The Orphans of Davenport

by Marilyn Brookwood

  Typically, residents at the Glenwood Home had been born into ordinary Iowa families. When it became clear that they had limited intelligence, their parents were advised that life would be “better . . . for the rest of the family if this child [were] cared for someplace else.”1 Many agreed and placed their children at Glenwood, where they usually remained for the rest of their lives. But unlike the Davenport Home, Glenwood was led by a physician, Harold Dye, someone experienced in the care of those with low intelligence. Dye oversaw a program described as a “veritable beehive of activities,”2 with school classes geared to residents’ abilities; daily religious services; trips to local attractions; dances; and even movie showings. Skeels and Dye had worked together years earlier when Skeels had tested the intelligence of some Glenwood residents, so it was natural for them to collaborate on the Davenport children’s entry into Glenwood’s adult wards.

  Set on a hill overlooking the Missouri River valley, in clear weather Glenwood’s turreted administration building, resembling a minor castle, could be seen from miles away. On its carefully landscaped thousand-acre campus, the institution was a self-contained village that included a hospital, fire station, several playgrounds, a chapel, and a 160-acre farm where some residents worked. Others cared for the grounds and cultivated its apple orchards and gardens.3 Glenwood housed more than twice as many residents as the Woodward Home, in all about 1,800, including 623 children.

  Unlike Woodward, where only two wards housed women with mental ages from 5 to 9 years, Glenwood had seven wards that were home to women whose mental ages ranged from 5 to 12 years, making it suitable for the number of Davenport children Skeels would transfer. As the children arrived, staff members, along with low-intelligence women whom the courts had ruled were too intellectually limited to live freely, marry, and have their own children, helped settle them into their new home. Although the women’s chronological ages ranged from 16 to 52, everyone referred to them as “girls.” Residents with mental ages from 5 to 8 years comprised the lowest-level wards; women whose mental ages were from 7 to 10 years lived in the next level. The highest-functioning women, with mental ages between 9 and 12 years, lived together in two wards. There is no evidence that Skeels sought to have children placed with women of certain mental age groups; in fact, he wrote that the placements were “largely a matter of chance.”4

  Several times during their Glenwood stays, Skeels tested each child’s intelligence and found some relation between their progress and the intelligence of the women with whom they lived. While he suspected that placements with brighter women, or with women who were more caring, might promote higher intelligence, he held that information back, perhaps because his Glenwood investigation included many other variables. For example, at Woodward BD and CD experienced almost pseudo-adoptions, but at Glenwood a few Davenport children did not become the special charge of one woman, nor receive added attention from a staff member. Instead, they became “everyone’s” boy or girl.

  Had Harold Skeels’s contemporaries known of his Glenwood experiment, they would not have taken it seriously. Almost no mental test psychologists doubted the principle that IQs were constant, that children’s intelligence remained stable from birth to adulthood. However, at the time of Skeels’s experiment, there were at least three significant problems with existing investigations of constancy. First, psychology’s leaders denied that intelligence in young children could change, and therefore studies of that theory were considered a waste of resources and were seldom done. Second, almost no investigators tested whether IQ constancy held when the subjects’ environments were altered.5 And third, most psychologists neglected biology’s evidence that environment influenced development.6 Hence, studies that appeared to confirm constancy might only reflect the stability of the child’s environment. Skeels recognized that BD’s and CD’s shift from Davenport to Woodward defied the constancy interpretation. If he were right, his findings could profoundly call into question psychology’s understanding of intelligence. The outcomes in the children he transferred to Glenwood could be pivotal to discovering new knowledge about this process.

  Lewis Terman, recognized as the leader of mental test psychology, rejected out of hand the study of children younger than 4 years and scorned suggestions that early stimulation influenced later outcomes. Terman never addressed his reasons for this position, and psychologists who differed may have chosen to avoid an argument with someone known for sarcastic, even vitriolic attacks on those who disagreed with his ideas.7 However, in 1922, public intellectual Walter Lippmann suggested in the New Republic that “the earlier the [environmental] influence the more potent,” and that “the effects of superior and inferior environments persist,” and he stressed that Terman had conjectured about a period of child development that neither he nor anyone else had studied.8

  In reply, Terman sneered,

  And just to think that we have been allowing all sorts of mysterious, uncontrolled, chance influences in the nursery to mould [sic] children’s IQ’s this way and that. . . . It is high time that we were investigating the IQ effects of different kinds of baby-talk, different versions of Mother Goose.9

  Although Terman had little interest in the study of young orphanage children, he was quite interested in promoting the use of intelligence tests, and in 1918 he studied 68 children in a California orphanage, the Home of Benevolence, research with no defined hypothesis. Here he asserted, “Several clinical studies which have been made of children in orphanages, have revealed a large amount of mental inferiority and even of feeble-mindedness.” His remarks are of special interest because he provided no references and 74 percent of the children who entered the home were there because of a parent’s death.

  Curiously, Terman tacitly recognized the possible role of environment when he suggested that a good orphanage should stabilize or raise its residents’ intelligence because “the well conducted orphanage offers a cultural environment as good as that of the middle grade home, if not better.”10 Terman’s study found that the median intelligence for the 68 children was 92.3. Presuming their good intelligence was hereditary, he did not examine intelligence shifts experienced by the children during their orphanage stays. His paper stated no conclusions, but its interest lies in his suggestion that a control group could be used in future studies, one of the first occasions in psychology that a control group was proposed.11 It is also worth noting that although Terman claimed in this research that environment was a variable, he concluded that there was “no support to the environment hypothesis.”12

  During the 1920s, the decade before Harold Skeels and Marie Skodak started their work at Davenport, intelligence tests had begun to reshape American’s ideas about the evaluation of children’s ability. Where parents, teachers, and psychologists previously trusted subjective estimates, most now believed that IQ test instruments yielded reliable, scientific appraisals, and nearly all of the nation’s schools used them to evaluate students. That shift from impressionistic assessment to mental test results could not help but unite the fields of education and psychology.

  Terman published his IQ test in 1916 and in 1921 leading educators, educational psychologists, and mental test psychologists gathered to discuss the test at the decadal meeting of the National Society for the Study of Education (NSSE). The NSSE investigated many aspects of education and the psychology of learning and for each meeting published and presented for discussion a Yearbook collection of the latest research in the area under discussion.

  Along with many positive reviews of mental tests, the IQ test critiques of several psychologists took a more critical position. For example, Edward L. Thorndike, of Columbia University, advised that “only the baser parts of education can be counted and weighed, that the finer consequences for the spirit of man will be lost in proportion as we try to measure them.”13 Stephen S. Colvin, of Brown University, suggested that “the testing movement might grow out of all bounds; that it may be misunderstood . . . and [be] even harmfully applied in practice.”
He suggested those with “the greatest potential . . . will never become highly intelligent in an environment that affords scant opportunity to learn.”14 And a harsh warning came from another Columbia scholar, Marion Trabue. “The tests at present . . . are so inadequate and crude,” Trabue said, “that one who uses a single test score as the sole basis for a vital decision . . . is guilty of . . . unscientific practice and possibly of a great injury to the child.”15

  Those 1921 reports became a prelude to a 1922 skirmish when Columbia University psychologist William Bagley attacked Lewis Terman’s declaration that intelligence tests revealed ability that was hereditary and fixed from birth. “No theory in the whole history of science,” Bagley wrote, “has been based on . . . assumptions so questionable.”16 Bagley then challenged Terman’s nativist bias that low intelligence was “very common among Spanish-Indian and Mexican families . . . and also among negroes.”17 He also charged that Terman’s suggestion that schools provide “new types of secondary education which would be better suited to inferior intellects”18 subverted democracy.

  Terman’s riposte accused Bagley of preferring “to believe in ‘miracles’ rather than search for the order . . . which would make prediction [of intelligence] possible.” Terman then asked, “Does Dr. Bagley really believe that native differences in intellectual endowment are mythical? . . . [If so] his vision is blurred by moist tears of sentiment.”19 In a response to Terman, Bagley asserted that his criticism was “mild . . . compared with . . . appalling fallacies [of mental tests] pointed out by recognized authorities.”20

  Following Bagley’s criticism, investigations about possible environmental influences in the intelligence of young children modestly increased, with one of the most salient from Helen Thompson Woolley at the University of Chicago. In 1925, Woolley compared the IQ scores of children admitted to nursery school with a matched group of children on a waiting list and found that those who attended the program made significant gains. She noted, “Young children seem to be more rapidly and spectacularly affected by changes of environment than . . . older children.”21 But Woolley’s later work did not include studies of very young children, and for health reasons she left the field prematurely.

  By 1927, preparations were underway for the 1928 NSSE review of mental tests (the scheduled meetings were only roughly decadal), and Lewis Terman, selected to plan and chair the meeting, chose as its theme “Nature and Nurture.” As chair, he would also edit the meeting’s Yearbook. Although his NSSE leadership role required that he balance invitations and Yearbook articles representing all viewpoints, the Yearbook was slanted toward the hereditarian view, with four of thirty-eight articles, about 10 percent, from a single scholar, Terman’s prize graduate student Barbara Burks. No other psychologist on either side contributed more than one.

  The contentiousness of the 1921 debate led psychologists to anticipate that this meeting would be a call to arms. But concerned that his past responses to critics might work against him, Terman preemptively advised all invitees, “We should avoid everything that would savor of appeal to emotional bias. In making this suggestion I am not unconscious of the fact that I myself have ‘sinned’ in this respect.”22 Nevertheless, two outstanding psychologists rebuffed Terman’s invitations. The first, William Bagley, was no surprise. Bagley advised Terman to invite someone “who has not made the question so much of a personal issue.”23 Pressured by a Terman ally, Bagley agreed to attend but he would present no research.

  The second refusal came from one of Terman’s own, Carl C. Brigham, of Princeton University. Following World War I, Brigham and his colleague, Robert M. Yerkes, had analyzed IQ test data collected from army recruits. Along with Terman, Brigham and Yerkes were members of the American Eugenics Society (AES), the largest and best-funded eugenics organization in the nation.24 In the foreword to Brigham’s army report, A Study of American Intelligence, Yerkes advised that “none of us . . . can afford to ignore the menace of race deterioration or the evident relations of immigration to national progress and welfare.”25

  Brigham’s study asserted that the gradient of recruits’ intellectual superiority to inferiority measured by their test scores tracked with their national backgrounds and “race.” His conclusions fell into the envelope of nearly identical “facts” published in 1917 by his mentor, eugenicist Madison Grant.26 Heredity, Brigham wrote, explained why the highest intelligence scores were found in recruits from Nordic backgrounds. He claimed that the lowest scores were from eastern and southern European recruits, followed by African Americans.27 But Brigham went even further when he claimed to have calculated “the proportion of Nordic, Alpine and Mediterranean blood in immigrants from each of the European countries.”28 Brigham also found that immigrant recruits who had lived in the United States for over sixteen years had test results “almost as high as native born” recruits, but reasoned, “we are measuring native or inborn intelligence and any increase . . . due to any other factor may be regarded as an error.”29

  Yet Brigham declined Terman’s invitation to join an NSSE panel on nature and nurture, sending a telegram that said he “positively cannot undertake criticism.” In an explanatory letter, he wrote, “My own opinions on the subject have undergone radical alterations and I have not yet reached any solution which is satisfactory.”30 By 1929, Brigham told eugenicist Charles Davenport, “The more I work in this field the more I am convinced that psychologists have sinned greatly in sliding easily from the name of the test to the . . . trait measured.”31 Then, in an astonishing 1930 paper, The Intelligence of Immigrant Groups, Brigham recanted his 1923 findings about national and racial groups and in a self-scorching confession said, “One of the most pretentious of these comparative racial studies—the writer’s own—was without foundation.”32

  What led Brigham, who earlier assured Yerkes, “I am not afraid to say anything that is true, no matter how ugly the facts,”33 to fundamentally revise his view of what the facts of hereditary intelligence actually were? Brigham’s reconsideration reflected his recent discovery that the Mendelian argument that human traits were inherited as single units was insufficient to explain behavioral outcomes. According to eugenics scholar Garland Allen, of Washington University in St. Louis, during the 1920s American geneticists—Thomas Hunt Morgan and Herbert S. Jennings were two—“had advanced arguments against the scientific credibility of eugenics.”34 Then, in 1927, biologist Raymond Pearl, of Johns Hopkins University, once a fervent adherent, also disavowed eugenics, portraying it as “a mingled mess of ill-grounded and uncritical sociology, economics, anthropology and politics, full of emotional appeals to class and race prejudices, solemnly put forth as science, and unfortunately accepted as such by the general public.”35 Brigham paid attention to these defections, and they moved him.

  In that dynamic context, Brigham told Terman that recent discoveries—he labeled them “very much more important than any investigation included in the Yearbook”36—suggested that the army intelligence test erred in presuming that intelligence was a unitary inherited function. Here he referenced the recently published work of Truman L. Kelley, of Stanford University, later of Harvard, who posited that verbal, quantitative, and spatial abilities were probably separate from one another,37 that mental ability was not a single characteristic but an assembly of skills. Brigham also recognized that accurate test results had to compare “only individuals having equal opportunities” to acquire the language used on the test.38 By 1934, his analysis had turned even more damning: “The [mental] test movement . . . accompanied one of the most glorious fallacies in the history of science . . . that the tests measured native intelligence without regard to training or schooling. I hope nobody believes that now. . . . The ‘native intelligence’ hypothesis is dead.”39 Brigham did not reject ability testing per se. His was a moral argument—he opposed the tests because he recognized their potential to injure individuals who had not had the opportunity to learn the information or skills being tested.

  While 1920s ge
neticists and biologists had abandoned eugenics en masse, unlike Brigham, few mental test psychologists followed, and into the late 1930s desertions from the Terman camp were rare. Later, Brigham assumed a leadership position at the College Board, where he advanced another test, “a metamorphosis from the Army . . . test,” the Scholastic Aptitude Test.40

  Although Lewis Terman weighted the NSSE 1928 Yearbook to favor hereditarian views—it came to be known as the Terman Yearbook—he also included challenges to that dogma. For example, Barbara Burks’s fair-minded analysis of every research paper in the area of nature versus nurture from the 1869 work of Sir Francis Galton to papers published in 1927—a total of 279—was often skeptical of data that supported claims of an association between intelligence and heredity. Burks also criticized the findings of her mentor, Lewis Terman, whose 1918 paper showed that environment had negligible influence on orphanage children.41

  Following the 1928 NSSE meeting, two new reports suggested that environment counted. The first, a modest 1930 examination of young orphanage children, came from Helen E. Barrett and Helen L. Koch, at the University of Chicago, and showed that the mean intelligence score for children who attended nursery school rose significantly compared with a matched group that was not in school: after six to nine months, the mean IQ score of the schooled children rose from 91.71 to 112.57. The mean IQ of children in the control group rose from 92.59 to 97.71.42

  Subsequently, in 1932, Mandel Sherman, a psychologist at the University of Chicago, and Cora B. Key, from the Washington Child Research Center, published an ambitious field investigation of a deprived population. In The Intelligence of Isolated Mountain Children, the researchers compared residents of four Blue Ridge Mountain hollows, each separated from the next by a high mountain ridge. The hollows’ populations descended from immigrants who, as newer immigrants arrived, moved deeper and deeper into the mountains.43 Nearly all of the hollows’ adults, including the parents of the study’s subjects, were illiterate, and almost none were employed.