Classics in the History of Psychology

An internet resource developed by
Christopher D. Green
York University, Toronto, Ontario
ISSN 1492-3713

(Return to Classics index)

Differential Action upon the Sexes of Forces which Tend to Segregate the Feebleminded.

Leta S. Hollingworth (1922)
Teachers College, Columbia University

First published in Journal of Abnormal Psychology & Social Psychology, 17, 35-57.

Posted October 2000


It has often been stated that among individuals of the lowest two or three percentiles in the distribution of human intelligence there are more males than females. This statement has been based chiefly on the census of state training schools for mental defectives, and of special classes for subnormal children. Such enumerations practically always show a preponderance of males. Kuhlmann[1] gives figures, which may serve us as fair samples, collected by him in 1915, as the result of a questionnaire returned from seven states in this country. In the state schools for feebleminded were 4,046 males (53.5%), and 3,518 females (46.5%). The United States Report for 1910 showed in institutions for the feebleminded 11,015 males (53.8%), and 9,716 females (46.2%o).

Special classes for subnormal children in local school districts also show more boys than girls on the registers. The reports of the Inspector of Ungraded Classes, in New York City, show a ratio of more than two boys to one girl. Recent reports from England show three boys to two girls on the records of the special schools after-care committee of The City of Birmingham,[2] a ratio "which is frequently found in the various special schools for the mentally defective."

From reports of which those cited are fair samples, inferences like the following are drawn. "Idiocy is almost everywhere recognized as more common in males than in females."[3] "There is no doubt that there are more feebleminded boys than girls. There are more boys in the special classes in London, and in the Hilfschulen in Germany, and in the special classes so far found in the United States, and everything agrees with this."[4]

In working among the children in the special classes for mental [p. 36] defectives in Philadelphia, Sylvesters[5] noticed that when mental tests were made by him, the girls of these classes were revealed as more defective than the boys, though fewer in number, and he commented as follows: "For reasons not of interest here, a relatively small number of girls are placed in the special backward classes. It is a matter of observation, confirmed by these results (in mental tests) that the girls of these classes as a group are more backward than the boys... Obviously the girls of a mental grade corresponding to the brighter boys in the backward classes, were left in the regular classes."

Reference should here be made to the findings of The Royal Commission of Great Britain, of 1904.[6] This Commission sought to enumerate the feebleminded outside of institutions, by sampling certain sections of the country, and applying a social-economic criterion to the inhabitants. The definition of "feebleminded person" which they adopted was, "One who is capable of making a living under favorable circumstances, but is incapable from mental defect existing from birth or from an early age (a) of competing on equal terms with his normal fellows or (b) of managing himself and his affairs with ordinary prudence." Upon application of this criterion, the Commission found more feebleminded males than females, but there is every reason to suppose that this criterion applies unequally to the sexes. The same supposition holds for their definitions of "imbecile" and "idiot."

Within the last five years, fairly large samplings of school children, chosen as nearly as possible at random, have been measured objectively for intelligence, with results which are not consistently the same. Terman[7] found no sex difference in numbers of the very stupid identified, in measuring 905 school children, selected at random. Pressey,[8] in testing a random sampling of 2,544 school children in Indiana found boys to be more variable, and to preponderate among the very stupid. In another sample of 1,408 school children, the same investigators found a similar result, while in a sample of 787 children limited to the ages 12 and 13 years, no sex difference in variability was found, but boys preponderated among the very low scores (girls [p. 37] preponderating at the opposite extreme). Frasier[10] in mental tests of 1,550 unselected school children in California, found no sex difference in variability, and no preponderance of inferior intelligence among boys.

Studies by means of educational and mental measurement, made previous to 1914, which bear upon the question of sex differences in intellectual deviation, have been reviewed elsewhere by the present writer,[11] and will not be noted here. In general, these studies were based on inadequate samplings, and yielded no conclusive results.


The present investigation proposes (1) to throw additional light upon the question of the frequency of extreme deviations in intelligence, as related to sex, (2) to pass upon the validity of the census of the segregated, as a measure of sex differences in mental variation, and (3) to give an account of the extent to which segregation may be differential, as it affects boys and girls.

The method is to analyze large samplings respectively of those who are brought for mental examination because they are thought to be deficient, and of those who have been actually segregated. As preliminary to the collection of data, the following specific questions were formulated:

1. Are males and females brought in equal numbers from the general population, for examination as suspected mental defectives?

2. Are males and females committed in equal numbers to institutions for mental defectives?

3. Are males and females, brought for mental examination as suspected mental defectives, and those committed as such, equally distributed in birthday age?

4. Are males and females so brought, and those so committed, of equal birthday age, mental age for mental age?

5. Are males and females so brought, and those so committed, equally stupid, age for age?

6. What inferences are to be drawn from a census of the segregated, regarding the relative frequency of extremely low intelligence as related to sex? [p. 38]


The present writer undertook the investigation of some of the above propounded questions first in 1913, and in that year some of the data here included were published.[12] The material was, therefore, gathered in two separate parts, at two different times, as follows.

1. In 1913, 1,000 consecutive cases of suspected mental deficiency were transcribed from the files of The Clearing House for Mental Defectives, at The Post Graduate Hospital, in New York City. This was a public clinic, from which children or adults, if found to be feebleminded, might be officially committed to appropriate institutions. Individuals of any age, from any borough of Greater New York, if suspected of inadequate intelligence, were admissible to this clinic, for mental examination.

Measurement of intelligence was made at that time by means of Goddard's Revision of Binet's Measuring Scale. The mental examinations were made by psychologists duly appointed to serve on this clinic, among whom was the present writer, who made about one third of the measurements recorded here. Of the 1,000 consecutive cases tabulated, 117 had no record of mental measurement, because of sensory defect, language difficulty, non-co-operation, and so forth.

2. In 1921, 1,142 cases were taken from the official files of The Psychological Laboratory, in the institution maintained by The Department of Public Welfare, of New York City, called The Children's Hospital, of Randall's Island. This is the municipal institution for the detention of the feebleminded. The mentally defective of the City are received here without limitations as to age, and receive training in the schools conducted there, though the institution is called The Children's Hospital, the great majority of inmates being children of school age, as will be seen from the data presented.

The data here used were obtained by transcribing all commitments, resident at any time since the establishment of The Psychological Laboratory, alphabetically by surname, through A, B, C, D, E, F, G, and H. All cases falling under these categories were transcribed in order, both those who were inmates at the time of transcription, and those who had been inmates, except epileptics, and those whose mental tests were incomplete. The latter were cases either just admitted to The Island, or too ill to be brought for mental examination, and were distributed in chance proportion between the sexes. [p. 39]

The method of mental measurement was The Stanford-Binet, in about 90 per cent of cases, the remainder having been measured by Goddard's Revision of Binet's Scale, by Pintner's Scale of Performance Tests, or by Kuhlmann's Extension of Binet's Measuring Scale. The last named was used in grading cases under 3 years mentally. The mental measurements were made by the psychologists duly appointed to serve in The Psychological Laboratory, after civil service examination, or by assistants directly under their supervision.

These two distinct collections of data may be regarded as "check samples" for each other. In each case the date (1913 or 1921) refers to the year in which the transcription of cases was made.

There are certain reasons for supposing that the results here obtained from the study of segregated cases might differ somewhat from those obtained from a study of all segregated mental defectives in the state. The institution studied is located in a city, and there probably is a tendency to send boys and men to country institutions, where they can work upon the land. More males than females are transferred to up-state schools, as the files show. It seems highly probable that more males may be sent to the country in the first place, since that is permissible. Again, the state maintains a separate institution for feebleminded women of child-bearing age, and this may affect somewhat the number of women of that age found in this study of inmates (1921 data).

These possible selective influences do not, however, affect the data from those brought for examination (1913), as these cases were uncommitted at the time of examination.


Are males and females brought in equal numbers for mental examination, as suspected mental defectives?

Are males and females segregated in equal numbers in an institution for mental defectives?

Of the 1,000 consecutive individuals examined at The Clearing House for Mental Defectives, 568 were males and 432 were females. Of the 1,142 cases listed in the Randall's Island institution alphabetically, 603 were males and 539 were females.

The preponderance of males brought for examination is greater than the preponderance of males actually segregated here, probably for two reasons. The first of these is that already mentioned, namely [p. 40] that there may be a tendency to commit males to state schools, where agriculture is possible. The second probability is that considerable numbers of the males presented are not committed, because they are found on test to be of a status higher than that where commitment can be justified.

It is true that more males than females are born, in a proportion of about 106 to 100.[13, 14] Statistics of mortality, however, show a somewhat higher death rate among male infants, so that by adolescence the numerical equality of the sexes is usually considered to be established. No statistics are available to the present writer from which the exact numerical ratio of the sexes at each birthday age, in New York City, over the decade studied, can be learned. The federal census in 1910 gave a proportion of 1,029 males to 1,000 females, in New York City, and in 1920 gave the proportion as 998 males to 1,000 females. The preponderance of males examined, and the preponderance of males segregated, in New York City, is greater at any rate than the preponderance of males born, or living in New York City during the time studied, as follows:



Are males and females, brought for mental examination as suspected mental defectives, and those committed as such, equally distributed in birthday age?

This question is answered in the tables and graphs immediately following. [p. 41]

[p. 42]

[p. 43]

[p. 44]

From the data it is clearly seen that the preponderance of males is due to the presence of boys under 16 years of age. Males are brought for examination, and are segregated, at a relatively early age. Girls escape the pressure that brings about identification longer than do boys. It is obvious that many women who were in childhood of a status warranting segregation, escaped till the age of twenty or thirty years, or longer, in ways which were not open to the boys and men in question. Since many die before the age of 16 years, it follows that many females escaping examination and segregation at the ages when males are most frequently examined and committed, never come under those influences which cause the curves of distribution to cross at 16 years, because they have died in the meantime.

Also, there is no reason to suppose that factors which finally become operative to bring feebleminded girls for examination and commitment after 16 years of age, are equal in pressure to those which bring feebleminded boys before 16. Even if all the feebleminded lived to be just 70 years old, let us say, there is no reason to suppose that as many females as males would be ultimately identified by present social pressures.

Our second question relating to comparative age is this: Are males and females, brought for mental examination as suspected mental defectives, and those so committed, of equal birthday age, mental age for mental age? [p. 45]

[p. 46]

From Table II, Parts 1 and 2, it is clear that females who survive in the ordinary schools, and in society, till past the age of adolescence, do not do so because they are of better mentality than the corresponding males. Females of any given mental age escape examination and segregation longer than do males of the same mental age, -- this difference becoming very marked above the mental age of 6-7 years. A female with a mental age of 6 years has as good a chance to survive inconspicuously in the educational, social and economic milieu of New York City as has a male of a mental age of 10 years.

That the greater age of females at each mental age does not result simply from greater longevity of females, is clear from the distributions in Figures 1 and 2. [p. 47]


Are males and females, brought for mental examination as suspected mental defectives, and those so segregated, equally stupid, age for age?

In seeking the answer to this question, the IQ's have been distributed by sex, and by birthday age, and the average IQ under each birthday age has been found for each sex separately. Also, the IQ's have been distributed by sex, disregarding birthday age. (IQ is here taken to mean the ratio between birthday age and mental age, by whatever scale of, measurement was used.) In almost all of the cases here distributed, the scale was either Stanford-Binet (1921) or Goddard's Revision (1913).

In these tabulations a difficulty arises through the lack of exact measurement below the level of 3 years. As before stated, a few cases testing below 3 had been measured by Kuhlmann's method, and the IQ's for these can be included. All who remained unmeasured, being listed in the records simply as "below 3", were omitted from the distributions.

Very few of the children sufficiently stupid to be presented or segregated early, register as high as 3 mentally by their sixth birthday. Not enough children under 7 by birthday age registered 3 or more mentally, so that distribution of IQ's under that age would be worth while. The calculations begin, therefore, with the 7-year-olds. There is no reason to suppose that the comparison would result differently, if all individuals below 3 mentally could be included in the distribution of IQ's. More males than females are below 3 in mental age, but this does not mean that there are more very low-grade intelligences among the males. Inspection of Table IV will show that more of the males than of the females below 3 mentally are relatively young children, so that their IQ's, when they have developed through the period of growth, may be relatively high.

By methods at present known, it is, of course, not possible to calculate the lowest existing IQ's. It is an interesting theoretical question as to how low IQ can run in the human species, and as to what "just not any intelligence," or zero intelligence, would be.

The tables and graphs which follow show the facts with respect to the comparative stupidity of the males and females, in the samples here studied. In making the graphs in Figures 5 and 6, all individuals [p. 48] over 16 years of age have been grouped together as adults, because after that point the cases at any one age are so very few that comparison from year to year has practically no reliability.

It will be seen that at the points which include the great majority of cases, namely among children of school age, the females have less intelligence than the males. After the age of 16 years, the difference disappears. For cases presented for examination after that age, the average IQ is 52, for both males and females. Among segregated cases the average IQ for adult females is 43 and for adult males, 45 with a large measure of unreliability.

It is interesting to see what the differences are between the presented, and the finally segregated. The IQ's of those presented run about ten points higher on the average than the IQ's of the actually committed. The difference in stupidity between the sexes is much more marked among those-actually segregated, than among those presented for examination. A girl must be relatively more stupid than a boy in order to be presented for examination, in the first place, and she must be still more stupid, comparatively, in order that she may be actually segregated, as unfit for social and economic participation.

 [p. 49]

[p. 50]

[p. 51]

[p. 52]

[p. 53]


The data here collected show that the pressures which bring about segregation of feebleminded school children are differential in their action upon the sexes, but they do not tell us what the pressures are. So far as the present study is concerned, interpretation of the nature of those pressures remains in the state of conjecture.

Girls segregated as mentally deficient are more stupid than boys segregated as mentally deficient; fewer girls than boys are segregated; and the girls who are segregated escape longer than do the boys.

These phenomena may possibly arise because (1) parents are more interested in their daughters than in their sons, and are more reluctant to part from them; (2) parents and other relatives are more interested in boys than in girls, and counsel is, therefore, more readily sought for deviating boys than for deviating girls; (3) girls and women have legal and customary means of economic survival, which boys and men do not have, and which require a low minimum of intelligence; (4) intelligence is not so important an element in the social-economic survival of girls and women, as it is in that of boys and men, so that girls who are extremely stupid do not give concern, and are not noticed, to the same extent as are boys of the same degree of actual stupidity; (5) boys, because they are less restricted, come more often into conflict with the law than do girls, and are thus scrutinized and referred more often by courts; (6) the subjective notion as to what constitutes intelligent behavior is different in the case of girls from what it is in the case of boys; (7) factors not included in any of these conjectures may be operative.

Data showing the social-economic status of the feebleminded over 16 years of age, among the cases presented for mental examination, are given herewith. They are suggestive as to a line of research which might prove fruitful, in an attempt to determine what are the forces, the differential action of which has been demonstrated. [p. 54]

The above tabulation shows that there is certainly a sex difference in the occupations followed by the feebleminded. Housework is the basis of survival for a very large percentage of feebleminded women. This may be for relatives, as paid domestics, as married women, or as common law wives. On the other hand, housework accounts for the survival of but one boy over 16 years of age. [p. 55]

Prostitution appears to offer the next most favorable basis for economic survival, among the feebleminded girls here found.

There seem to be no occupations which support feebleminded men as well as housework and prostitution support feebleminded women.

Data showing the social-economic status of feebleminded women have been gathered and presented elsewhere,[15] and are in harmony with the tabulation here made.

It is a matter of common knowledge that girls, as a group, are not expected continuously to follow competitive careers for a living. It is expected that their work will be in the household, domestic work and child-rearing, performed by the majority non-competitively (as wives or daughters) and without a stipulated wage.

Men, on the other hand, form a competitive group, working in rivalry with each other for a wage. This is true of even the simplest work that men do, such as digging with a pick, or loading sand on carriers. It would be expected a priori that the boy who cannot compete would become an object of concern.

The girl who cannot compete mentally need not become an object of concern to the same extent, because she may drop into the non-competitive vocational life of the household, where she "naturally" performs many routine tasks, requiring but rudimentary intelligence, such as peeling vegetables, washing dishes, scrubbing, carrying fuel, and so forth. If physically unobjectionable, as may be the case, she may marry, thus fastening herself to economic support in a customary fashion. Or she may become a prostitute, for survival is possible in that pursuit at a low intellectual level. Thus feebleminded girls fit in with the existing folkways, in a relatively inconspicuous fashion.

It should be worth while, from the standpoint of school and society, to determine just what the forces are which act to bring about differential segregation, and what the relative weight of each force is.

A question of incidental interest, arising from the graphs in Figures 1 and 2, concerns the infrequency of individuals over 30 years of age. Why are there so few "old people" in the samples studied ?Of those who are inmates at 12, at 13, or at 14 years of age, but few are left two decades later. Whether they die, or find adjustment elsewhere by that age, our present study does not inform us. It may be that most of the persons as low-grade mentally as those here studied have perished by the age of 40 years. [p. 56]

Another question of interest, not answerable from the data at hand, arises from Figures 5 and 6. Why should the IQ's become lower, as the birthday age of the individuals studied increases, so that adults have the lowest average IQ of all?

It may be that the higher IQ's among the children win discharge from the institution by the age of maturity. This would not, however, explain the same phenomenon in the 1913 data. The slight and gradual decrease in IQ among the very stupid, during the course of development, may enter as a factor.[16]

By intensive study of complete histories, these questions could doubtless be answered.


From the foregoing study the following conclusions are drawn:

1. More males than females are brought in New York City for mental examination, as suspected mental defectives.

2. More males than females are committed to the municipal institution for mental defectives, in New York City.

3. Males and females brought for mental examination as suspected mental defectives, and those so committed, in New York City, are not equally distributed with respect to birthday age. From 0 years to 16 years of age, there is a decided preponderance of males. From the age of 16 years, the reverse is true. The frequency of cases after 16 years is greatly reduced, and almost no cases past the age of 40 are involved.

4. Males and females, brought for mental examination as suspected mental defectives, and those so committed, are not equally old, mental age for mental age. At all degrees of stupidity females survive longer without examination or commitment than do males of the same mental status.

5. Males and females suspected of being mentally defective, and those segregated as such, are not equally stupid, age for age. The females are of lower grade than the males, up to the age of about 16 years. Among the adults here studied, the segregated males have an average IQ of 42.6, and the females have an average IQ of 45.0, while among the adults presented for examination both males and females have an average IQ of 52. [p. 57]

Since the great majority of ail cases falls under 16 years, the central tendency of IQ is lower for females than for males.

6. Institutional statistics, showing merely the numerical ratio of the sexes to each other among inmates, are invalid as an index of sex differences in frequency and amount of mental deviation, for the population at large. Such institutional statistics serve merely as an index of the extent to which it is comparatively easy and convenient for females who are extremely unintelligent to survive in the performance of their functions, in ordinary society.


The implications of this study are of interest for school and society. Among public school children there are probably more feebleminded girls than boys, since the girls corresponding to the higher grade boys under institutional supervision are evidently left in the population at large. Yet the registers of special classes for mentally deficient children show more boys than girls in such classes. Feebleminded girls will doubtless continue to drift with the regular classes, unless selection for special classes is made in a rigidly objective manner.

From the standpoint of society it is of interest to know that feebleminded girls fit into the existing social and economic order more conveniently than do boys of the same mental quality. Extremely stupid girls thus survive, and presumably reproduce their kind, more easily than do extremely stupid boys. The social order is such that survival for the former depends less upon intelligence than it does for the latter.

In so far as the preponderance of males among segregated feebleminded persons has been thought to support the theory of greater male variability in intellect, it must be said that a mere census of such persons is without validity as evidence.


Acknowledgments should be made of the courtesy and co-operation of Dr. M. G. Schlapp, Director of The Clearing House for Mental Defectives, The Post Graduate Hospital, New York City; of Dr. John S. Richards, Medical Superintendent, The Children's Hospital, Randall's Island, New York City; and of Dr. Louise E. Poull, Psychologist, The Children's Hospital, Randall's Island, New York City. They facilitated in every way the collection of data on which the study is based.

Table I, Part 1, Table II, Part 1, and Figure l, are here reproduced by courtesy of Wm. Wood and Co., publishers of The Medical Record.


[1] Kuhlman, F., The Part Played by the State Institutions in the Care of the Feebleminded. Journal of Psycho-Asthenics, 1916.

[2] Report of The Education Committee, City of Birmingham, England. .Quoted in School and Society, Oct. 22, 1921.

[3] Ellis, H., Man and Woman, 1909 ed.

[4] Goddard, H. H., Two Thousand Normal Children Measured by the Binet Measuring Scale, Pedagogical Seminary, 1911.

[5] Sylvester, R., The Form-Board, Psychological Monographs, 1913.

[6] Tredgold, A., Mental Deficiency, Wm. Wood and Co., New York, 1920.

[7] Terman, L. M., The Measurement of Intelligence, Houghton Mifflin, Boston, 1916.

[8] Pressey, L. W., Sex Differences Shown by 2,544 School Children on a Group Scale of Intelligence, with Special Reference to Variability, Journal of Applied Psychology, 1918.

[9] Pressey, L. W., Further Data with Regard to Sex Differences, Journal of Applied Psychology, 1921.

[10] Frasier, G. W., Sex Differences in Mental Traits and in School Achievement, Master's Essay, Stanford University, 1919.

[11] Hollingworth, L. S., Variability as Related to Sex Differences in Achievement, American Journal of Sociology, 1914.

[12] Hollingworth, L. S., The Frequency of Amentia as Related to Sex, Medical Record, 1913.

[13] Williams, J. W., Obstetrics, D. Appleton and Co.. New York, 4th ed., 1917.

[14] Rauber, A., Der Ueberschuss an Knabengeburten u. seine biologische Bedeutung. A. Georgi, Leipzig, 1900.

[15] Schlapp, M. C., and Hollingworth, L. S., An Economic and Social Study of Feebleminded Women. Medical Record, 1915.

[16] Kuhlmann, F., The Results of Repeated Mental Examinations of 639 Feebleminded Over a Period of Ten Years. Journal of Applied Psychology, 1921.