According to the Philadelphia-based Institutes for the Achievement of Human Potential, it has produced a noninvasive treatment for brain damage. Details are contained in a popular book by Glenn Doman, founder, in 1955, of the Institutes.1 Doman-Delacato patterning therapy (DDPT) was developed in the 1940s and the 1950s by Doman, a physcial therapist; Temple Fay, a neurosurgeon; Robert Doman, a physiatrist; and Carl Delacato, a psychologist. Its core assumption is that brain damage causes a blockage in the normal pattern of brain development. The consequences of this blockage can allegedly be eliminated through what is termed “patterning” therapy, exercises that supposedly rewire the brain.
Before reviewing empirical tests of the Doman/Delacato exercises, I will first outline the theoretical bases of the technique. Central to the patterning approach is the long-discredited view that “ontogeny recapitulates phylogeny.” According to Delacato,2(p5) “The ontogenetic development of each individual’s nervous system, in general, recapitulates that phylogenetic process.” Thus, the therapy is based on a view of the development and organization of the brain that is simply wrong. This is most apparent in the “developmental profile” found in Doman’s 1999 book.1 Here, he divides the central nervous system into 7 areas: (1) spinal cord and medulla, (2) pons, (3) midbrain, (4) initial cortex, (5) early cortex, (6) primitive cortex, and (7) sophisticated cortex. These divisions of the cortex correspond to no accepted cytoarchitectonic regions. Doman’s “initial” cortex, for example, includes the inferior occipital and posterior inferior temporal lobes. His “early” cortex is above that. The “primitive” cortex is above that, running in a strip from the inferior frontal lobe back to the middle parietal and superior occipital lobes. The “sophisticated” cortex sits like a cap atop the brain.
Compounding these errors, Doman’s profile is divided into 6 areas of “competence”: visual, auditory, tactile, mobile, language, and manual. In the resulting 7-by-6 table, each of the 42 cells is assigned 1 or more functions said to be characteristic of that division of the brain for the “competence” in question. Doman’s chart is so full of neurological misinformation that I use it as an exam question in my junior/senior-level physiological psychology class. Students get 1 point for each error they identify. A few examples will suffice. Outline perception is said to be done by the pons, as is “vital response to threatening sounds.” The midbrain does “creation of meaningful sound” and, in the area of visual competence, it is credited with “appreciation of detail within a configuration”; and, in audition, with the “appreciation of meaning sounds.” Thus Doman and his collaborators demonstrate unfamiliarity with even the basics of neuroanatomy and localization of function. It is as if someone offering a wonderful new auto repair system were to assert that fuel combustion takes place in the tires and that the transmission makes the radio work.
As I have noted elsewhere,3 a reliable sign of a quack remedy is the claim that the same intervention will be effective for a variety of different diseases with diverse etiologies. Note that the full title of Doman’s1 book is: What to Do about Your Brain-Injured Child or Your Brain-Damaged, Mentally Retarded, Mentally Deficient, Cerebral-Palsied, Spastic, Flaccid, Rigid, Epileptic, Autistic, Athetoid, Hyperactive, Down’s Child. In effect, Doman is claiming that patterning therapy will cure dozens of different disorders with causes ranging from genetic to traumatic. Such a sweeping assertion is absurd.
A few of Doman’s published examples of patterning treatments will provide a feel for the nature of DDPT. Doman1(p61) describes the technique of “homolateral patterning” as it is applied to a child “hurt in the pons”:
This pattern was administered by three adults. One adult turned the head, while the adult on the side to which the head was turned flexed the arm and leg. The adult on the opposite side extended both limbs. As the head was turned, the flexed limbs extended while the extended limbs flexed. We found that when this pattern was administered rigidly enough and frequently enough, in a timed pattern, many children hurt in the pons began to crawl. When they became able to crawl, they would rapidly move through creeping to walking and do well.
The same patterning approach is also allegedly applicable to sensory impairments. Consider the case of a 10-month-old girl named Mary who is, according to Doman,1(p157) “for all practical purposes deaf.” Here, patterning therapy takes advantage of the fact that Mary still has a normal startle reflex, which means, of course, that she is not profoundly deaf. In this version of DDPT, Mary’s mother will “stimulate her auditorially every waking half hour. . . . Mother will do so by unexpectedly banging two blocks of wood just behind Mary’s head. She does so ten times at three-second intervals in each of twenty-four sessions.”
Or take the case of Sean, who apparently can distinguish hot water from cold, but not warm from cool. This indicates to Doman that he has a blockage at the third stage of “tactile competence.” To overcome this blockage, Sean’s mother will dip his hands, alternately, in warm and cool water 600 times every day.
On the face of it, such far-fetched therapeutic regimes as DDPT would seem quite unlikely to yield any improvement, although the cruel and reckless treatment of Mary might very well engender a good deal of fear and anxiety. One would expect that proponents of such implausible techniques would feel a special obligation to back up their therapeutic claims, but this is not the case. In fact, the Institutes for the Achievement of Human Potential have shown very little interest in providing empirical support for their methods. In 1967, a “well-designed, comprehensive study (supported by both federal and private agencies) was in the final planning stage when the Institutes withdrew their original agreement to the design.”4(p1215) Here, the Institutes foreshadowed the sort of behavior that has come to typify the majority of practitioners of what is now known as “complementary and alternative medicine” (CAM); i.e., pay lip service to the need for the empirical evidence they ought to have amassed before selling their treatments, promise to cooperate with critics in supplying it, then renege on the commitment after reaping the public relations benefits for having made the promise.
The founders of DDPT have published only one study of its effectiveness in the medical literature.5 In this 1960 paper in the Journal of the American Medical Association, the subjects were 76 brain-damaged children with both “traumatic and non-traumatic lesions,” but none who were “genetically defective.”5(p261) The children were given patterning training for at least 6 months. The authors report that they “found significant improvement when we compared the results of classic procedures we had previously followed with the results of the procedures described above.”5(p261) Unfortunately, the paper does not contain a description of what the “classic” procedures were. At the beginning and end of the study, the severity of movement impairment for each child was rated. The article does not state that these ratings were blinded, and they almost certainly were not. It is not even reported who the raters were. Nor were any statistical analyses of the results presented.
In three books, Delacato2,6,7 summarizes 11 studies that he claims support the effectiveness of DDPT in treating reading problems. None of these studies was published and only 2 were by Delacato himself. All suffer from serious statistical and/or methodological flaws. These shortcomings were discussed in detail by Glass and Robbins,8,9 on whose critiques much of the following is based.
In his 1959 book, Delacato6 reports an 8-week “case study” of 30 reading-retarded third through fifth graders who were “taken from a number of schools.”6(p99) Among other things, the children “all had the subdominant eye occluded for one hour each morning and one hour each evening at home.”6(p99) The children were forbidden to listen to music or sing because music, according to Delacato, is an activity of the nondominant hemisphere and thus musical activities will result in “unnecessary use and therefore activation of the subdominant area.”6(p25) The children were also lectured about the proper posture to maintain during sleep. Apparently the same reading test (or tests) were used at the end of the program to evaluate change as had been used at the start to classify the children. No untreated control group, let alone a placebo control group, was run. The median “reading growth” was nine months. No further statistical analysis was presented. There is no hint that Delacato saw the need to control for confounds such as practice and placebo effects. Rather, he concludes that the observed change is due to “changes in the neuro-organization of the children.”6(p100)
Delacato’s 1963 book7 describes two more studies of DDPT. The first was by Gayle Piper, a special-education teacher at an Arizona high school. Subjects were 14 special education students. Prior to patterning therapy, they took the Gates Basic Reading Test (GBRT). Subsequently, for 6 weeks, the “sub-dominant arm of each student … [was] tied down to the body during special education classes every day.”7(p157) A second version of the GBRT was then given, followed by 6 more weeks of therapy, and a third GBRT form, which was administered at the end of the school year. A fourth form of the GBRT was given at the start of the fall term “to see what effect a three month summer vacation had on the reading abilities of the student.”7(p166) The means and standard deviations (in parenthesis) for the four testing sessions were, respectively: 5.07 (1.28); 5.55 (1.33); 5.43 (1.41) and 6.24 (1.48).
Delacato’s book7 offers no statistical analysis for these means. Robbins and Glass9 computed change scores from the first to the second test administration and computed a t-value based on these scores, which was significant at the .01 level. However, a t-test comparing only the first and second tests omits half the data from this study. A more appropriate analysis would be a one-way analysis of variance including the data from all 4 testing sessions. The present author performed such an analysis. One data point was missing, because one child had transferred to another school. This missing cell was filled using the method of Winer et al.10(p480)
As expected, there was a significant effect of testing session [F(3, 39) = 20.26, p < .001]. This shows only that the 4 means differed statistically. A Tukey post-hoc test was then applied. This allows comparisons of individual means, but controls for the increasing likelihood of finding significant differences by chance alone as the number of such comparisons grows. The critical difference at the .05 level was 1.4. Thus, while the overall pattern of test score changes across the 4 tests was statistically significant, no 2 pairs of means, considered individually, were significantly different. It is, nonetheless, important to note that the largest difference between 2 adjacent means occurred for tests 3 and 4, given at the end of the spring term and the beginning of the next fall term, respectively. The only “intervention” during this interval was summer vacation.
Delacato7 nonetheless takes these results as support for his methods. In reality, they provide virtually none because the study lacks both placebo and no-treatment control groups. In the absence of these comparisons, it is impossible to distinguish any putative benefits of the therapy from those of several possible confounds. The latter include practice effects from taking similar versions of the same test four times, simple maturational effects, the effects of additional experiences over time, and improvements due to regular school instruction. That maturational factors had a major effect can be seen from the fact that the biggest single improvement (.81) is found between the means for the assessments before and after summer vacation, a period in which no therapy (or classroom instruction) was given. By contrast, additional therapy was given between tests 2 and 3, but there was no improvement in reading scores. In fact, scores decreased by a statistically insignificant amount.
Robbins and Glass9 also fault Piper’s study with respect to a possible regression artifact, arguing that because the students scored lower than normal on their initial reading assessment, their scores would be expected to regress toward the mean and thus improve on the next test. It is not clear, however, that regression is a problem here. The participants were special education students and thus scored below the average of the entire school population on this reading test. However, they were not compared to normal students in this study, nor is there any indication that they were picked because they were below the mean of the special education population in the school. Thus, regression does not seem to be an issue. Nonetheless, this study suffers from enough other methodological deficits to render it useless as support for DDPT.
If regression effects are not a problem for the Piper study, they dominate the second study reported by Delacato7 This study was apparently done by Delacato’s group itself, as the pronoun “we” is used throughout. The participants consisted of all 25 students in the junior class of an all-male private high school. The verbal score on the Scholastic Aptitude Test (SAT) was the dependent measure. On the basis of their first SAT results, the boys were divided into two groups: a patterning therapy group containing 9 students and a control group containing 16 students. No rationale was offered for not dividing the students more equally. Astonishingly, the experimental group was intentionally chosen to include the 9 students who scored the lowest (X=398) on the verbal section of the SAT. The control group consisted of the 16 boys who scored the highest (X=547). After six weeks of DDPT, the SAT was readministered. On the second test, the control group showed an improvement of only 7 points, the new mean being 554. The experimental group, however, showed an increase of 66 points, their new mean rising to 464. It is almost as if the study had been designed to demonstrate a regression artifact.
Delacato7 does address the issue of why the control and experimental groups were not matched, but his rationale shows no awareness of the problems regression artifacts can create. He says, “We did not match the groups because of the great importance to each boy of the test scores. We felt that it would not be ethical to possibly sacrifice the opportunity to go to college for a group of boys merely for the sake of purity of experimental design.”7(p172–73) Did they not consider that if the therapy had really proved effective in a “pure” experimental design, it would have been a simple matter to provide it to the control group following completion of the experiment?
Moreover, it is unlikely that the subjects in the experimental group actually received much therapy. Delacato describes the therapeutic intervention, which “varied according to individual needs from creeping to visual training”:7(p172) “Each boy in the experimental group was diagnosed and was taught to follow a program of neurological organization for one-half hour per day without supervision for a six week period” (emphasis added).7(p171–72) In other words, no one ensured that the experimental subjects crawled around, occluded one eye, and so on, during this 6-week period. How many male high school juniors would comply, on their own, with an odd regimen like this for 6 hours, let alone 6 weeks?
It is in Delacato’s 1966 book2 that the majority of the 11 unpublished studies he considers support for DDPT appear. Eight, none conducted by Delacato, are described. The relative inexperience of these investigators in conducting and evaluating experimental research apparently did not concern Delacato.
The first study was conducted by a Sister M. Edwin at a Chicago-area Catholic school. The study started with 108 kindergarten children but, for unspecified reasons, only 84 actually took part in the entire study, 41 in a control group and 43 in the experimental group. Over a 5-week period, the experimental group received 105 minutes daily of patterning activities such as creeping and cross-pattern walking. The control group received no special treatment whatever, except that their mothers or older siblings were “asked to read or tell a story to the child for at least ten minutes daily.”2(p51)
A battery of 7 unspecified tests was given at the start and the end of the 5-week period. It is not stated whether these evaluations were blinded. No statistical analysis is presented but it is claimed that there was a 79% increase in something called “controlled attention span” and a 70% increase in “uncontrolled attention span.”2(p52) It is not clear whether these differences are between the experimental and the control group at the end of the program or between the experimental group at the start compared to the same group at the end of the program. Either way, the amount of personal attention that children in the experimental group received, rather than any specific benefits of the patterning procedures, could easily account for these observed differences.
The second study reported by Delacato2 used middle-class children who had reading problems. They were attending summer remedial reading classes but their ages and grade levels were unspecified. Nineteen teachers participated in the experiment. Each taught 1 class of children from the experimental group and 1 class of children from the control group. DDPT was provided outside of class, requiring that children in the experimental group come to school 15 minutes early and stay an additional 15 minutes after class for the 6-week duration of the trial. Probably because of this, each teacher’s first morning class was designated the experimental class while the second class was always the control group.
The study began with 422 subjects, though only 242 completed the 6-week trial. The relative dropout rates of experimental and control subjects was not reported but Robbins and Glass9 note that this high attrition was unlikely to have been distributed evenly between the 2 conditions. Because the regimen imposed on the experimental group was much more onerous than that of the control group (i.e., nothing), we can be reasonably sure that the dropout rate was substantially higher in the former. Disproportionate thinning of the experimental group could have affected the outcome in various ways. First, the classes in the experimental group would become smaller, allowing the teachers to give additional attention to each child, compared to those in the larger control classes. Second, it is likely that those who dropped out were, on average, less motivated and less able readers than those who stayed. Finally, differences in parental involvement could have been related to the differential attrition rates, making it likely those who stayed in might have received more help and encouragement at home as well. Such differences could have produced a spurious treatment effect for the remaining experimental subjects.
In addition to the foregoing problems, the absence of a placebo control and blinding of evaluators also suggests alternative explanations for the observed results. It is likely that some combination of these artifacts was responsible for the small, but statistically significant, improvement for the experimental group.
A Father Francis McGrath conducted the next study reported by Delacato.2 The 92 third- to eleventh-grade participants were recruited from those who were reading below grade level. They were given about 45 minutes of DDPT a day, 5 days per week, for 6 weeks. There was neither a nontreated control group nor a placebo control group. At the end of the 6-week program, there was a statistically significant improvement, amounting to sixtenths of a grade level. However, due to the lack of experimental controls, once again, any improvement could easily be accounted for by regression to the mean and placebo effects.
In his chapter 14, Delacato2 recounts another study of DDPT, by Ruth Kabot of the Morton Street School of Newark, NJ. Twenty-two third graders were equally divided into an untreated control group and an experimental group. The latter received daily half-hour patterning treatments over an 8-week period. The 2 groups were “carefully matched as to IQ, reading scores, reading retardation and laterality.”2(p119)
A major flaw in this study was that different tests of reading were used before and after the therapeutic intervention. The Stanford Reading Test was used as the pretest and the California Reading Test as the post-test. Given that different tests were used, it is hard to see how pretest and post-test scores could be compared. Kabot reported that the control group showed a gain of “6 months” over the 8-week intervention, but this figure is reported as a decimal (“.6”) in the table in the report. The experimental group showed a gain of 8 months (or “.8”). This difference was not significant.
In a footnote, Delacato himself reports a follow-up of these subjects, one year later, to see if the experimental group showed further improvement over that time. However, only 7 pairs of students appear in the followup. No mention is made of why 5 pairs were not included. Nor does Delacato specify what test was used for the follow-up evaluations. He does not report the mean scores at follow-up, but simply asserts that the experiments group showed a .54 (units unspecified) greater improvement than the control group and that this was statistically significant at the .05 level.
Chapter 16 of Delacato’s book2 is devoted to a study of high school students enrolled in a summer remedial reading clinic. Students attending the 8 A.M. class were designated the experimental group. Every day for 7 weeks, these students spent 1 hour in the gym doing patterning therapy. This period represented half of their total daily class time. The control group was made up of students in a 10 A.M. class who received no special treatment and spent the entire 2-hour period, with the exception of a short break, in the same classroom. There was no placebo treatment group. At the beginning and the end of the 7-week program, both groups were given the Nelson Test of Mental Abilities and 7 subtests from the Stanford Achievement Test. On 2 of these (paragraph meaning and word meaning) the experimental group showed significantly higher scores than the control group at the end of the program. The experimental group also scored significantly higher than the control group, post-treatment, on the Nelson Mental Abilities test.
The most obvious of the many problems with this study was the lack of a placebo control. As the report itself admits, the fact that the control group had to stay in the same classroom for 2 hours, while the experimental group was allowed to go to the gym for an hour, had a major differential effect. The authors of the study noted that “there was a problem of class control” in the control group: “There was an undertone of restlessness and poor attitude that was not present in the experimental group.”2(p141) In addition, some subjects were moved from one group to the other. And finally, 12 subjects (8 experimental and 4 control) unaccountably vanished from the data analysis. The report states, with no explanation, that there were 28 control and 22 experimental subjects, but data were analyzed for only 14 experimental and 24 control subjects.
A study by John Noonan Jr. of the Boston University School of Education occupies chapter 17 of Delacato’s book.2 Eleven sixth- and seventh-grade students participated. Nine were reading between 2 and 5 years below grade level. For the entire school year, 9 students (not necessarily the 9 reading below grade level) received 45 minutes of DDPT each school day. The remaining 2 students started the program late and so received patterning training for only half the year. The Iowa Silent Reading Test was the dependent measure. There was neither an untreated control group nor a placebo control group. On 4 of the Iowa test’s 6 measures of reading ability, there was a significant improvement over the course of the year, averaging 3 years and 3 months. There was an improvement of 1 year and 3 months on the other 2 measures, but this was not statistically significant. These results can easily be explained by a combination of regression effects (since most of the students were reading well below grade level initially) and the normal effects of a year of schooling, as well as a placebo effect. Once again, the study provides no convincing support for the patterning exercises.
Unlike any of the foregoing unpublished studies summarized by Delacato, the one by a Sister Alcuin (chapter 18) does include appropriate control groups. She employed 3 groups of 40 children each, ranging from 6 to 14 years old. They were enrolled in a 6-week summer school course. A nontreated control group received only the standard curriculum. A placebo control group of sorts received 3 20-minute periods per day of unspecified “psychological training” and some calisthenics. The experimental group received 3 20-minute periods per day of “neurological training,” i.e., DDPT.
Given the relative sophistication shown by the inclusion of 2 control groups, it is unfortunate that assignment of students to the 3 conditions was apparently not done randomly. Assignment was based on students’ pretest scores on the Stanford Reading Achievement Test and the individual teacher’s judgement. It appears that students who were the least able readers may have been placed in the experimental group. This group showed an improvement of .73 years reading level over the 6-week program. The nontreated and placebo control groups showed improvements of .40 and .42 years, respectively. The difference between the experimental and the 2 control groups was statistically significant. To the extent that the experimental group consisted of the more disabled readers, the observed differences would be due, again, to regression effects.
The final study in Delacato’s 1966 book2 is an edited version of a EdD dissertation by Brian Miracle at the University of Wyoming.11 Forty fourth- and fifth-grade students, all reading below grade level, were subjects. They were divided into 4 groups—2 received “neurological training” (DDPT) and 2 did not. Crossed with this variable was another in which students did or not receive a remedial reading program—2 groups did and 2 did not. In this 2-by-2 research design, the Iowa Test of Basic Skills was given before and after the 8-week remedial program. The test yeilds two relevant measures, vocabulary skill, and reading ability. Miracle reported mean pre- and post-intervention scores for each of the 4 groups on these 2 variables. Oddly, neither in Delacato’s2 edited version nor in the original dissertation did Miracle report doing an overall analysis of variance on his data. Instead, he presents a rather unhelpful table giving 24 Fisher’s t-values for each of 6 intergroup comparisons for each of the 4 variables (reading and vocabulary scores, pre- and post-intervention). Robbins and Glass criticize this procedure because it capitalizes on the chance significance expected with multiple testing. They apparently did not realize that Miracle (as he stated) was using Fisher’s t-test, which (unlike Student’s t) is a post-hoc procedure that controls for the effects of multiple comparisons (see Keppel12).
Nonetheless, Miracle’s table of 24 t-values, of which 7 are significant at the .05 level, is not very informative because he never reports comparisons of the 4 groups before and after the program. That is, he does not say whether the crucial within-group changes are significant. For the reading ability data only, Robbins and Glass9 estimate the variances of the pre- and post-test means and performed their own analysis of variance. They found that the scores of the 2 groups that received neurological training were significantly greater (using Tukey post-hoc tests) than the scores of the 2 groups that did not.
The question, of course, then becomes, why was this the case. Robbins and Glass9 correctly point out that important details about the experimental methodology are missing from Miracle’s reports. Specifically, they say, “One cannot learn from the research report whether the four groups had the same or different teachers, whether the groups met at the same or different times of the day, whether the subjects were treated individually or as intact groups.”9(p366) The issue of whether each group had its own teacher is especially important. If this was the case, the observed group differences could well be due to differences between the teachers and not to differences in the effectiveness of the treatments.
In summary, the studies reported in Delacato’s 3 books provide essentially no convincing evidence in favor of patterning therapy. Robbins and Glass9(p347) aptly concluded that these studies are
… exemplary for their faults. They were naively designed and clumsily analyzed. They suffer from a multitude of sources of invalidity. They appear to have been executed and reported in an atmosphere of relative insensitivity to basic considerations of empirical, experimental research.
The present author has found these experiments to be excellent sources of exam questions in his introductory statistics class. Students must read Delacato’s description of a study and then outline its design flaws.
Melvyn Robbins, a critic of the studies reported in Delacato’s books, published his doctoral dissertation in 1966.13 Interestingly, he notes that he spent 2 months “in residence with Carl Delacato.”13(p57) at the Institutes in Philadelphia. In Robbins’s experiment, 126 second-graders were divided into 3 groups. For 3 months, an experimental group received 30 minutes of several types of patterning during the half hour before school. A placebo group danced, played games, and received nonspecific patterning for a comparable period. The nontreated controls followed their normal before-school routine. All 3 groups were tested on their reading and arithmetic skills, intelligence, and laterality (hand, foot, and eye preferences), as well as their creeping ability, before and after the 3-month program. The creeping test was designed by individuals trained at the Philadelphia Institutes.
Robbins tested 6 hypotheses based on the Delacato approach. First, at pretest, creeping ability should correlate positively with reading ability. Three raters rated the creeping ability of all participants. Interrater reliability was high (average r = .76), but there were no significant correlations between rated creeping ability and reading scores, even at the .10 level. The average correlation was –.03.
The second hypothesis concerned laterality. Subjects who were lateralized should, according to Delacato’s perspective, show higher reading scores than those who are not. This was not the case. The third hypothesis, a variant of the second, stated that the relationship between reading and creeping should emerge when creeping ability was statistically controlled for. In fact, this did not yield any greater relationship between reading and laterality.
The fourth hypothesis was that the DDPT group would show greater improvement after training on reading than would the other groups. Since there had been significant differences in reading (and arithmetic) scores between the groups at pretesting, an analysis of covariance with pretest score as the covariate was used to test this hypothesis. It was not supported. The fifth hypothesis was based on the claim that patterning treatment is effective for boosting reading but not math skills. Analysis of covariance showed this not to be the case. Finally, the sixth hypothesis predicted that there would be more lateralized experimental subjects after training than before. Although the primary aim of DDPT is to increase lateralization, this was not found. Robbins’s study stands out in that not only did it test the therapeutic claims of Delacato’s method, it also tested specific theoretical predictions made by the underlying theory. DDPT was found wanting on all counts.
In his first experiment, Robbins studied normal students. In his second study,14 he used third through ninth graders who were attending a summer remedial reading program. A total of 149 students was divided into 3 groups. The experimental group received DDPT and related training, both at home and at school. The placebo control group took part in nonpatterning physical activities (games, sports, music, dancing, and so on), both at home and at school, for the same amount of time.
There were no differences between the three groups in the amount of reading improvement over the course of the program. Replicating Robbins’s13 earlier findings, neither creeping ability nor laterality measures correlated with reading ability. Robbins’s second study was published in the Journal of the American Medical Association. It was accompanied by a short commentary by Freeman15 pointing out the empirical and theoretical shortcomings of the Doman-Delacato approach. Robbins16 has also published a shortened summary of his work.
A study by Kershner, allegedly supporting DDPT, has appeared in three different versions. The original was a 1967 master’s thesis done at Bucknell University.17 It was published in booklet form the same year by the State of Pennsylvania Department of Public Instruction. And finally, it appeared in the scientific literature in 1968.18 I have been unable to obtain the first 2 versions of this study, but Freeman19 reviewed the Pennsylvania state publication of it and the following is based on his review. Kershner’s study examined 2 groups of children classified as “trainable” retarded. Thirteen children were given DDPT every school day for 74 days. Sixteen control children engaged in normal physical activity during the same period. At the end of the study, the DDPT group was found to be better than controls on measures of “perceptual motor proficiency in areas not practiced”19(p914) and on the Peabody Picture Vocabulary Test (an intelligence measure). The DDPT group gained 12 Peabody IQ points while the control group lost 3.
There are two major defects in this study.19 The 2 groups were not equated on pretest IQ scores. The mean score for the experimental group was 40 and that for the control group was 62. Thus, the observed changes could well have been due to regression to the mean. To make matters worse, it seems that the experimental children received much more enthusiastic intervention than the control children. Freeman18 quotes from a local newspaper story about the study in which teachers of the children in the experimental group extol the “wonderful” patterning training. Differing zeal of the teachers, combined with the regression problem already noted, is sufficient to account for the observed group differences, negating any support for DDPT.
Kershner admitted as much in a later recounting of this work,18 noting that “the extent to which differential teaching effects entered into the findings is unknown” and that the increases in IQ after DDPT should be viewed with “caution” because of the pretest group differences. Freeman20 notes that the conclusions in this account were more conservative than those of the above-mentioned booklet published by the State of Pennsylvania, which was widely distributed in promotions of the Institutes for the Achievement of Human Potential.
As implausibly optimistic claims generally do, the Doman-Delacato technique received much favorable attention in the popular press in the early 1960s.14 Concern for the public prompted the American Academy of Pediatrics (AAP) to issue a statement, in 1965, condemning DDPT. By the fall of 1968, this statement had been approved by 9 other professional organizations, including the American Academy of Neurology,4 the American Academy of Physical Medicine and Rehabilitation, the American Association on Mental Deficiency, and the National Association for Retarded Children. In addition to faulting the theory and empirical support for DDPT, the AAP was critical of the highpressure tactics used to promote it. The AAP concluded that these sales methods “appear to put parents in a position where they cannot refuse such treatment without calling into question their adequacy as … parents.” In addition, “it is asserted [by DDPT promoters] that if therapy is not carried out as rigidly prescribed, the child’s potential will be damaged.” The AAP was also dismayed that “restrictions are often placed upon age-appropriate activities of which the child is capable, such as walking or listening to music.” An updated version of the AAP’s censure appeared in Pediatrics in 198221.
Several more studies of DDPT appeared after the original AAP critique. A 1969 dissertation by Fredericks22 (cited in Foreman and Ward23) showed that a standard 9-week motor training program was more effective than a patterning program for improving motor control in children with Down’s syndrome.
In 1970, Cohen et al.24 criticized DDPT on theoretical grounds, attacking, for instance, its central tenet that children must pass through lower stages of a skill to become competent at higher stages(e.g., one must crawl before one can walk). These critics cited, to the contrary, a developmental study of Hopi children.25 Hopi infants are bound to a carrying board that prevents crawling for the first year of life. But a comparison of Hopi children raised in the traditional way with those allowed to crawl showed no differences between the 2 groups in walking ability. It appears that the neural systems that control walking develop in the absence of crawling.
In 1974 Newman et al.26 reported a study using reasonably high-functioning retarded subjects. Excluded were those with very low IQs and those who were deaf, blind, or otherwise severely impaired. Sixty-six subjects were divided into 3 equal groups. The experimental group received patterning and sensory training for 2 hours a day, 5 days a week, for 2.5 months, and then 7 days a week for another 3.5 months. A placebo group participated in general physical activity and interacted with volunteers and supervisors for the same amount of time. There was a nontreated control group that was merely tested at the same time as the 2 other groups. All groups were tested 4 times during the course of the 6-month experiment, with the first test being at the outset of the study. The authors claimed that the experimental group improved more than either control group on perceptual, motor, and linguistic measures, but not on the intelligence measures.
Zigler and Seitz27 found serious methodological and statistical flaws in the design of the Newman study. The patterning training that the experimental group received was administered by trainers from the Dallas Academy, an organization strongly committed to the validity of DDPT. The physical activity the placebo control group received was administered by psychology graduate students. There may also have been a failure to blind evaluators as to which group any given child had been in. The part of the evaluation requiring close physical proximity to the children was carried out by individuals trained by the Institutes for the Achievement of Human Potential, individuals with a large stake in proving patterning therapy valid.
There were also statistical problems. Many of the 46 dependent variables had high inter-correlations; i.e., they were not independent. Separate univariate analyses were carried out on each variable. Although an initial analysis of covariance was done, Zigler and Seitz note that this does not protect against “an inordinately high likelihood of falsely rejecting the null hypothesis in their analysis of all 46 measures.”27 In fact, while 46 different dependent measures were examined, many more than 46 tests were done. Zigler and Seitz obtained a copy of the unpublished report of the full study, of which the published version was a summary. They found in the unpublished version that 276 statistical tests had been performed, but only 23 (8.2%) reached significance at the .05 level. Since many of the individual measures were correlated among themselves, even this 8.2% figure overstates the case. Newman28 replied to Zigler and Seitz,27 but his lengthy reply concerns itself largely with interpretative issues and generally ignores the more damaging statistical criticisms.
In 1978 Sparrow and Zigler29 published their own study of patterning treatment effectiveness. An experimental group received DDPT for a year. A placebo group received a “matched motivational” treatment designed to increase self-esteem. This included various games and other activities individual children participated in with foster grandparents. There was also a nontreated control group. Over the year, all 3 groups showed some improvement on various measures, but “in no case did the pattern of change of the treatment group differ from that of its crucial comparison, the motivation group.”29(p137)
Although no peer-reviewed experimental research on patterning has been published since 1978, it has been mentioned in a few nonexperimental papers. In 1981, Zigler30 published a “plea to end the use of the patterning treatment for retarded children,” calling it “useless, expensive, and possibly harmful.”30(p389)
In 1979 the National Academy for Child Development (NACD) was founded by Robert Doman in Riverside, California. In 1983 Holm31 reviewed the program of the NACD (now headquartered in Ogden, Utah) and concluded that nothing had been added to the old, discredited patterning approach.
In a 1986 review of treatments for Down’s syndrome, Foreman and Ward23 mentioned patterning, but only to emphasize the lack of evidence for its effectiveness. They reiterated the condemnation of DDPT by the American Academy of Pediatrics. Foreman and Ward32 also surveyed pediatricians regarding their preferred therapies for Down’s syndrome. Of the 204 respondents, 97.5% said that they never recommended patterning and 92.2% reported that they advised parents against using it. Only 2.9% said that they were “sometimes in favor” of patterning and the remaining 4.9% were unaware of its existence.
Although no research on patterning therapy has been published in the medical or psychological literature for over 10 years, it has not disappeared. Discredited treatments rarely do. Instead, they are perpetuated by testimonials and uncritical media reports that fuel the marketing efforts of profit-driven promoters. True to fashion, Doman’s 1994 book1 was essentially a republication of an identical book from 1974.33 The only difference was that the word “Down’s” was added to the title, broadening the potential clientele. The Institutes for the Achievement of Human Potential continue to promote patterning therapy in their publications and courses, all listed on the IAHP Web site (www.iahp.org). One of their courses is titled “What to Do About Your Brain-Injured Child” and uses Doman’s book as the text.
- Doman G. What to Do about Your Brain-Injured Child or Your Brain-Damaged, Mentally Retarded, Mentally Deficient, Cerebral-Palsied, Spastic, Flaccid, Rigid, Epileptic, Autistic, Athetoid, Hyperactive, Down’s Child. Garden City Park, NY: Avery; 1994.
- Delacato C. Neurological Organization and Reading. Springfield, Ill: Thomas; 1966.
- Hines T. Pseudoscience and the Paranormal: A Critical Examination of the Evidence. Amherst, NY: Prometheus Books; 1988.
- American Academy of Neurology. The Doman-Delacato treatment of neurologically handicapped children. Neurology. 1968; 18: 1214–1216.
- Doman RJ, Spitz EB, Zucman E, Delacato CH, Doman G. Children with severe brain injuries. JAMA 1960; 174: 257–262.
- Delacato C. The Treatment and Prevention of Reading Problems. Springfield, Ill: Thomas; 1959.
- Delacato C. The Diagnosis and Treatment of Speech and Reading Problems. Springfield, Ill: Thomas; 1963.
- Glass GV, Robbins MP. A critique of experiments on the role of neurological organization in reading performance. Reading Research Quarterly. 1967; 3: 5–51.
- Robbins MP, Glass GV. The Doman-Delacato rationale: a critical analysis. In: Hellmuth J, ed. Educational Therapy. Vol. 2. Seattle, Wash: Special Child Publications; 1969: 321–377.
- Winer BJ, Brown DR, Michels KM. Statistical Principles in Experimental Design. 3rd ed. New York, NY: McGraw Hill; 1991.
- Miracle BF. The Linguistic Effects of NeuropsychoLogical Techniques in Treating a Selected Group of Retarded Readers [dissertation]. Laramie, Wyo: University of Wyoming; 1964.
- Keppel G. Design and Analysis. 2d ed. Englewood Cliffs, NJ: Prentice-Hall; 1982.
- Robbins MP. The Delacato interpretation of neurological organization. Reading Research Quarterly. 1966; 1: 57–78.
- Robbins MP. Test of the Doman-Delacato rationale with retarded readers. JAMA. 1967; 202: 389–393.
- Freeman RD. Controversy over “patterning” as a treatment for brain damage in children. JAMA. 1967; 202: 385–388.
- Robbins MP. Creeping, laterality, and reading. Academic Therapy Quarterly. 1966; 1: 200–206.
- Kershner JR. An Investigation of the Doman-Delacato Theory of Neuropsychology As It Applied to Trainable Mentally Retarded Children in Public Schools [master’s thesis]. Lewisburg, Pa: Bucknell University; 1967.
- Kershner JR. Doman-Delacato’s theory of neurological organization applied with retarded children. Exceptional Children. 1968; 34: 441–450.
- Freeman RD. Review of Kershner JR, An Investigation of the Doman-Delacato Theory of Neuropsychology As It Applies to Trainable Mentally Retarded Children in Public Schools. J Pediat. 1967; 71: 914–915.
- Freeman RD. Letter. Exceptional Children. 1968; 35: 94–95.
- American Academy of Pediatrics. The Doman-Delacato treatment of neurologically handicapped children. Pediatrics. 1982; 70: 810–812.
- Fredericks, H. A Comparison of the Doman-Delacato Method and Behavior Modification upon the Coordination of Mongoloids [dissertation]. Eugene, Ore: University of Oregon; 1969.
- Foreman PJ, Ward J. Treatment approaches in Down’s syndrome: A review. Aust N Z J Dev Disabilities. 1986; 12: 111–121.
- Cohen HJ, Birch HG, Taft LT. Some considerations for evaluating the Doman-Delacato “patterning” method. Pediatrics. 1970; 45: 302–314.
- Dennis W, Dennis MG. The effects of cradling practices upon the onset of walking in Hopi children. J Genet Psychol. 1940; 56: 77–86.
- Newman R, Roos P, McCaan BM, Menolascino FJ, Heal LW. Experimental evaluation of sensorimotor patterning used with mentally retarded children. Am J Ment Defic. 1974; 79: 372–384.
- Zigler E, Seitz V. On “An experimental evaluation of sensorimotor patterning”: a critique. Am J Ment Defic. 1975; 79: 483–492.
- Newman R. A reply to Zigler and Seitz. Am J Ment Defic. 1975; 79: 493–505.
- Sparrow S, Zigler E. Evaluation of a patterning treatment for retarded children. Pediatrics. 1978; 62: 137–150.
- Zigler E. A plea to end the use of patterning treatment for retarded children. Am J Orthopsychiatry. 1981; 51: 388–390.
- Holm VA. A western version of the Doman-Delacato treatment of patterning for developmental disabilities. West J Med. 1983; 139: 553–556.
- Foreman PJ, Ward J. A survey of pediatric management practices in Down’s syndrome. Aust Pediatr J. 1986; 22: 171–176.
- Doman G. What to Do about Your Brain-Injured Child or Your Brain-Damaged, Mentally Retarded, Mentally Deficient, Cerebral-Palsied, Spastic, Flaccid, Rigid, Epileptic, Autistic, Athetoid, Hyperactive Child. Garden City Park, NY: Avery; 1974.