Comparative Analysis of False Domestic Violence Claims

face of a prisoner

Different communicative fields support and propagate criminal suspicion of men in different ways. Define as S the claim “domestic violence is the leading cause of injury to women” or the many closely related formal variants of that claim. S has been prevalent throughout public discourse. S supports highly damaging criminal suspicion of men. S has no rational basis in good reason. S instances by year, the use of frightening comparators with S, and the share of debunking S instances differ across communicative fields (news sources, web texts, law journals, Congressional documents, and judicial opinions). Those S differences provide a quantitative index of different communicative fields’ reasoning about criminal acts. They indicate the extent that different communicative fields support criminal suspicion of men.

Simple textual searches through general-purpose databases of different communicative fields provide a dataset of S instances for quantitative analysis. The dataset of S instances is called the Domestic Violence Claims Dataset (DVCD). The S statistics below, summarized from the DVCD, should be consider lower bounds on the actual number of S instances. Collecting instances of S through general searches of textual corpora not pre-selected for S helps to ensure the representativeness of S statistics relative to communicative fields. Because the scope of searching is difficult to measure and standardize across communicative fields, count differences between fields are less informative than time-series trends and S instance shares within fields.

"Domestic violence is leading cause of injury to women":
S Instances in Different Communicative Fields, 1985-2010

yearscommunicative fields
Source: Domestic Violence Claims Dataset.
newslaw journalsCongresscourts
peak-years comparison
2006-10 as share of 1993-727%28%1%20%

While S concerns criminal acts, law journals were relatively slow to produce S. The Surgeon General reportedly claimed S in 1984. The earliest directly documented S instance occurred in a news article in 1987. Two S instances occurred in the Congressional Record in 1988. From 1986 to 1990, news articles and Congressional records contain at least 44 and 13 instances of S, respectively. From 1986 to 1990, only one instance of S occurred in a law journal (in 1989) and no instances occurred in judicial opinions. That judicial opinions lagged in producing S isn’t surprising: judicial opinions producing S have typically done so through citations to Congressional records and law journal articles. Law journals are commonly thought to be a leading forum for thinking about issues of law. Law journals’ relative late performance in producing S might be attributed either to stronger reason or less attention to current issues of legal practice. Law journals’ subsequent frequent production of S favors the latter explanation.

S instances per year peaked about 1994 in news articles, law journals, and Congressional documents. The 1994 peak is particularly sharp for news articles. S instances in news articles in 1994 (126 instances found) were about double the level in 1993 and 1995. In 1994, O.J. Simpson was charged with murdering Ronald Goldman and Nicole Brown Simpson, O.J. Simpson‘s wife. That sensational event probably stimulated news articles asserting S.

S instances in U.S. judicial opinions peaked at 6 in 1997. Judicial opinions are a communicative field with different characteristics from news articles. Judicial opinions typically don’t respond to high-profile current events such as the O.J. Simpson case. No S instance in a judicial opinion was essential to the argument in that judicial opinion. Judicial opinions produced S in the context of documenting, in a supporting way, authoritative claims about domestic violence. The reduction in S instances in judicial opinions after 2000 is difficult to interpret. Fewer S instances may reflect fewer cases that presented judges with the opportunity to assert S.

After peaking in the mid-1990s, the frequency of S instances dropped much less in news articles and law reviews than in Congressional documents. Congress has regular opportunities to produce S through, for example, statements associated with yearly proclamations of Domestic Violence Awareness Month. Nonetheless, only one instance of S occurred in Congressional documents from 2006 to 2010. Many instances of S occurred in news articles and law journals from 2006-2010. Instances of S in that period amounted to about 27% of instances during the general peak from 1993 to 1997. Politicization is commonly thought to work against good reason. But with respect to reducing production of S, deliberation in Congress outperformed news articles and law journals. Commercial, market-driven news arguably favors sensational crime stories and encourages punitive criminal justice policy.^ However, status-driven competition for publications in law journals has responded no better than market-driven news in moving away from sensational, false claims of S.

"Domestic violence is leading cause of injury to women":
S Instances Associated with Frightening Comparators & Debunking, 1996-2005

fieldtotal instancesinstances with frightening comparatorsshare with frightening comparatorsinstances with debunking claimsshare with debunking claims
Source: Domestic Violence Claims Dataset.
law journals2047336%31%

Debunking claims against S have been relatively rare and not influential. The first documented debunking instance was in a news article in 1994. That news article included a clear statement disavowing S from a co-author of the Surgeon General’s 1992 statement on domestic violence. Nonetheless, many instances of S occurred in news articles, web texts, law journals, and judicial opinions after that debunking article was published.

News articles published between 1996 and 2005 contain 14 debunking S instances. Debunking S instances account for about 11% of total S incidents in news articles. Debunking instances occurred in news stories with titles such as “Ending Bias in Domestic Assault Law,” “Bashing Boys Is, Like, Not Ok,” “In Abuse, Men Are Victims, Too,” and “Reports of domestic violence exaggerated.” All the identified authors of the S debunking articles are women. Just two women wrote 10 of the 14 debunking articles. Stories about the need to help battered women or about praiseworthy efforts to do so are the typical context in newspapers for affirmative incidents of S. While men did not explicitly author any of the debunking articles, men wrote about 40% of news articles affirmatively stating S. Compared to news article affirmatively stating S, news articles debunking S were much less common and only authored by women.

Among the 204 instances of S in law journal articles published between 1996 and 2005, only three of those instances were associated with debunking claims. One instance put the debunking claims in a footnote to an affirmative S claim. It sourced the debunking to one woman and did so in a way tending to undermine her credibility.^ Another instance put both the affirmative claim and the debunking claim in a footnote. The debunking claim was a newspaper reference and the statement that S was based on a sample of black women. The third debunking instance declared:

While no one would argue that the statistics show that domestic violence is one of the leading causes of injury among women, there is polarization in the field regarding what constitutes domestic violence, its causes, and appropriate treatment to end that violence. ^

That statement debunks S, but, as the dataset of S instances shows, it does so with a grossly false claim. Many instances of S have been produced across communicative fields. At least with respect to S, law journals show surprisingly little exercise of good reason.

The predominance of women among authors of law-journal articles asserting S is probably best understand in terms of the different professional interests of female and male law professors. In 2000, roughly 30% of law professors were female.^ About 70% of law journal articles asserting S had female authors. Hence female law professors were more likely to assert S than were male law professors. Nonetheless, as the paucity of debunking claims against S in law journals indicates, neither female nor male law professors have seriously challenged S. Asserting S in law journals appears to be largely a convention of writing about domestic violence or family law more generally. Female law professors predominately in the highly under-appreciated field of family law.

Among searched communicative fields, the web had the highest frequency and quality of documents with debunking S instances. Structured web searches for S in early 2006 found 11% of S instances to be associated with debunking claims. The second-highest result on the Google search result page for S instances containing the phrase “leading cause” was a web page that provided detailed injury statistics, by cause, based on data from the 1996 National Hospital Ambulatory Medical Care Survey Data File. That web page noted that the relevant data file could be freely downloaded from the website of the National Center for Health Statistics. Tables of leading causes of injury, based on the National Electronic Injury Surveillance System, are currently easily available online via WISQARS, an online system of the National Center for Injury Prevention and Control. Debunking claims, and authoritative injury data, exist on the web along with many affirmative instances of S. Affirmative S instances exist on web sites of colleges, major non-profit organizations, and organizations involved in the administration of criminal justice. Low-cost, readily accessible communication across the web hasn’t resolved with good reason important, contrasting factual assertions.

Frightening comparators used with S provide an additional indicator of non-reasoning. Among communicative fields searched, the web had the highest share of S instances associated with frightening comparators. Law journals had the second-highest share. Congressional documents, in contrast, had the lowest share of S instances associated with frightening comparators. The difference is large: the share of S instances associated with frightening comparators is 43% and 36% for the web and law journals, respectively, while Congressional S instances had a frightening-comparator share of only 7%. At least with respect to S, freedom to think and write without respect to political and economic interests (the web and law journals) didn’t produce better reasoning than a high-pressure political field (Congressional documents).

Comparative analysis of false domestic violence claims in different communicative fields suggests that criminal suspicion of men won’t be mitigated through the growth of more free, more information-rich fields of public communication. Such communication, however, is not the only possible direction of development of democratic communication. New fields and new forms of communication offer the best hope for better democratic governance of criminal justice and punishment.

Rationalizing Domestic Violence As the Leading Cause of Injury to Women

face of a prisoner

High-quality, highly credible, U.S. nationally representative injury surveys indicate that domestic violence is far from the leading cause of injury to women ages 15 to 44, or to women generally. Nonetheless, public discourse has successfully rationalized the many variants of the claim that domestic violence is the leading cause of injury to women. Common standards of public reasoning — assessments of interests, credibility, representativeness, etc. — have not been applied to highly damaging claims directly related to criminal acts. Instead, these false claims have been rationalized in a variety of ways.

False claims about domestic violence have been rationalized by claiming under-reporting. An influential claim about domestic violence being the leading cause of injury to women concerned injuries prompting visits to hospital emergency departments. Extensive public investment exists in emergency services for treating injuries and in nationally representative surveys for measuring the incidence of such injuries. Domestic violence being the leading cause of women’s injury-related visits to hospital emergency departments would imply massive under-reporting of such injuries. Under common standards of reasoning, such a claim would require weighty evidence. No weighty evidence supports massive under-reporting of domestic-violence injury-related visits to hospital emergency departments.

False claims about domestic violence have been rationalized by re-description. For example, the leading cause of injury to women ages 15 to 44, as measured by U.S. national injury surveys, is unintentional falls. One might claim that a large share of what injury surveys record as women’s injuries from accidental falls are actually domestic violence against women: Humpty Dumpty was pushed. However, a cause category “Injury Undetermined Whether Accidentally or Purposely Inflicted” was explicitly included in one such survey. That cause category amounted to only 1% and 5% as many injury visits as accidental falls in 1992 and 2001, respectively. In 2001, women and men ages 15 to 44 suffered 1.2 million and 1.4 million emergency-department visits, respectively, from accidental falls. Under common standards of reasoning, weighty evidence would be necessary to claim that a large share of domestic violence injuries to women are misreported as accidental falls and that the true incidence of unintentional falls is much less for women than for men. No such weighty evidence has supported rationalizations that a large share of women’s, but not men’s, accidental falls resulted from them being pushed, i.e. domestic violence.

False claims about domestic violence have been rationalized through pettifoggery. With sufficiently fine subdivisions of all other causes of injury, any given cause of injury can be made the leading cause of injury. A more subtle variation is to call domestic violence “a leading cause of injury” or “one of the leading causes of injury” and to define implicitly the number of leading causes to be sufficiently large so as to include domestic violence. Alternatively, one can define injury in such a way that good evidence isn’t available to measure alternate causes of injury. Shifting the definition of injures to a broader class of injuries than serious physical injuries increases the number of injuries from domestic violence and from all other causes. That can create desired confusion about comparative numbers. Broader categories of injuries are more difficult to measure and thus favor arbitrary claims. Shifting the definition of injury to encompass psychological harm helps to create measurement difficulties and to advance arbitrary claims. For example, here’s a sweeping, practically unfalsifiable claim: “Domestic violence affects more people than any other health-care problem in the United States.”^ For the most part, such pettifoggery has not been necessary. Without good reason, public discourse has widely supported the claim that domestic violence is the leading cause of injury to women.

Good reason is unlikely to overcome the claim that domestic violence is the leading cause of injury to women. From 1988 to 1997, instances of that claim (including variants ) have been asserted as true at least 780 times in U.S. Congressional records, federal and state judicial opinions, law journal articles, and newspaper articles. Nonetheless, in 1998, a psychotherapist working domestic-violence cases wrote in the law journal Arizona Attorney:

While no one would argue that the statistics show that domestic violence is one of the leading causes of injury among women, there is polarization in the field regarding what constitutes domestic violence, its causes, and appropriate treatment to end that violence.^

That statement’s introductory clause denies a large public record of just such claims. The article making that statement addressed domestic violence in terms of power and control. The claim that domestic violence is the leading cause of injury to women has remained prevalent in public discourse. That’s an astonishing example of the power and control of thinly veiled anti-male gender animus.

Formal Rational Basis for False Claims about Domestic Violence

face of a prisoner

In U.S. v. Morrison (2000), a dissenting Supreme Court opinion reproduced false claims about domestic violence against women in finding a rational basis for provisions of the Violence Against Women Act. That Supreme Court opinion, which four justices supported, observed that Congress passed the Violence Against Women Act after “extended hearings,” “vast amounts of testimony and documentary evidence,” “years of the most thorough legislative consideration,” and the assembly of a “mountain of data.” The opinion explicitly presented twenty Congressional statistical findings. Among that list of statistical findings were the following two:

  1. “Violence is the leading cause of injuries to women ages 15 to 44….” S.Rep. No. 103-138, p. 38 (1993) (citing Surgeon General Antonia Novello, From the Surgeon General, U.S. Public Health Services, 267 JAMA 3132 (1992)).
  2. “[B]attering ‘is the single largest cause of injury to women in the United States.’ ” S.Rep. No. 101-545, at 37 (quoting Van Hightower & McManus, Limits of State Constitutional Guarantees: Lessons from Efforts to Implement Domestic Violence Policies, 49 Pub. Admin. Rev. 269 (May/June 1989).^

Both of these statistical findings are variants of the prevalent false claim that domestic violence is the leading cause of injury to women. While the Supreme Court opinion cited these findings from two different Congressional reports, they also occur within a single Congressional report.

The justices reviewed to some extant these findings. The Senate report cited for the first finding above incorrectly attributed the finding to Surgeon General Antonio Novello. The Supreme Court opinion correctly substituted Antonia for Antonio. The opinion’s presentation of the Senate’s finding also suggests some substantial evaluation. The Senate’s finding was actually reported by the Senate thus:

Violence is the leading cause of injuries to women ages 15 to 44, more common than automobile accidents, muggings, and cancer deaths combined.3 {footnote 3} Surgeon General Antonio Novello, “From the Surgeon General, U.S. Public Health Services,” “Journal of the American Medical Association,” vol. 267, No. 23, at 3132 (June 17, 1992).^

In presenting that Senate finding, the Supreme Court opinion omitted the frightening comparators “more common than automobile accidents, muggings, and cancer deaths combined.” Those frightening comparators underscore the irrationality of the reported finding. The Supreme Court opinion’s editing and selective quotation of the Senate’s finding suggests some consideration of that finding.

The Supreme Court opinion did not provide additional, simple review of the above two Congressional findings. The frightening comparators in the Senate’s finding (“more common than automobile accidents, muggings, and cancer deaths combined’) don’t actually occur in the Senate’s cited text for the finding. Moreover, the finding as reported in the Senate report, even without the frightening comparators, isn’t a fair representation of the statement in the Senate report’s cited text. Thus the Supreme Court opinion produced, as formal evidence of a rational basis for a law, a specific Senate finding that is formally false at the first level of reference.

The findings that Supreme Court opinion cited are formally false in more sophisticated ways. The source (“Van Hightower & McManus”) cited for the second of those findings actually attributed that finding to the Surgeon General:

In fact, the United States Surgeon General identifies battering as the single largest cause of injury to women in the United States.^

MacManus and Van Hightower’s article provides no citation to support attributing this claim to the Surgeon General. The Senate report eliminated that problem by moving down the chain of claim reproduction. In addition, the first finding refers to violence, while the second refers to battering. Battering (repeated, severe abuse) is a subset of domestic violence, which itself is a subset of violence. The first finding addresses the population “women ages 15 to 44,” while the second finding addresses women in general. Causes of injury, not surprisingly, vary substantially between young women and elderly women. The two findings are inconsistent at their surface levels. An exercise of reason could easily reveal that the these Congressional findings are formally factual, but also formally inconsistent, formally ill-supported, and substantially false. Such an exercise of reason apparently was not part of the Supreme Court opinion’s marshaling of those findings to indicate a rational basis for the Violence Against Women Act.

In reproducing such findings, the Supreme Court opinion did not analyze how gender affects motivation, but rather expressed an impression of gender motivation. The opinion noted:

It is true that these data {findings} relate to the effects of violence against women generally, while the civil rights remedy limits its scope to “crimes of violence motivated by gender” – presumably a somewhat narrower subset of acts. See 42 U.S.C. § 13981(b). But the meaning of “motivated by gender” has not been elucidated by lower courts, much less by this one, so the degree to which the findings rely on acts not redressable by the civil rights remedy is unclear. As will appear, however, much of the data seems to indicate behavior with just such motivation.^

In specific cases of violence, motivation is typically described with case-specific facts. Impressions of motivation for highly aggregated classes of behavior reflect more general feeling. Consider, for comparison, a true finding: as measured by visits to hospital emergency departments, men suffer 45% more injuries from violence than do women. In aggregate, a man is more likely to attack a man than to attack a woman. Is violence against men motivated by gender? The current impression in public discourse seems to be that violence against men is not motivated by gender. That factually incongruous impression seems best understood, like the presentation of findings on domestic violence, as motivated by gender.

Prevalent reproduction of false claims about domestic violence against women indicates men’s vulnerability to criminal suspicion in public discourse. In U.S. federal and state courts, 21 judicial opinions have explicitly reproduced claims like the claim that domestic violence is the leading cause of injury to women. False claims about domestic violence against women have gained support at pinnacles of public reasoning such as the U.S. Congress and the U.S. Supreme Court. Such institutions have supported false claims about domestic violence against women through highly formal processes of reasoning such as extensive public hearings and lengthy court proceedings. These results, which are examples of a more general pattern of judicial reasoning about domestic violence, are weighty evidence for gender motivation in public reasoning. These results should inform public consideration of sex disparity among persons incarcerated.