Chapter 4

Results

Overview

This chapter provides general results from all of the data collected and combines results for certain areas of interest, such as autism causation by decade or the types of causations being investigated from one country to the next. The quantitative results are broken out by type, time period, and location. Other pertinent findings—such as the emphasis on childhood, direct conflicts in the literature, and specific peculiarities of the databases used—are incorporated at the end of this chapter.

Quantitative Results

Overall Statistics

Besides the four causal categories previously mentioned—environmental, genetic, neurological, and psychological—two other categories emerged during data collection and analysis: allergy/auto-immunological malfunctions and biological influences (i.e., a lack or abundance of some chemical present in autistic individuals). As mentioned under “Environmental Basis,” immune weaknesses often manifest themselves as environmental causations (i.e., in utero exposure to rubella; ingesting lead or adversely reacting to antibiotics during the toddler years). However, the abstracts are more likely to pair immune system weaknesses with allergies than with environmental factors. Similarly, the biological impacts become evident through brain chemistry unique to autism. Therefore, they appear to be neurological in nature.

Realistically, there could have been more than these six causations. However, the researcher limited the study to these two additions to the four apparent from the literature review since the lines of distinction among the sources blur as their numbers increase. When other causations’ investigations were identified, these were mapped to the six using correlations shown in Table 4.1.

Table 4.1. Characterizations Used for the Many Autism Causations in the Literature

Most of the causations labeled “environmental” are in fact related to complications surrounding the pregnancy. Purely environmental causations—such as a chemical factory or nuclear plant in the autistic individual’s home town—have started to be investigated only in the 1990s. However, 8 of the 1,543 causes clearly single out allergic insults (independent of environmental causation) as triggering autism. Ninety-six causes are specifically attributed to autoimmune system deficiencies. Thirty-eight of the 103 biological causes asserted are neurobiological (and so counted under both causes). Only 47 of the articles—such as those that specifically addressed body chemistry imbalances in the brain like an excess of the chemical serotonin—are recorded as solely biological causes. (The remaining 18 biological etiologies are paired with other causations.) Table 4.2 gives the total number of causes counted for all sources. (Appendix B, Figure 1, provides a pie chart of this same information to clarify visually the percentages and totals involved.) The 1,543 causes for autism appear in 1,259 abstracts and primary sources. The following sections further break out these totals by decade of publication, by the country in which each document was published, by source (at both the database/bibliography level and at the journal, monograph, conference proceeding, etc. level), and by author.

Table 4.2. Total Number and Percentage of Autism Causes

By Decade

In Appendix C, Table C-1 gives a year-by-year summary of the causes found and Figure 2 provides a decade-by-decade summary of the same information. Table C-1 makes obvious the year-span information for each of the causation. The second half of the table makes clear that once a potential cause surfaces in the literature, it never goes away. The first half of this table shows that a psychological causation was the first proposed, in 1943, with neurological investigations coming soon after. A possible environmental causation first appeared in 1949, while genetic or biological causations were not initially proposed until 1953. Allergy/immunological is the most recent research tack, first mentioned in 1965.

However, psychological causation research accounts for more than half of the pre-1970 writings, with neurological causations appearing about half as frequently, and the other four causations each in single-digit percentages. The issue of whether an improved understanding of genetics and brain mechanisms contributed to the increase in research into these causations and the declining percentage of psychological etiologies being investigated is addressed in the next chapter under “Bandwagon Appeal.” The following sections provide more specific narrative overviews of the causation data for each decade.

1943 to 1956

For layout convenience, this “decade” includes fourteen inclusive years. However, this does not skew the overall data for the ten-year decades that follow. Immediately after autism’s first appearance, follow-up articles were sporadic and sparse. In fact, the researcher found no articles on autism causation for 1945, 1946, and 1948, and only one source for both 1947 and 1949. Table 4.3 is a by-decade and by-etiology summary of all causes found.

Table 4.3. Instances of Each Autism Etiology by Decade of Publication

Two articles appeared in 1943, both attributing autism to psychological maladjustments. One of these is Kanner’s “Autistic Disturbances of Affective Contact,” the seminal article describing autism as “a new form of emotional disorder. . . present from birth.” Prior to 1943, autism received passing mention as a substrate of schizophrenia. From 1943 through 1956, 31 more autism causation articles were published. Twenty-two of these attributed autism to psychological disturbances. Four proposed a neurological causation. One author investigated a genetic cause, while two articles offered an environmentally induced causation and two others saw autism’s causation as biological.

1957 to 1966

The causal attribution trend from the previous decade continued, with the overall number of articles nearly doubling (from 32 to 60). More than half (31) of these were psychological causation-based, with neurological causes still second in number but with a more compelling 18 instances. The other four causes remained in single digits: genetic (5), environmental (2), biological (3), and allergy/immunological (1).

1967 to 1976

The number of causes given during this decade increased nearly five times from the previous ten years, with 292 autism causation possibilities proposed. Psychological causes still led the numbers, with 105, but neurological causation articles were closer behind at 102 total. This time, the other four causations were all in double-digits with genetic investigations still a distant third at 34 causes given. Environmental causes of autism were written up 21 times; biological causes 13 times. The theory that allergy/auto-immunological malfunctions trigger autism appeared 17 times.

1977 to 1986

The number of autism causation articles continued to grow with the increase in the overall literature on autism, though not at the same rapid pace as in earlier decades. The 440 causes given from 1977 through 1986 represent a little more than one and a half times the total from 1967 through 1976 (the 292 causations covered above). This time neurological causations account for 35.68% (157), while psychological causes dropped to 91 (or 20.68%). Genetic research became a closer third in this decade, with 89 causations so cited. The other causations each remained less than ten percent of the total causes under investigation, with 42 environmental, 33 biological, and 28 allergy/auto-immunological etiologies of autism appearing in published articles.

1987 to 1996

Like the first, the time span for coverage of this last decade is not as exact as for those in between. Perhaps not all of the autism causation articles published in 1996 are represented as yet, especially those from foreign language sources and conference compilations (an interesting thought considering that 1996 had the second highest total causes for a single year at 95). Seven hundred nineteen of the 1,543 causes found (nearly 47%) come from this last decade. Neurological causations alone number 270. With 176 genetic causes cited, this area of investigation moved ahead of the psychological causations published (116) for the first time. Only 47 environmental causations were offered, with the remaining two etiologies each accounting for about eight percent of the decade’s total as well: biological investigations numbered 52 while allergy/immune studies accounted for 58 of the 719 causes addressed.

By Country

Forty-four countries are represented in the autism causation literature appearing between 1943 and 1996. (See Table D-1, Appendix D, for the specific number of sources for all countries.) The United States accounts for nearly half, 621 of the 1,259 sources found (and 746 of the 1,543 causes). There are 121 sources listing the “United Kingdom” as the place of publication, with a different 121 specifying “England.” France is the country of origin for 71 of the sources used and Switzerland 47. Forty-two abstracts are of articles published in West Germany, plus two from journals published in the former East Germany and four recent articles from the reunified Germany. There are 32 articles from Japan, 24 from the Netherlands, 22 each from Canada and Italy, 20 from Denmark, 15 from Australia, and 13 from Spain. Eleven articles in this study come from journals published in the Union of Soviet Socialist Republics, not counting a more recent one from Russia. Norway has 8 sources in the sample. This thesis used six or less articles from each of the remaining 26 countries. (See Table D-2, Appendix D, for the complete by-country rank-order numbers.)

Variance of Cause by Country

Table D-3, Appendix D, provides a by-causation breakout for each country. Among the ten countries with twenty or more articles published, only three do not have all six causations represented. Italy has no articles attributing autism to biological causes. Canada and France have no allergy/auto-immune-based causations published. That most of the countries frequently publishing autism causation research findings cover the spectrum of causal attributes addresses the international scope of this lack of consensus.

Even for those countries with a moderate amount of articles—like Australia with 15, Spain with 13, or the USSR with 11—the span and the range of causes and years show no clear uniqueness. Other than an interesting aside for the three mentioned—that none of them published any allergy/immune articles—the variety of causes, their mix and span of years, and the percentage for each causation are reasonably consistent with the overall totals. (More discussion on this point appears in the section following.)

The two clearest examples of by-country variance are the slow growth of autism causation literature worldwide and the differences in the number of possible causations researched for each country. Of the thirty articles published before 1960, journals in the United States published twenty of them. Many of the countries with fewer than five sources used do not enter the debate until after 1970. Only 14 of the 44 countries in which autism causation articles were published have pre-1970 research appearing. Another statistic that can be gleaned from Table D-3’s tally by country and by cause is that in only 14 countries were allergy/immunological causation articles published; biological causations articles were published in 19 countries; articles on environmentally induced autism appear in 19 different countries; and 22 of the 44 countries account of all of the genetic causation literature published. The neurological causation is most widely internationally attributed, with 36 countries having articles on this etiology. Sources on psychological causations of autism have been published in 26 countries.

Variance of Cause Within Certain Countries

Even more interesting than the lack of international consensus regarding autism’s causation is a comparison of how the percentage for each cause within each country matches up with the overall percentages given in Table 4.2. Table 4.4 provides the numbers and percentages for the five countries publishing the most source information on autism’s etiology. The issue of how the “United Kingdom” is different from “England” and how “Germany” was counted is discussed in the next chapter along with other geographical issues. For this breakout of the information, the focus is on how these percentages compare to those given in Table 4.2. The relative order of the top three causations is almost the same in all cases while the rank order of the bottom three is more inconsistent. What is more consistent is a noticeable gap in numbers between the highest three and lowest three causes evident for most countries.

Table 4.4. Number and Percentage of Autism Causes for the Five Most Prolific Countries Compared to the Overall Numbers/Percentages

Some notable exceptions to the neurological, psychological, genetic order for those heavily referenced are: for the United States, genetic causes slightly outnumber psychological causes; for articles published in France, psychological causes emerged as the etiology most written about, with neurological causes distinctly second, and genetic causes a distant third (36, 29, and 13 appearances, respectively). For Switzerland, both genetic and allergy/auto-immunological causations placed ahead of psychological etiology research.

The least referenced causations from the overall list are environmental, biological, and allergy/auto-immunological research articles. This order of incidence is not consistent for many of the individual countries, though. For articles published in the United States, the allergy/auto-immune etiology appears 62 times versus 41 for biological causations. For England (and the United Kingdom), the bottom three are tightly grouped, with 21 biological cause articles, 16 environmental, and 15 allergy/auto-immune. France’s least referenced three appear in this same order with far fewer numbers: 4 biological, 1 environmental, and zero allergy/auto-immune references. Switzerland published 11 allergy/auto-immunological autism causation articles, with only 8 citations each for biological or psychological causes; the environmental etiology appears only 4 times. Only 3 biological, 2 allergy/auto-immune, and 1 environmental causation article were published in Germany. The more inconsistent order of the less frequent causations may be more attributable to their meager numbers than to any lack of causal consensus regarding autism. When a single occurrence represents almost two percent of the total, as is the case with Germany, the utility of these percentages is difficult to gauge. This is why only the top 5 among the 44 countries appearing in the literature are discussed. (Refer to Appendix D, Table D-3 for the countries not discussed.)

By Source

Almost 93 percent of the sources are abstracts of journal articles. Only 21 conference proceedings, 21 book chapters, 15 monographs, 27 letters, 4 editorials, and 2 web site-based articles were found. One thousand one hundred sixty-nine journal articles came from 397 journals. The background information on the databases (see the “Data Collection” section of the previous chapter) indicates that for most of the databased used, journals are what the database vendors concentrate on. Even so, a credible inference to make from the journal source percentage found is that a statistically significant portion of autism causation research is published in journal article format. Not as much can be inferred from the numbers per database because these numbers were influenced by the order in which the information was collected (see “By Database or Bibliography”).

By Journal

As stated, 1,169 abstracts came from 397 journals. The 27 letters, 4 editorials, and 11 of the 21 conference proceedings were also published in this same group of journals. (The Journal of Autism and Developmental Disorders has a category called “brief reports” and some of the other journals have categories called “research” or “reports,” but this study made no distinction between these categories and journal article types.) The most frequently referenced journals appear in Table E-1, Appendix E. It should be no surprise that journals account for more than 9 of 10 autism causation sources. The timely dissemination of formally published information favors journals, which is why bibliometric studies of scientific or medical information would favor journals. One issue is their currency of information, but other reasons journals are favored in bibliometric studies are easy to reckon: their articles are brief, more narrowly focused than monographs or even conference proceedings, and usually indexed individually.

The 53 journals having 5 or more articles account for 686 of the 1,211 articles, conference write ups, editorials, and letters counted in the 397 journals, or 56.65 percent. The United States-based Journal of Autism and Developmental Disorders and the Journal of Autism and Childhood Schizophrenia, its predecessor, account for 155 articles between them, or 12.31 percent of the total sources in this study. This finding has significance in the area of distribution and the next chapter contains a discussion of how Bradford’s Law of Scattering applies to the autism causation topic.

Variance of Causation by Journal

All 13 of the early Excerpta Medica references came from Section VIII, “Neurology and Psychiatry.” This is not too surprising considering this broad combination. (In recent years, the section is called “Neurology and Neurosurgery.”) What is surprising is that no autism causations came from the biological or other general medical sections that did exist prior to the creation of EMBASE. What is less surprising—especially considering that the articles were published between 1950 and 1969, with only three published in the 1960s—is that psychological causations dominate the Excerpta Medica selections. Nine of the 16 causes given in the 13 volumes used cite a psychological causation. However, all of the other causations except allergy/auto-immune are represented (neurological 3 times, biological twice, and one instance each of an environmental and a genetic causation).

The same spread of causations was evident across all of the journals with multiple instances of citations. Of course, one would expect a journal such as the Journal of Autism and Developmental Disorders and its precursor, the Journal of Autism and Childhood Schizophrenia, to reflect the same trends as the whole collection, but it does not (except that all causations are represented). The percentages for allergy/auto-immunological and environmental causations were more than twice as high (25 instances or 13.66%, contrasted with the 104 instances and 6.74% for the overall allergy/immunological causations and 28 environmental causations, or 15.30%, in contrast with the 114 environmental causations representing 7.39% of the overall totals). By contrast, the psychological causations attributed in the Journal of Autism and Childhood Schizophrenia and the Journal of Autism and Developmental Disorders combined make up less than half of the percentage that psychological causations have of the overall totals (21 for 11.48% of the former versus 366 overall psychological causations, 23.72% of the total—see Table 4.5 following). Part of this discrepancy can be explained by the timing of the Journal of Autism and Childhood Schizophrenia. The initial publication of the journal was 1971, nearly thirty years into the evolution of autism causation research.

Table 4.5. Number and Percentage of Autism Causations for Journals With the Highest Number of Autism Etiology Research Items Published

Although the rank order of the causations is different for this primary autism publication, all causes are represented, with biological causations accounting for a little more than seven percent and neurological etiologies claiming nearly a third of the total. One would expect the other top journals cited in this research to favor a specific causation, if only based on their titles: Developmental Medicine and Child Neurology (neurological); the Journal of the American Academy of Child Psychiatry, the Journal of the American Academy of Child and Adolescent Psychiatry, and the Journal of Childhood Psychology and Psychiatry and Allied Disciplines (psychological); and the American Journal of Medical Genetics (genetic). However, this association holds true only for the last mentioned, in which 31 of its 35 causations are indeed attributed to genetics. Twenty-two of the 52 causations (42.30%) covered in the Developmental Medicine and Child Neurology selections are neurological; however, 12 are genetic and 10 are attributed to psychological causations as well (23.08 and 19.23 percent, respectively). The Journal of the American Academy of Child Psychiatry and its successor the Journal of the American Academy of Child and Adolescent Psychiatry combine for a total of 48 causations. Surprisingly, most of these are genetic, not psychological (17 versus 15, or 35.41% to 31.25%). Eight of the causations mentioned are neurological (16.67%). Similarly, psychological causations are the second most cited etiology in the Journal of Childhood Psychology and Psychiatry and Allied Disciplines. This time, though, the most cited causation is not genetic, but neurological (15 of 34, or 44.12%). Psychological causations account for only 9 (26.47%) and the genetic causations are third in number (6, for 17.65%). Again, these peculiarities may be explained at least in part by the time span the journals in question cover. Developmental Medicine and Child Neurology began 1965. The Journal of the American Academy of Child Psychiatry ran from 1963 to 1986, before being renamed the Journal of the American Academy of Child and Adolescent Psychiatry (which is still in publication). The Journal of Childhood Psychology and Psychiatry and Allied Disciplines was first published in 1977. Another influencing factor is that three of these most cited journals are published in the United States, and two in England (Developmental Medicine and Child Neurology and the Journal of Childhood Psychology and Psychiatry and Allied Disciplines). Therefore, the Anglo-American perspective is over-represented. However, this Western influence is mitigated by the fact that all of the most cited journals publish works contributed from research conducted internationally.

Non-Journal Sources

Subtracting the articles, letters, and editorials which came from journals leaves only 48 non-journal sources. Ten of the 21 conference proceedings, 21 book chapters, 15 monographs, and 2 web site-published articles are the sum of nonjournal sources used. The two web-based sources are two of many Internet articles that are repostings of work previously published the traditional way. However, the researcher did not find these two indexed in any of the bibliographic databases searched. The 21 conference write-ups are mainly indexed by BIOSIS Previews (9) and PsycLIT/PsycINFO (4). All but 2 of the 21 book chapters are abstracted in PsycLIT, the only database that makes “book chapter” its own index category. Seven of the 15 monographs are indexed in the National Library of Medicine’s Current Catalog. BIOSIS and the Autism Resource page on the World Wide Web http://web.syr.edu/~jmwobus/autism/autism.faq each abstracted 3 monographs. The two remaining books are from the researcher’s own library.

By Database or Bibliography

Discussion of the numbers per database needs to be prefaced with an explanation of how the researcher treated duplicate appearances of usable articles. Nicholas and Ritchie claim that duplication from one database to the next is no more than 75 percent. In fact the OneSearch (of BIOSIS Previews, EMBASE, MEDLINE, and PsycINFO) run before the “remove duplicates” command was entered returned 2,905 articles; after the “rd” command, 2,086 articles remained (or 71.81%, meaning that nearly 28% of the articles were duplicates). However, the researcher also filtered out many false drops to finish with only 1,259 source items used. This final number is 60.35 percent of the 2,086 abstracts with duplicates removed. But the 1,259 also includes 157 sources not searchable automatically. For example, an April 1964 article appears in both the 1964 and 1965 runs of the Cumulated Index Medicus. Where the duplicate is a language translation or a conference report that later became an article or book chapter, these were not always filtered out automatically but manually removed (usually prior to entry in the project-specific database).

But even with those sources for which the duplicates could not be removed automatically, the order in which they were searched controlled the outcome to a degree. The researcher started with the OneSearch, then searched each of these database files’ print-based counterparts for the pre-digital years of bibliographic collection. Other sources—such as the National Library of Medicine’s Current Catalog, PsycLIT (to complement Psychological Abstracts and DIALOG’s PsycINFO for complete American Psychological Association source coverage), and genetic studies abstracted on the Web—were added next. For each source, more and more usable abstracts were repeats. The researcher noted these repeats, but did not count (or account for) them in the totals used from each source. Repeat occurrences of usable articles were viewed no differently from other false drops. In fact, articles that appeared in several databases were (of necessity) counted only the first time. The thought going in was that only the final tally matters, not individual source totals. However, since source totals are the subject of this section, this background may help explain Table F-1 (Appendix F).

By Author

Of the 1,259 articles in the study, 632 come from authors appearing only once. Eighty-seven authors have two articles published in this research. Thirty-four authors were published three times each. Twenty-four authors contributed four articles each to the total. Eleven authors have published five articles on autism causation. Five authors have six articles each. Only three authors have had seven articles published on the topic. Five researchers have had their work published eight times. Only two authors were published nine times. One has been published ten times (Sanua). Ritvo has contributed 13 articles to the autism causation debate. Rutter and Wing each have had 15 articles published. Gillberg is clearly the most prolific on the topic with 38 articles published.

An explanation of how the authors were counted is useful here. Especially in the more recent years of autism research, the articles have been published by multi-author teams rather than individual researchers/writers. Therefore, only the lead author is tracked for this tally. While this may miscalculate instances in which a contributing author later became a lead author, the composition of the research teams has generally remained constant enough over the years that correcting for this instance is more time consuming than its benefit warrants.

Part of the reason such an effort would be laborious involves a shortcoming in the researcher’s database design. However, another driver behind counting the research teams together was due to an unavoidable lack of information. The database designed for the researcher’s data collection allows for only one author(s) field instead of several. The result of this early miscalculation is that the author-counting process is automated only down to the individual article level, not individual coauthors. However, many of the databases and bibliographies do not even include the entire list of authors, but use et al. after listing one to three contributors. There can be no work-around for this if the source is cited only in such a database or bibliography (short of the even more time-consuming task of finding the source itself).

A second fault of the “author(s)” field is its not including first names, or at least first initials. The researcher suspected this might be troublesome for common surnames, but did not predict wherein the larger problem would manifest itself. Many of the coauthors are married couples. The DIALOG-based files and their print precursors use only first initials of authors, but PsycLIT (a SilverPlatter database) and Genetic Studies Abstracts supply authors’ full names. Of course, using the first initial would not have helped with Christopher and Carol Gillberg, but it would have with the Wings, Ritvos, and many other couples collaborating on autism causation research.

That stated, the method of author collection data is not so flawed as to be useless. More often than not, married researchers collaborate rather than publishing independent of each other. This and other factors make a case for emphasizing the lead author. The authors, after all, put themselves in a particular (rather than an alphabetical) order usually based on the percentage of their contribution to the published article or research project. The databases use author order as it appears in the source document.

Do the Same Authors Stay with the Same Causes?

Individual authors tend to be consistent in the causes they espouse, but only slightly more consistent than are individual countries. The range of autism causations being investigated, even by the same author, is surprising. Taking the five most published authors, Table 4.6 provides a breakout of the autism causations each has reported. Only Rutter and Wing show clear causal preferences. Perhaps Gillberg and Ritvo, like many of the authors with fewer articles published, are drawn toward different avenues of research depending upon the team with which they are co-investigating. For example, seven of Gillberg’s neurological causation articles are written alone. However, articles written with Hagburg and Wahlstrom tend toward genetic causality. The Ritvo articles written alone or with one partner tend to offer neurological causes, while most of those in which he is leading a six (or more) person team tend toward a genetic causation. This explanation does not account for the range given for Sanua’s articles, all ten of which are solo efforts. However, as one can judge from the 18 causes in 10 articles, Sanua tends to be multicausal even in each single write-up.

Table 4.6. Number of Articles/Causes for the Most Prolific Authors

Other Findings

Miscellaneous results worth noting include the emphasis on “childhood,” some direct conflicts in the literature, and a few database peculiarities of interest.

Childhood Autism

With the study of autism now more than fifty years old (not including investigations of the disability before it had this label), one would expect to see the emphasis on childhood gradually dissipate. However, despite the assertion from many sources that autism is a life-long disability for which there is no known cure, autism’s emphasis is and remains on childhood. The rationale for this could stem from the DSM-IV’s categorization of autism as a condition “Usually First Diagnosed in Infancy, Childhood, or Adolescence.” It should be no surprise, too, that those studying the cause of a condition are going to start at the beginning. Even so, the world of adult autism is under-represented in the literature. Only 9 of the 1,259 articles address adult autism. Twenty-one articles in this study begin with “Childhood Autism. . .,” “Childhood Psychosis. . .,” or “Childhood Schizophrenia. . ..” “Early Infantile Autism” is the subject of 29 articles and “Infantile Autism” appears in 119 titles. Even the titles of the journals in which the autism causation articles appear reflect this childhood emphasis: Developmental Medicine and Child Neurology, Journal of the American Academy of Child and Adolescent Psychiatry, Journal of Autism and Childhood Schizophrenia, Journal of Childhood Psychology and Psychiatry, Neuropsychiatrie de l’Enfance et de l’Adolescence, and Praxis Kinderpsychologie und Kinderpsychiatrie.

Other “autism” phrases encountered in the literature include “early childhood autism,” “autismus infantum,” and “infantile dementia.” Of course, some of the autism characterizing words were broad enough to incorporate adult autism, like Kanner Syndrome (for “classic” autism) and Asperger Syndrome (for certain high-functioning autism, especially that in which the individual has normal speech). Earlier sources called these primary and secondary autism, or low- and high-functioning autism. The appearance of autistic-like traits in a patient not diagnosed with autism (or diagnosed with something else, like cerebral palsy or mental retardation), is often termed “pseudoautism,” especially in the early years of research into autism. (However, all of the “autistic-like” and “pseudoautism” abstracts are listed as false drops for this research.)

It could be that parents are more likely to have their recently diagnosed children participate in studies, or that adults with autism have settled into the routine of it. Autism could be called something different—like schizophrenia—in adults. Or it could be simply that the formal diagnosis of “late onset autism” is much more rare than “childhood autism” or “infantile autism.”

Direct Conflicts in the Literature

In the previous chapter, the researcher addressed the handling of articles that were negating a cause without offering an alternative hypothesis, for example stating that autism is not genetic. What was not mentioned during that explanation is that at least 42 articles in this study directly oppose positions taken by many of the other articles. Four describe autism as not environmentally induced; one of these four offers no alternate proposal. One article states that autism may or may not be the same as childhood schizophrenia; two others claim that autism is the same as childhood schizophrenia. One article negates the biological causation. Four articles assert that autism is not neurological. Two sources specifically counter the cognitive causal theories and one other, the idea that autism is caused by hearing trouble in childhood. Two articles negate the immunological etiologies. At least five articles counter the psychological causations, including one that insists autism is not caused by what the parents may have done or not done.

Especially for the genetic, biological, and psychological causations, this opposition demonstrates a voice, if not a majority. Also noticeable in the abstracts is the surety of the neurological proponents matched against the more open-ended causation wording of other proponents (wording such as “autism is a neurologically-based developmental disability” versus “allergies play an important role” or “. . .indicates a genetic component”).

Those proposing a connection between autism and Fragile X Syndrome, Rett’s Syndrome, or tuberous sclerosis have even more opposition in the literature. Two claim there is no link between Fragile X Syndrome and autism. Ten claim that autism is not genetic while four of these assert that Fragile X is genetic. Sometimes when two such articles appear in close proximity, the juxtaposition introduces the appearance of chaos in autism etiology research. All that these direct conflicts clarify is how little is supported about autism’s causation.

Database Peculiarities

The strengths and weaknesses of the online databases became a noticeable side issue during data collection. MEDLINE has a dearth of abstracts for its pre-1975 articles. BIOSIS Previews has more conference papers indexed but fewer journal articles than either MEDLINE or EMBASE. Even considering that PsycLIT was run after the four DIALOG databases used, it yielded the fewest false drops (both clear false drops and unusable items in the researcher’s opinion). PsycLIT is also the only source to index chapters of books.

In the on-paper sources, the pointers to autism vary greatly. Biological Abstracts is the only one that added key words in context to the subject listing. This proved extremely useful in that the researcher could distinguish immediately those that deal with treatments from those that cover causation. Improving the Biological Abstracts search even more is the inclusion of “Autism” as its own index category starting with the 1960 annual. Prior to this, autism fell under various subjects: infancy and childhood, psychiatry, and psychopharmacology. Psychological Abstracts characterized autism similarly. Although “autism” can be found alone in the indexes at the back of each annual volume, autism abstracts generally appear under the “Childhood and Adolescence,” “Behavior Deviations,” or “Receptive and Perceptual Processes” subject headings. In the National Library of Medicine’s Current Catalog and Cumulated Index Medicus “autism” has been its own subject from its beginning (although 1964’s is the earliest one used for the former and 1944 for the latter in this research). It has already been noted that all of the Excerpta Medica articles came from Section VIII, “Neurology and Psychiatry.” However, the subsections used within this section include: “Psycho-pathology” and “Child Psychiatry.”

Also noticeable is the progression of euphemisms from the early tomes to the digital databases, and the sensitivity with which disability wording has evolved. The early abstracts (and by assumption, the articles they index) use wording that would be considered offensive, or at least harsh, in the 1990s: “autists,” “retardates,” and (for experimental control groups) “normals.” Control groups in the 1980s and 1990s began to be called by the more accurate term “neurologically typical.” “Handicapped” has all but given way to “disabled.” And “subnormal” (mercifully enough) has become “exceptional,” “special,” and more commonly “individual with autism” in recent years.

The database files that give full journal titles are a boon (EMBASE and especially BIOSIS Previews, which makes “full journal title” its own field). However, for certain year-spans in their indexing, all but PsycINFO omit punctuation from their titles. For titles that should include colons or question marks, this could change the meaning, such as in the following examples: Gillberg’s “Autism and Anorexia Nervosa: Related Conditions?” and Goodman’s “Infantile Autism: A Syndrome of Multiple Primary Deficits?”.

BIOSIS Previews and MEDLINE use a full list of authors, while EMBASE, PsycINFO, and PsycLIT tend toward et al. after listing the first one to three authors. Besides having the fewest multiple authors listed, EMBASE has the most misspellings in its database records. Also, fewer letters and editorials are cited than are used in the other DIALOG database files included.

Summary of Results

This chapter includes general results from all of the data collected and combines results for certain areas of interest, such as autism causation by decade or the types of causations being investigated from one country to the next. The quantitative results are broken out by type, time period, and location. Other pertinent findings—such as the emphasis on childhood, the appearance of direct conflicts in the literature, and specific database peculiarities—close this chapter.

The final chapter applies bibliometric analyses to the results described above. Bradford’s Law of Scattering (concerning the number of journals that publish articles on autism causation), Goffman’s Epidemic Theory (concerning how autism etiology article ideas are generated), and Lotka’s Law (concerning author productivity on the specific topic of autism causation) will be explored. Other implications drawn from the data collected will also be discussed.

This page is maintained by
Jeff and Cathy Romanczuk, lukate@charter.net