The 8mm Microscope

Home Movies' Uses in the Medical Community and Beyond

Snowden Becker
Editor for Interactive Programs, J. Paul Getty Museum
Chair, AMIA Small Gauge and Amateur Film Interest Group
snowdenbecker@yahoo.com


Notes from my August 9 presentation at the NHF symposium follow. I would like to thank several people for providing invaluable assistance during my research and investigation of these ideas. Melinda Stone, and Louisa Trott shared materials and their own very helpful related research; Michael Friend's input on a very early draft of this research motivated the more thorough research review that is summarized here. Special thanks are due to Pamela Wintle for providing addition direction, and for being the first to personally encourage me to investigate the phenomenon of home movies as tools for scientific research and inquiry.

Francis Poole assisted me in finding a copy of "The Natural History of a Psychotic Illness in Childhood" (1958), which was screened after this talk to the symposium attendees. A very productive discussion of the film followed, and focused particularly on its success as medical propaganda, and its failure as objective documentation of an autistic child's case.


Despite the efforts of the archivists and scholars gathered here, the scholarly literature in the liberal arts on home movies remains scanty. The short list of books devoted to home movies and home movie making have concentrated on technology and social context, and have largely avoided engaging in deep content analysis of specific amateur film works or bodies of work. A search on indexed journals for “home movies” will retrieve far more articles about the home display of commercially produced movies than films made by amateurs for personal enjoyment or documentation of family activity.

In the social and medical sciences, however, there is a small vein to be mined of articles that are genuinely about home movies. An initial search in Medline or other scientific journal databases will yield a cluster of about a dozen studies that use the phrase “home movies” in the title. Citation chaining from this initial result set brings the total list to about three dozen articles, essays, and studies published since 1966 in which home movies figure as a tool for observation, diagnosis, and study—primarily of two major disorders, autism and schizophrenia.

The first piece in this body of literature was published in 1935. Dr. A. Gessell’s collection of notes in the Journal of Genetic Psychology, called “Cinemanalysis: A method of behavior study,” provides a detailed argument for the use of the still-new home movie technology for recording and making repeated observations of patient behaviors. Gessell’s idea of “cinemanalysis” seems to have had little immediate impact; the next three decades certainly did not witness any widespread use of the term, although the practice of recording patients under observation was adopted in due course by various branches of the health sciences for case review and teaching purposes. Gessell’s paper was cited in the 1966 article by Drs. Allen and Goodman, writing in the Archives of General Psychiatry in a piece called “Home movies in child psychodiagnostics: The unobserved observer.” In their article, Allen and Goodman present a case study of a girl diagnosed in childhood with autism, whose parents happened to have taken home movies of her in infancy. The authors of the paper describe film’s merits as a way to confirm a standing diagnosis, and go into technical detail that is lavish enough to suggest that at least one of the authors might be an amateur filmmaker himself. Allen and Goodman also note the approximate cost of a home filmmaker’s outfit of equipment, and note that it is accordingly a method of family record-taking that is most likely to be used by the upper middle class.

From this paper by Allen and Goodman, there is another longish gap in the literature on home movies, to 1973, when Dr. Henry Massie picks up the thread of “cinemanalysis” in his studies of autism. From 1973, studies incorporating home movies continue to appear fairly regularly up to the present (given scientific journal lead times, this represents work completed as late as about January 2003). My point, then, is to look at the bulk of these published papers, examine the role that home movies played in the studies, the social/historical and clinical context of their use for this purpose, and extrapolate from this some further ideas to which we as archivists should perhaps devote serious thought.

Studies on autism

I will start with a brief discussion of autism, with apologies for my lack of professional expertise in this area—I am not a medical expert, and until beginning this research, I knew little more about autistic disorders than anyone else who had seen “Rain Man” or read a book by Oliver Sacks. Dustin Hoffman’s portrayal of Raymond the autistic savant in Barry Levinson’s film is hardly representative of the disorder. In fact, no one case is representative of the population of sufferers; autism’s nature as a spectrum disorder, affecting a number of areas of development to differing degrees in each individual, makes each case unique.

As classified in the Diagnostic and Statistical Manual of the American Psychiatric Association (DSM-IV) autistic disorder is one of a family of five—the Pervasive Developmental Disorders—which also includes Asperger’s Syndrome, Rett’s Syndrome, Childhood Disintegrative Disorder, and Pervasive Developmental Disorder – Not Otherwise Specified. This last provides some hint of the difficulty that persists in describing and identifying this family of ailments, which are characterized generally by “severe and pervasive impairment in several areas of development,” which include social behaviors and communication skills and abilities. Autism in particular is associated with repetitive or “stereotyped” movements (rocking, obsessive touching of objects, hitting oneself when agitated), inappropriate or extreme responses to sounds or sights, or aversion to physical contact, identification or obsession with objects, lack of social smile or other normal emotional expressions, and several other symptoms, all of which may vary in severity. The diagnosis of autism is typically made on the strength of a composite score from a trained rater, who evaluates a host of characteristics, usually on a five-point range. It is a nuanced assessment, and subject to a certain degree of unreliability based on the rater and any input they may receive from parents or other caregivers.

The differentiation of autism from general psychosis was definitively made in 1943 by a Dr. Kanner, in an article called “Autistic disturbances of affective contact” that appeared in the medical journal Nervous Child. The title of the publication itself says much about the general state of medical and psychiatric knowledge at the time. A diagnosis of autism was relatively unhelpful to the parents and families of those affected for, I’d say, a good forty years or more—many patients with autism were simply institutionalized, with little hope of treatment or targeted therapies that might have an effect on their prognosis.

It is my sense, though, that the home movie studies carried out during the last three decades may have contributed significantly to improved understanding of this illness—and the shift from treating autism as one of many largely undifferentiated psychotic conditions, to its current description as one of a family of developmental disorders that have distinct signifiers, visible very early on in infancy or childhood, and in which early intervention and targeted therapy have a huge impact on prognosis.

One of the most obvious (and to my mind, most salutary) shifts in approaches to autism over the thirty-plus-year history of research involving home movies is the movement away from blaming poor parenting—and specifically, bad mothering—for autism. An old headline from the satirical newspaper The Onion, (RESEARCHERS LINK CRIB DEATH TO BAD PARENTS WHO DIDN’T LOVE THEIR CHILDREN ENOUGH) comes appallingly close to describing the conclusions of some of the early studies in this body of literature. Dr. Henry Massie’s work and case studies were among the first to use family home movies in the study of autism, and are frequently cited in the studies that followed, but they betray a persistent and probably inappropriate bias against what Massie judges to be flawed parenting behaviors. In his 1978 study, Massie seems to be going out of his way to find fault with the mothers in their filmed interactions with their children, concluding that “In 9 of 10 cases, the parents’ behavior with their child was inappropriate from the earliest weeks of life.” In some cases, both parents are targeted for criticism, but Massie notes that “the fathers are relatively absent,”—perhaps because they are typically behind the camera—and dismisses their behaviors, when visible, as being “generally parallel to the mothers’.” This assessment furthers an overall sense that the author holds the mother largely responsible for the disturbance of her child. Of particular note is his description of a Mrs. S: “[she] looks as if she were ensconced in armor. She is deanimated, recalling the wire-frame surrogate mothers which the Harlows (1965) used for the experimental production of maternal deprivation and subsequent psychosis in monkeys.” The audacity of such a comparison on the basis of only a few minutes’ worth of film seems horrifying now, and it also makes a curious contrast with the case study presented in Allen and Goodman’s 1966 paper. In the case of Allen and Goodman, the authors seem to be using the evidence of the films at hand to mitigate the mother’s self-blame, and they make a cautious, revised diagnosis of a condition with a more organic and neurological, less environmental and psychological, origin. “The parents seemed to pay more attention to her than they reported” is among the few judgments Allen and Goodman make of parental behavior on the film.

At the time, however, Massie’s value judgments and the weight he ascribed to mother-infant interaction seems to have been fairly typical. The early child development specialist Bruno Bettelheim was one of the first clinicians to focus on autism, and it is his mistaken theories of its origins that are perhaps most responsible for the long-standing tendency to blame a mother for her child’s autism. The 1958 documentary I will be showing later, The Natural History of a Psychotic Illness in Childhood, has some of the tone echoed in Massie’s later work—the narrator remarks at one point that “Mother’s determination to make normal contact with [her daughter] is apt to drive her further and further into her autism”—and, like Massie’s analysis, this work tends not to balance the scant evidence of parental misconduct or neglect to be found on the films with the more compelling fact that the parents did care for and attend to their children enough to create detailed records of their early childhood, in the form of photograph albums and diaries as well as home movies. David E. Simpson’s recent documentary “Refrigerator Mothers” (2002) deals directly with the devastating diagnosis of “frigid” mothering, and the life stories of several mothers who came to terms with such a diagnosis for themselves and their autistic children.

Based on the full range of the literature to which I referred, I think the mother-blame is largely or even wholly invalidated. This is not only due to later home-movie-based autism studies that objectively identify distinctive sensorimotor abnormalities in earliest infancy (which indicates an inborn/organic origin, rather than a primarily environmental one), but also to the flaws in the early studies themselves. Many of these papers do note that the fathers, being behind camera, are not present in enough footage to rate. More importantly, they acknowledge that the footage represents only a tiny fraction of the full lives of the patient and family, and parental behavior offscreen could simply never be verified. “Normal” parents may behave badly when not called on to smile for the camera, and parents of the autistic infants may display much more relaxed, normal behavior when not preoccupied by the concern that their child would behave strangely on film.

This latter conundrum—does the parent behave differently because the child is abnormal, or is the child abnormal because of the parent’s behavior?—is later addressed much more thoughtfully by Dr. Grace Baranek, in 1999. She points out that caregiver compensation techniques apparent in the home movies of children with autism may be unconscious reactions, but still provide important clues to the nature of the child’s earliest symptoms. Since some of these symptoms may seem to disappear as the child’s own development compensates for abnormalities, these clues of caregiver behavior can provide important guidelines for designing effective, personalized treatments and therapies.

Generally speaking, the findings in the autism home movie studies range from mere confirmation of the researchers’ hunches to startling and sometimes sad revelations about autism as a phenomenon and its treatment. For instance, Baranek’s 1999 study established that nine rating criteria, when combined in reviewing family films, could differentiate with 93.5 – 100% accuracy between children with autism, children with mental retardation, and normal children. Aside from the immediately evident merits of a system that can differentiate between autism and mental retardation, which are often confusingly similar in early childhood, one feels it would be particularly reassuring for a parent in doubt to receive a such a strong diagnosis of normality. A later study achieved results nearly as robust by rating four key actions or behaviors observed in videotapes of autistic and normal children’s first-birthday parties.

The saddest findings by far were in those studies that confirmed parents’ earliest feelings that something was wrong. In Maestro, et al., 1999, the results-blind clinicians who were evaluating the films tended to agree with parents’ estimates of the time at which symptoms were noticeable; however, it was found that on average, a full year elapsed before professional help was sought for the child. In Teitelbaum, et al., 1998, the authors looked exclusively at subtleties of movements and sensorimotor development in very young infants, and they isolated several striking abnormalities particular to the children who would later be diagnosed with autism. These authors' summation speaks volumes about the distance autism research has come from the mother-blame of the 1950s, and the long way it has yet to go: “The child’s mother is usually aware very early that something is wrong, but, because she is unable to specify something diagnostic, the pediatrician she consults often tends to reassure her that this is a minor problem that the child will grow out of…It is axiomatic that the earlier the therapy, the more effective it will be…[but] Time and time again…we have heard that the mother suspected…something was wrong…but that the pediatrician told her she need not worry. The pediatrician should be the earliest, not the last, to know that the child might be autistic.”

Studies on schizophrenia

The home movie studies on schizophrenia, though fewer in number—they account for only about a quarter of the total number of studies that incorporated home movies—presented similarly striking results. Schizophrenia as a disease has a much longer and better-documented history, one that dates back to the earliest analyses of patients with “dementia praecox,” as it was first called, in the last years of the nineteenth century. Cases of autism are most often diagnosed in early childhood, around 2-4 years; this is the time when children are beginning to socialize with same-age peers in the nursery-school setting, and their abnormalities are suddenly more evident. It is also the time when their developmental differences, especially undeveloped speech and motor functions, are conclusively discernible. Schizophrenia, on the other hand, manifests most commonly with a dramatic psychotic break in the late teens or early twenties, and has long been thought to have no discernible precursors in childhood or infancy.

Family factors are not a significant determiner for schizophrenia; 94% of schizophrenia patients have no family history of it. As a result, “high risk” studies of children of one or more schizophrenic patients are inherently flawed. The mental illness that crops up in a very small number of these high-risk subjects cannot necessarily be attributed to genetics any more than to the long-term effects of having been raised by a schizophrenic parent, especially one who may have been taking antipsychotic medications during pregnancy or nursing. Because of those family factors, it is especially useful to compare schizophrenics with their normal siblings in the shared family environment. It is here that home movies’ utility is especially evident, since most movies of multi-child families do at some points document two or more siblings interacting, often with peers or other relatives.

The home movie studies of children who were diagnosed with schizophrenia in early adulthood provided robust evidence that distinguishing behaviors and patterns of affect were visible, and rated at above-chance levels. In several studies where clinicians viewed edited segments of video that featured a pre-schizophrenic child with his or her normal siblings, the preschizophrenic child could be reliably singled out at as early as 5-7 years. Walker and Lewine, 1990, note that their study “represents the first demonstration that preschizophrenic subjects can be distinguished from sibling controls within the first 8 years of life by observing their behavior” (emphasis mine). The researchers were typically blind to the outcomes of all children featured in the footage, and their determinations, although accurate, were sometimes in direct defiance of parent, sibling, or patient memories of the period covered on tape.

Grimes and Walker went on to note in their 1994 study of childhood home movies of preschizophrenic children that “premorbid affective blunting is linked with poorer prognosis”—that is, preschizophrenic children who displayed more emotion in the films, whether that emotion was positive or negative, happy or sad, had a later average date of onset and better response to treatment than children who displayed little or no emotion before the onset of their illness. As they put it, “several reports indicate that adult schizophrenia patients who manifest depression have a better outcome than those who do not…negative affect in childhood may be both a sign of vulnerability and an indicator of more favorable prognosis.”

The ease with which researchers in these studies could correctly identify the child who would later have a psychotic break or serious mental illness is striking, and the use of home movies in these retrospective analyses of particular cases holds a lot of promise—especially now that the children of the video age are the focus of these inquiries. In a reply to a letter about her 1993 paper on the subject, in which the correspondents point out that abnormal or inappropriate verbalizations were not considered in her study as diagnostic factors in choosing a preschizophrenic child from the “lineup” on the tapes, Dr. Elaine Walker notes that the films were silent, and that “home videocameras came into widespread use in the mid-1980s, and they will provide data on language. The first cohort for whom childhood videos are available will reach adulthood in about 10 years [2003-2004]. Presuming that the origins of major mental illness have not been identified by that time, these videos may be invaluable sources of data on precursors.” That future is now, and we can probably expect to see a new crop of studies that do rely on those home videos—and data on language—within the coming years.

Critique of methods and conclusions

As exciting as the outcomes of the research incorporating home movies may be, there are serious flaws in these studies. First and foremost, the sample size is invariably very small—ranging from less than a dozen individuals to about 40, including controls. At these sizes, the sample is somewhat too small to provide reliable indicators that apply across a population, particularly in the case of autism or other spectrum disorders that can vary from patient to patient. And because the home movie footage is recorded without the intent of using it for clinical research, the sample material is difficult to normalize—a tricky issue, and one with which the social sciences frequently grapples. On the other hand, Osterling and Dawson’s 1994 study, which focused exclusively on first-birthday footage of autistic children, neatly accomplished this normalization by taking maximum advantage of the similarities in many children’s home-movie histories. First-birthday parties have been noted as one of the most common subjects of home movies. This study, though normalized in terms of footage and subject context, does not address the socio-economic narrowness of the samples; the expense of home movie equipment has rendered it available to a relatively small proportion of the population. Allen and Goodman pronounced their concern about this in 1966, and that unresolved worry recurs among other study authors up to the present. It remains to be seen whether digital video and increasing accessibility of this technology will ever do away with this concern entirely, but my sense is that it will not.

Other weaknesses of the studies result from the inherent technical and physical limitations of the medium used to document the subjects. The authors of these studies especially bemoan the lack of sound in early 16mm and 8mm films. Certainly, when children on film are being evaluated for “inappropriate responses to stimuli,” the absence of sound could account for dramatically incorrect assessments (imagine a Christmas scene in which a child begins crying for no apparent reason—although a parent behind the camera may have said “That’s it, no more presents!” in order to create a "more dramatic" subject for filming). The sheer quality of the image being reviewed is rarely, if at all, discussed—although in at least one case the authors' description of methodology made clear that second-, third-, or even fourth-generation video dubs of small-gauge films were being used for the study. Add to this the nature of the films as home movies—personal documents not made with the clinician’s eye or focus on particular events, behaviors, and interactions—and the fact that researchers must confine their studies to those subjects of whom there is sufficient usable film, and the already-small, self-selecting samples shrink further.

Finally, although some studies reported extremely reliable and robust results from a range of rating methods, others were reporting interrater reliability numbers considered “acceptable” at as low as .71. A firm diagnosis that is nonetheless a wrong diagnosis is far from desirable. Consider the fact that homosexuality was formerly listed in the DSM as a psychological disorder, and only removed in 1973. Pretending for a moment that that clinical description was still on the books, would we be willing to accept a nearly 30% variance in estimates or “diagnoses” of a condition that continues to carry real social stigma? Would we accept expert testimony about this sort of thing in court if experts agreed with one another only 71% of the time?

The openly gay pop singer-songwriter Rufus Wainwright, speaking about his sexual identity in Rolling Stone, said "You can be as liberal as you want, but once your son or daughter is actually gay, it’s hard for a lot of people. My father was fine with it. I mean, he claims to have known when I was four. One of my favorite toys as a kid was an apron that I would call my Put-It-On, and there’s this great home movie of me dancing around in my Put-It-On and my father in his lawn chair, freaking out. So he was prepared.” What if, instead of supporting a clinically debatable diagnosis of autism or schizophrenia, home movies were being used to support a practically undebatable diagnosis of homosexuality? Or used to show evidence of criminal tendencies, suicidal thoughts, spousal neglect, or thought crimes of any kind?

Another headline from The Onion seems apt here: AREA WOMAN HAS NO PHOTOGRAPHS WHERE SHE’S NOT HUGGING SOMEONE. Are we, individually or as a population, prepared to be living in a time when our personal information—including medical records and home movies—may be unsecure and subject to reexamination that leads to surprising, curious, or spurious conclusions? This is one point to which we should give long, careful thought.

Also bearing further consideration is the need to maintain and care for these personal documents. For parents of an emotionally disturbed or developmentally disabled child, records of a childhood that seems normal (or worse, once seemed normal, but is later found to have the seeds of a sad fate present all along) may be very painful. Reminders of that painful reality may be hidden, avoided, denied, or destroyed. Several authors of studies that used home movies note that the reviewing of those records may actually have a therapeutic effect for the family of the patient. Parents may learn that their parenting is unremarkable, and they are not to be blamed for their child’s illness; or they may feel vindicated when they find that a diagnosis they themselves had made in their hearts, though perhaps refuted by a professional consulted early on, was a correct one. Certainly the ground breaking results achieved in several of the home movie studies of autism and schizophrenia have contributed to a better understanding of those illnesses, and it would be a shame if such research ceased for lack of materials to study.

Lastly, if we by some lucky effort manage to profoundly mitigate or eradicate autism, schizophrenia, and other disorders entirely, might we not want to retain some record of what they once looked like? Or would films that capture this information perhaps be too dangerous in the wrong hands—again, as a foundation for spurious claims about the evidence of thought crimes, or unconscious clues about future behaviors? Or the means by which innocent people could be accused of having an unverifiable potential for disease? None of these questions have real answers, in my mind or elsewhere. They do, however, seem bound to come up as home movies and other private documents enter the public trust and are increasingly studied by audiences other than the one for which they were made. Let us think about these questions carefully, and make our own decisions before they're made for us.