale of language tests alone?
1 Yerkea, R. G., Bridges, J. W., and Hardwick, R. S. A Point Scale for Measuring Mental Ability. Baltimore: Warwick and York, 1915. ‘ Pintner, R., and Paterson, D. G. A Scale of Performance Testa. New York: Appleton and Co., 1917.
These are the questions involved in the comparison of the Yerkes-Bridges with the Performance Scale. A complete answer to all of them is impossible, but some light may be thrown upon them as a result of this study.
The Subjects.
The children used for this work were ninety-seven inmates of a County Children’s Home, including all from the first grade through the eighth grade.
On the whole the intelligence of this group is somewhat lower than the general level of a public school. However, as we are not concerned with the intelligence of the child but rather with the correlation of his ranking on the two different scales, the lower average intelligence can make no difference in the results of this work. Another group includes twenty-six subjects from the first grades of a public school in a very poor district of the city. The children range from five to twelve years of age.
The remainder of the total number is made up of fifty-five miscellaneous cases examined at the University Psychological Clinic. In this group the range is from five years to nineteen. The total number of children tested is one hundred and seventy-eight.
Scoring.
The total credits on the Point Scale were counted according to the standard method. The mental age derived from the total credits was computed from the table of age norms first published by the authors. The intelligence quotient of each child was then determined. On the Performance Scale the responses were evaluated according to the four possible methods suggested by the authors, and intelligence quotients for the three methods, which express the results in dental ages, were calculated. The fourth or percentile method cannot be used for conversion into an intelligence quotient.
Results.
Tables 1 to 4 show the distribution of the total number of cases according to the mental age as determined by the Yerkes-Bridges Scale and the three methods of the Performance Scale. These tables give interesting pictures of the range of the two scales. It will be noticed that the mental ages on the Yerkes-Bridges Scale reach up to age eighteen. The median mental age and the point age methods of the Performance Scale stop at age fourteen, while Mental Aqe. 8 4 5 6 7 10 11 12 13 14 15 16 17 18 Total Total 1 7 15 11 11 45 28 17 26 3 4 2 2 1 2 1 178 TABLE 1.?DISTRIBUTION ACCORDING TO MENTAL AGES ON THE YE USES-BRIDGES SCALE. Mental Age. Total 3 4 5 6 7 8 0 10 11 12 13 14 15 16 17 18 Total 14 22 28 34 28 14 178 TABLE 2.?DISTRIBUTION OF CASES. PERFORMANCE SCALE. MEDIAN MENTAL AGE METHOD. the year scale method stops abruptly at a mental age of thirteen. This is due to the fact that our norms do not go beyond fourteen years. The scale as a whole might be discriminative above this point if the standardization were extended, although in some tests with which children above the age of fourteen have been tested, there are indications that the tests do not discriminate between individuals whose mentality exceeds the mental age of fourteen. If an extension of the norms were to show this to be true, then the scale would be limited in its use to children below the age of fourteen. POINT AND PERFORMANCE SCALE TESTS. 145 Mental Aoe. 3 4 5 6 7 10 11 12 13 14 15 16 17 18 Total Total 3 14 14 39 28 28 16 6 21 178 table 3.?distribution of cases, performance scale, point age method. Mental Age. 2 3 4 5 6 7 10 11 12 13 14 15 16 17 18 Total Total 1 3 3 10 12 21 27 22 28 27 24 178
table 4.?distribution of cases, performance scale, tear scale method. The distributions according to intelligence quotients, show very much the same features as the mental age distributions, that is, that the Yerkes-Bridges intelligence quotients show the most nearlynormal distribution. On the distribution according to the Performance Scale the point age method conforms most nearly to the Yerkes-Bridges, while the year scale method extends up to very much higher intelligence quotients, thus spreading the distribution ?f the cases out at the upper end. Conversely, the median mental age method masses more cases at the lower end.
Graph 1 shows these different distributions. The curves are fairly normal considering the total number of cases. On the whole the point age curve follows the Yerkes-Bridges most closely. The median mental age method shows a curve which conforms very well with the Yerkes-Bridges except at the lower end. The year scale method shows very few cases at the lower end of the distribution and far too many at the upper end. The cases are, furthermore, very scattered.
Correlations 6^(d2) The formula P = 1???- was used. Correlations have been n(n2-l) computed with the Yerkes-Bridges ranking in each case and they are as follows: Yerkes-Bridges and Median Mental Age. P = .50 r =. 52 P. E.= .037 Yerkes-Bridges and Point Age Method. P = .55 r= .57 P. E. = .037 Yerkes-Bridges and Year Scale Method. P = .59 r-.61 P. E. = .033 Yerkes-Bridges and Percentile Method. P= .41 r = .43 P. E.= .043
With the exception of the percentile method, the correlations between the Yerkes-Bridges and the Performance Scale are all above .50 and the probable errors are small. The difference between the three correlations above .50, that is, .52, .57, .61, is small, so that as far as these correlations are concerned we cannot say that any of the three methods is decidedly superior to the others. The fact that the three correlations are not higher would seem to indicate that we are not testing exactly the same kind of ability by the Performance Scale as by the Yerkes-Bridges Scale. Both scales are testing ability of some kind and on the whole rank the individuals in much the same order, but there is some difference between them. As to which scale gives the better ranking of the individual in accordance with his real intelligence, it is, of course, impossible to say. The conclusion for practical work would seem to be that where possible, both scales should be used. A combined rating of the individual’s achievement on both scales might give us the best index of his intelligence.
Correlations for Specific Age Groups.
The groups at ages ten and eleven were used because they contained the largest number of subjects. The Spearman Foot Rule Method was used. The correlations between the ranking on the Yerkes-Bridges Scale and the ranking on each method of the Performance Scale, and also between each method and every other method of the Performance Scale were computed. The correlations are as follows:
Age 10. No. of cases 24R r Yerkes-Bridges with Median Mental Age 45 .65 ” ” Point Age Method 49 .70 ” ” Year Age Method 50 .71 Median Mental Age with Point Age Method 88 .98 ” ” ” Year Age Method 75 .92 Point Age Method with Year Age Method 79 .95 Age 11. No. of cases 21. Yerkes-Bridges with Median Mental Age 21 .32 ” ” Point Age Method 33 .50 ” ” Year Scale Method 34 .51 Median Mental Age with Point Age Method 70 .89 ” ” Year Scale Method ‘.69 .88 Point Age Method with Year Scale Method 74 .92
The correlations of the different methods of the Performance Scale with every other method are higher in all cases than the correlations of the Yerkes-Bridges and the Performance Scale. This is exactly what we would expect, because even with the different methods of computing mental age, we are measuring the same abilities. Table 5 shows the median mental age for each chronological age according to the four methods. This table shows a rather consistent rise of the mental age with the chronological age. In reading across the table the relative differences between the different methods and the Yerkes-Bridges Scale is again brought out. In following down the column, or in other words, with the advance in chronological age, we see the increase in median mental age. The increase from year to year is not constant, nor would we expect it to be, but there is some increase all along, except at age thirteen on the Performance Scale and at ages fourteen and above on both scales. Our older children would seem to be more retarded mentally than our younger. The average of the differences at each age between the Yerkes
01 ?i 1 1 1 1 1 1 1 ii?’”???<?? ^? i JO-if p-tf frsj -lo-lf Jc-ff go-S<j fo-ff iK-iof “c-iiy no-/if iic-iif /w-iifj /sc/sf ti-‘lf i/i-Jy If Graph I. Graph I. , “I- ^ /? 9 / V 7* /s-r ,<[- /?? ‘_ -y Ytrkes. ? ? ?? ?. Tiled ian M4. 5W^ - Tent Ay. vl ._. )W tye. 3 2 / arf,Ap. 6 ? g 9 “>? a ? >*? rj-‘t Graph II. Bridges and the median mental age methods shows that the YerkesBridges, on the average, ranks a child about .39 of a year higher than the Performance Scale. Chron. Age No. of Medians Yerkes M. M. A. Point Age Year So. Year Scale minus Yerkes M. M. A. minus Yerkes Point Age minus Yerkes 7 8 9 10 11 12 13 14 15’ 16 17’ 18 19 4 14 19 19 21 24 21 17 12 14 4.6 5.75 7.8 8.16 8.64 9.0 9.9 10.5 11.15 10.10 11.9 9.42 5.0 6.0 7.0 7.0 8.0 9.0 9.25 10.5 9.5 11.0 11.42 8.50 5.0 6.47 7.31 7.63 8.32 8.9 9.37 10.12 10.69 10.82 12.07 4.8 7.40 8.60 8.93 9.80 10.95 11.26 12.4 12.26 12.7 13.0 11.09 +0.20 +1.65 +0.80 +0.77 +1.16 +1.95 +1.36 +1.90 +1.11 +2.60 +1.10 +1.67 +0.40 +0.25 -0.8 -1.16 -0.64 0 -0.65 0 -1.65 +0.90 -0.48 -0.92 +0.40 +0.72 -0.49 -0.53 -0.32 -0.10 -0.53 -0.38 -0.46 +0.72 +0.17 -0.54 Average Differences. +1.35 -0.39 -0.11
table 5.?median mental age for each chronological age by each method, WITH AVERAGE DIFFERENCE FOR EACH.
The same average difference for the Point Age Method shows the Yerkes-Bridges diagnosis only .17 of a year higher. The largest difference of all occurs with the Year Scale Method, where it is 1.35 of a year higher than the Yerkes-Bridges median. Graph 2 makes the point somewhat clearer, that the variation between the methods, except in the case of the Year Scale, is not very great. The graph shows the median mental ages for each chronological age calculated according to the four different methods. The curves for the Yerkes-Bridges, the Median Mental Age and the Point Scale are on the whole rather close together, while the Year Scale curve is very much above the others at every age. Diagnosis by the Different Methods.
The diagnoses shown in table 6 are made roughly from the tables of distribution according to intelligence quotients and so do not follow the usually accepted divisions. Intelligence quotients below .70 are supposed to denote feeblemindedness, those from ?70 to .89 backwardness. Subjects receiving quotients from .90 to
Feebleminded Backward Normal Bright VeryBright Totals Intelligence Quotients. 0-69 70-89 90-109 110-129 130Yerkes-Bridges Median Mental Age. Point Age Year Scale 26 32 25 52 73 65 48 63 51 64 62 28 18 20 40 9 4 4 20 178 178 178 178 Totals. 91 238 240 106 37 712
TABLE 6.?TABLE OP DIAGNOSES, SHOWING DISTRIBUTION OP CASES BY EACH METHOD.
1.09 are considered normal, from 1.10 to 1.29 bright, and from 1.30 upward very bright. The median mental age method has the advantage of being a very easy and quick way of determining a child’s mental age, but it grades a larger percentage of the cases feebleminded and backward, and fewer normal, bright, or very bright, than the Yerkes-Bridges Scale.
The year scale method goes to the other extreme. An abnormally large percentage of cases are thrown into the groups of bright and very bright.
The correlation of the percentile method with the YerkesBridges is the lowest of the four methods and at present does not seem to be of much value. Possibly this is due to the comparatively small number of cases tested at each age. There should be several hundred at each chronological age to give reliable norms according to this method. The advantage of such a distribution is that new cases can always be added. This addition may be accomplished at some future date.
In table 6, showing the distribution of cases as diagnosed, it will be seen that, on the whole, there is less variation in the size of the groups diagnosed feebleminded, backward, etc., by the YerkesBridges Scale and the Point Scale Method on the Performance Scale than in the case of the other two methods. The difference between the medians for these two rankings is only .11 of a year.
At present it would seem that either the Median Mental Age Method or the Point Age Method gives a close enough approximation to the Yerkes-Bridges Scale for all practical purposes, although just where the dividing lines between different grades of intelligence are to be drawn, it is difficult to say. The correlations between the various methods of the Performance Scale and the Yerkes-Bridges Scale, although not extremely high, are, however, high enough to indicate that both scales are roughly grading the children in the same way. There is, of course, a certain amount of discrepancy and this indicates that the two scales are not testing the same qualities, and that ability to make accurate language responses to the tests of the Yerkes-Bridges Scale does not imply similiar ability to make non-language responses, such as are required by our performance tests. This difference between ability on language and on performance tests has been repeatedly pointed out and has shown itself strikingly in this study. It is precisely for this reason that a performance scale is required to supplement our present intelligence scales with their emphasis upon tests requiring language responses. Summary.
1. A comparison between the Yerkes-Bridges Scale and the Pintner-Paterson Performance Scale has been made by comparing the records of 178 children tested on both scales. 2. The correlation between the two scales indicates that exactly the same abilities are not tested by both scales. 3. The scales would seem to supplement each other. 4. The Point Scale Method on the Performance Scale seems to correspond most closely to the results obtained on the YerkesBridges Scale. 5. The Median Mental Age Method shows a fairly close correspondence with the results obtained on the Yerkes-Bridges Scale. By the suggested method of diagnosis, it seems, however, a little too severe in comparison with the Yerkes-Bridges Scale. 6. The Year Scale Method of the Performance/Scale is much too lenient in comparison with the Yerkes-Bridges Scale. 7. The Percentile Method of the Performance Scale is of doubtful value with the present inadequate standardization.
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.