Methodological Issues — Archival Data versus Scientific Research by Tom James, Esq.
Methodological Issues — Archival Data versus Scientific Research by Tom James, Esq.© 2003 Tom James
Reprinted with permission from Domestic Violence: The 12 Things You Aren't Supposed To Know
| EJF Home | Join the EJF | Comments? | Get EJF newsletter | Newsletters |
| DV Home | Abstract | Contents | Tables | Index | Bibliography |
| Chapter 4 — Domestic Violence Statistics |
| Next — Battered women and battered statistics |
| Back — NIMH studies of domestic violence |
The Conflict Tactics Scale (“CTS”)
Criticisms of the Conflict Tactics Scale (“CTS”)
Ex-spouses and former intimates
Archival data means information that can be obtained by reviewing records kept by organizations and governmental entities in the ordinary course of their business or operations. With reference to domestic violence, archival data includes such things as law enforcement records; court records; emergency room records; and information obtained from battered women's shelters. The reliability of archival data is a function of the sample size, the randomness of the sample selected, the manner in which the inquiry is formulated and communicated, the absence of factors that may affect reporting, and the manner in which information from records is communicated and interpreted.
Survey data is information obtained by soliciting information from a sample population. Like archival data, the reliability of survey data depends on the size and randomness of the sample selected; it also depends on the kinds of questions asked and the manner in which they are asked.
It has been pointed out that the utility of archival data is inherently more limited than that of survey data because the samples from which archival data are obtained are never randomly selected. Data from battered women's shelters, for example, by definition exclude one-half of the population (males) right from the start. Similarly, data from medical facilities will mostly involve people who have not been socialized to suppress their pain, to “take it like a man.” Law enforcement records will exclude data about male victims for this reason, too — as well as for a host of other reasons.
These kinds of considerations are what led Lenore Walker, in her treatise on battered women, to state that archival data cannot be considered a legitimate database from which to make generalizations about domestic violence. 1 For that, survey information obtained from random samples of the population at large is needed.
The Conflict Tactics Scale is a survey-based measure of domestic violence that was originally developed by Murray Straus over twenty years ago. It has been revised several times since in the intervening years since then. The measure assesses how parents react in conflicts with their children, as well as how people react in conflicts with their partners.
The CTS has been criticized on the grounds that it only measures the incidence and severity of violent acts, not the causes or motivations behind the acts. 2 Criticisms of this kind usually cite as examples the woman who “playfully” punches her husband or “jokingly” kicks her husband in the groin without really meaning any harm by it, the premise being that not all violence is equally motivated by malice. This shortcoming is kind of a strange criticism to hear from the mouths of those who would urge reliance on archival data instead, since archival data are also simple tabulations of acts (and alleged acts) that do not usually entail the making of fine distinctions along motivational or contextual lines.
It is true that surveys employing the CTS are not always accompanied by psychoanalysis of the survey participants to ascertain the true motivations for and significance of each of their violent acts. Still, there are reasons why it can be expected that CTS-based survey data would be more reliable than archival data in this respect. Most of the items that the survey classifies as “severe” are by their very nature not likely to be things that are done in jest. It would be hard to imagine how stabbing someone with a knife, or shooting a gun at someone, or slamming someone against a wall, or beating someone up, could ever be regarded as “playful” acts. By distinguishing these kinds of acts from “mild” acts, such as pushing and shoving, the CTS gives us a better approximation concerning malicious intention than a record of a conviction for battery does. “Battery” is legally defined in such a way that it can include everything from tickling a person with a feather to beating a person senseless with a baseball bat. Thus, simply by building in measures of severity, the CTS provides a much more accurate estimate of malicious intention than archival data does.
It should also be noted that although the CTS does not delve deeply into the underlying causes of violence, other studies have been specifically designed for this purpose. For example, Fiebert and Gonzales 3 have specifically studied the motivations underlying female violence as measured by the CTS scales. They found that female violence is motivated by the same kinds of things that motivate male violence — jealousy, control, etc. Other researchers have reached the same conclusion. 4
Critics of the CTS sometimes contend that it doesn't provide accurate information about domestic violence because it only tells about the number of acts of domestic violence, without distinguishing which of these acts were committed in self-defense and which were committed with malicious intent.
This criticism could just as easily be made about the archival data, too. Law enforcement records also only tell about the number of reports of violence, not motives. Indeed, one of the primary criticisms of the data showing that the rate of arrests of women for domestic violence is catching up to the rate of arrests of men is the fact that arrest data do not necessarily reflect which of the two parties was “the primary aggressor” and which one was acting in self-defense. This is why women's advocates have secured the passage of legislation in some states (and are lobbying for similar legislation in other states) requiring police to make detailed findings concerning which party to a mutually combative domestic dispute is “the primary aggressor” before they may take the female combatant into custody.
Data obtained from emergency room records are also subject to this criticism. Findings about domestic violence that are based on emergency room records tally the number of patients who respond positively to what is essentially a survey-type of question: “Have you ever been a victim of domestic violence?” 5 This is the same type of question that the CTS asks. Neither question can tell us for certain whether the subject was acting in self-defense or with malice aforethought.
Information from battered women's shelters also does not provide reliable information about self-defense. Battered women's shelter services are usually provided only to women, not men, so one-half of the population is selectively excluded right from the start. In addition, battered women's shelters can only provide information about one small subcategory of abuse victims — women who are victims of life-threatening abuse. Even the archival data shows that the vast majority of domestic abuse is not life-threatening. Women seeking shelter in a battered women's shelter comprise the one category of victims that is most likely either to have been victims of lethal (or threatened lethal) violence — and therefore most likely to have been acting in self-defense if they themselves were violent, or to have a motive to claim self-defense in order to qualify for services. In short, it is impossible to draw any meaningful conclusions about domestic violence from information obtained from battered women's shelters.
Actually, the best source of information that is available, so far, about whether or not particular acts of domestic violence were perpetrated in self-defense comes from CTS-based research. It is probably true that the earliest versions of the CTS did not contain adequate measures for determining whether violence was perpetrated in self-defense or not. The CTS was subsequently revised, however, in ways that provide more information about the extent to which self-defense may have been involved.
Contrary to popular belief, a considerable amount of research has been conducted into the motives for domestic violence. The research goes well beyond simply tabulating the number of acts of violence committed by members of each gender. Research studies that have tested and disproved the hypothesis that female violence is always in self-defense are set out in Chapter 3 and include, among others: Billingham and Sack, supra; Bland and Orne, supra; Bookwala, supra; DeMaris, supra; Fiebert and Gonzales, supra; Gonzales, supra; Mann, supra; Mason and Blankenship, supra; McCarthy, supra; McKeowan, supra; Morse, supra; O'Keefe, Brockopp and Chew, supra; Sommer (1994), supra; Sorenson and Telles, supra; Stets and Pirog-Good, supra; Stets and Straus, supra; Straus (1996), supra; Tyree and Malone, supra; and the National Youth Survey, supra. Some researchers, such as Sommer (1994), supra, have found that in domestic violence cases, it is actually men who are most likely to be acting in self-defense when they are violent toward a member of the opposite sex.
One of the most oft-repeated criticisms of the CTS is that it only tallies up the number of acts, without distinguishing between “occasional” perpetrators and “repetitive” perpetrators. This may have been true of the earliest version of the CTS, but for several years now the CTS has contained a variable that specifically tests for chronicity. 6
Critics of the CTS sometimes claim that it only tabulates the number of violent acts, without telling anything about how severe or injurious the acts were. Again, this criticism could also be made about law enforcement and court records. Law enforcement and court records are rife with records of men who were reported for “abuse” for things like slamming a door in anger, or taking a woman's keys from her hand to prevent her from driving drunk. 7 In one documented case, an overzealous child protection caseworker who observed a man in a restaurant open his arms to invite a hug from his six-year-old daughter pursued charges of child sexual abuse against him on this basis alone. 8
The claim that the CTS does not distinguish between the number of acts committed and their relative severity is completely unfounded. The CTS, both as originally constructed and as subsequently revised, has always measured both the incidence and the severity of violence. In addition, Dr. Straus has always taken these kinds of criticisms to heart and has made revisions to the CTS from time to time, in an effort to improve its efficiency for distinguishing between mild and severe forms of violence. For example, after receiving criticisms of the CTS on the grounds that “throwing something” doesn't distinguish between throwing something relatively harmless (such as a pillow) and throwing something that could be more dangerous when thrown (such as a glass or a plate), Dr. Straus revised this particular item, so that it would instead ask respondents to tell whether they or their partner had “thrown something that could hurt.” (emphasis added.) The revised CTS includes even more specific measures of the kinds of injuries that have resulted from the specific acts of violence about which the survey asks. 9
Research studies that have specifically tested for and confirmed the severity of female violence include, among others: Arias, Samios and O'Leary, supra; Brinherhoff and Lupri, supra; George, supra (1999); Greenfeld, supra; Headey, Scott and De Vaus, supra; Kalmuss, supra; Langhinrichsen-Rohling and Vivian (1994), supra; Magdol, et. al., supra; McKeowan, supra; McLeod, supra; McNeely and Mann, supra; Russell and Hulson, supra; Rollins and Oheneba-Sakyi, supra; Rouse, supra; Stets and Henderson, supra; Steinmetz (1981), supra; Straus (1996), supra; Straus and Gelles (1986), supra; Straus and Gelles (1990), supra; and Vivian and Langhinrichsen-Rohling (1996), supra. The National Crime Victimization Survey (1994) also reached a similar conclusion.
Several writers have criticized the CTS on the grounds that it supposedly fails to measure the incidence and frequency of choking. 10 Perhaps these critics are confusing the CTS with the All-Alberta study. 11 The All-Alberta study did not include choking. Then again, the All-Alberta study also did not include wife-against-husband assaults as a possible category of violence, either! 12 The CTS included choking, as have most studies that have concluded that male and female rates of violence are similar (including the National Family Violence Survey), long before the 1996 expansion of the CTS. 13 The CTS still includes choking. Appropriately, the CTS not only includes choking, but has always classified it as a “severe” form of violence. 14
Given that scratching is a behavior more often engaged in by women than men, its supposed omission from the CTS seems an odd complaint to hear from those whose ultimate mission is to discredit CTS-based findings that women are as violent as men. Yet, the same people who criticize the CTS because it supposedly omits choking also criticize it for failing to include scratching. 15 According to Vera Mouradian, Ph.D., of the National Violence Against Women Prevention Research Center at Wellesley College for Women, however, studies using the CTS and other instruments have, in fact, measured scratching 16 as a specific form of domestic violence.
Moreover, even if “scratching” isn't specifically mentioned in a CTS-based survey, questions 11 and 12 of the CTS2 specifically ask whether the respondent ever received or inflicted even a small cut because of a fight with an intimate partner. 17 A “small cut” strikes me as being pretty much the same thing as a “scratch.”
The original version of the CTS covered only physical, verbal and emotional abuse. Consequently, it was sharply criticized for failing to specifically include measures of rape and sexual assault. 18
Given that the CTS was originally constructed thirty years ago, the omission of this category of violence is understandable. Historically, rape was a crime that was defined in such a way that it was only possible for a man to commit it, and the only victims of this crime could be female. This is because one of the essential elements was penetration of the female sexual organ by the male sexual organ. Therefore, any survey that included “rape” as a form of domestic abuse would necessarily find that 100% of the perpetrators are male and 100% of the victims are female — not because men are more violent than women, but because male sexual organs are constructed in such a way that they can be used to penetrate, while female sexual organs are designed in such a way that they cannot be used to penetrate. We don't need a survey to tell us that a penis is more likely to penetrate something than a vagina is. Such data would tell us only about anti-male gender biases in the definitions of crimes, not about the relative propensities of each gender for violence.
Between the time the CTS was originally constructed and its subsequent revision, the legal definition of rape was broadened and new categories of sex crimes were created that did not necessarily require the penetration of a vagina by a penis. After the concept of sexual assault was thus expanded, the CTS was revised to include measures of sexual assault, defining the term in this broader sense. 19
For the past several years, the CTS has included measures of sexual assault and has classified such an assault as a “severe” form of violence. This is as it should be. The CTS should classify an assault on one's sexual organs as a more “severe” form of violence than things like slapping, pushing, shoving, or calling someone a name. There should be no question but that inserting something (whether it's a penis, a physical object, or something else) into another person's vagina or anus against that person's will should be classified as a “severe” form of violence. By the same token, kicking, hitting, yanking or cutting a man's penis or testicles should also be considered “severe” forms of sexual abuse. For reasons having only to do with a deeply-rooted tradition of gender bias, however, the latter acts are normally classified — both in the law and in social science research — as “physical,” not “sexual” assaults. 20 As a result, the CTS classifies assaults on male sexual organs as sometimes “mild” and sometimes “severe,” whereas all assaults on (or “unwanted” contacts with) female sexual organs are automatically classified as “severe.” In this respect, the CTS, as it is currently constructed, is actually gender-biased in favor of women, not men.
The current version of the CTS measures behavior that is intended to cause one's partner to engage in unwanted sexual activity and covers three ranges of coercion: force, threats of force and verbal insistence. Thus, the revised CTS adopts the broad feminist definition of sexual coercion, under which “verbal insistence” resulting in sexual activity is considered a form of sexual assault, even though neither force, threat nor even an implied threat is employed. Even worse, the revised CTS specifically adopts the same gender-biased definition of sexual abuse that is employed by the Violence Against Women Grants Office, a definition that is couched in terms of violative penetrations of a vagina, anus or mouth, while omitting violent assaults upon penises and testicles. 21 The incorporation of the broad feminist definition of sexual abuse, while at the same time defining terms in a manner that arbitrarily excludes a significant category of male victims, can be expected to make it appear that the ratio of female-to-male victims of sexual abuse is far higher than is actually the case.
Despite the obvious anti-male gender bias in the current version of the CTS, it is important to keep the actual incidence of sexual abuse in perspective. Non-CTS studies on the incidence of sexual abuse show that it occurs with far less frequency than other forms of physical abuse. The most reliable survey data show that less than 1/10th of 1 percent of women are victims of rape, attempted rape, or some other form of sexual assault. If archival data is used, the figure would be 1/20th of 1 percent. 22 If, as the survey data, the National Organization for Women, the Violence Against Women Office and hundreds of other organizations claim, more than 3 million women are victims of domestic abuse each year — or if between 5% and 100% of all women are victims of domestic abuse 23 — yet less than 1/10th of 1% of all women are victims of sexual abuse, then mathematically it necessarily follows that sexual abuse must be involved in only an extremely tiny fraction of all domestic abuse. 24
None of this is meant to trivialize the experience of rape, attempted rape or sexual assault. These are serious offenses, and those who commit them should be severely punished. Nor is it my intention to imply that domestic violence never involves sexual abuse. It most certainly can. My point is only that it does not happen anywhere near as frequently as other forms of domestic abuse do. As was noted in Chapter 1, the data show that the higher rates at which men are victimized relative to women do not change that much even when rape and sexual assaults are factored into the computations.
Critics have also complained that the CTS does not provide information about murder. 25 Of course the CTS does not provide information about murder. It's a survey. Dead people tell no tales; they also don't respond to surveys. This is why murder is the one type of violence that has to be measured by archival data, not survey data.
Having said that, it nevertheless should be noted that although the CTS does not specifically measure homicides, it does measure the propensity to use lethal forms of violence (shooting, stabbing and choking, for example.)
Several writers have criticized the CTS for its omission of questions about stalking. 26 In defense of the CTS, it has been pointed out that some of the behaviors classified as “emotional abuse” may qualify as “stalking” behaviors as well. 27 Be that as it may, however, there are sound reasons for excluding data about stalking from the CTS.
The CTS is designed to measure violence between people who are currently involved in a relationship with each other. Inasmuch as stalking normally does not occur while a couple is together, and inasmuch as it is not necessarily an act that involves intimate partners, 28 it doesn't make much sense to include it in a study that is only designed to analyze the tactics that people who are involved in an intimate relationship use to deal with conflicts.
There are also significant definitional problems. In some states, simply existing in a public place with knowledge that one's presence there may be “unwanted” or “annoying” to another person who is also existing either in the same public place or in close proximity to the same public place where the first person is existing can make one guilty of “stalking.” Some states include the exercise of First Amendment rights within the definition of “stalking” if done in a residential neighborhood with knowledge that at least one of the residents disagrees with one's point of view (“residential picketing.”) Harmon, Rosner and Owens also adopt a broad definition of stalking, according to which anytime one commits two acts of any kind that are “unwanted” and “annoying” to another person, one is guilty of “stalking.” 29 Under this definition, a man who, seeing an attractive woman sitting at a bar, sits next to her (1 st act) and then asks her, “Do you come here often?” (2 nd act) would be guilty of “stalking” if the woman doesn't happen to find him attractive, but would not be guilty of “stalking” if the woman did happen to find him attractive. Perhaps I'm missing something, but this just doesn't comport with my intuitive sense of what is meant by “domestic violence.”
Following a person is also a form of stalking. Yet, is “following a person” really a conflict tactic? Perhaps in some cases it is. In many cases, however, people have entirely different reasons for following other people. Often, it is misguided courtship behavior (e.g., the women who have stalked members of pop groups such as the Beatles, or television personalities like David Letterman, and other wealthy and/or famous men.) In some cases, a person may follow another person to obtain information, not to harass, annoy or intimidate the person. Sometimes people follow other people simply for reasons of courtesy, as where one person follows another into a store or a theater rather than pushing ahead of him.
Moreover, since stalking is defined in terms of the subjective reaction to one's presence rather than the inherent dangerousness of the defendant's act, and since not all states require any proof of a specific intent to harass or frighten, it is essentially a strict liability, subjective thought-crime. In many states, being alive is the only overt physical act that is essential for a conviction. 30 A person may have a genuine interest in taking part in a public event, even though he knows that there's a chance he might run into someone who would prefer not to see him there. In some states, he may be guilty of “stalking” if he nevertheless decides to attend the event.
For all of these reasons, this particular kind of “violence” would be exceedingly difficult to measure by means of a survey. “How many times have you existed?” would not provide a researcher with much useful information. Yet, “How many times has somebody felt fear as a result of seeing you or knowing that you were close by?” is not the kind of question for which we can expect to get a reliable answer, since most human beings are not telepathic.
Actually, some efforts have already been made to measure the prevalence of “stalking.” What survey data there is on stalking wouldn't necessarily change the results of studies that are based on the CTS, though. For example, although some studies have shown a somewhat higher rate of stalking victimizations for women, surveys conducted by organizations that are not committed to the principle of gender-based discrimination, 31 such as the Statistics Profiles published by Statistics Canada each year, show that men and women are about equally likely to be victims of stalking. 32
As was discussed at some length in Chapter 11, the problem of over-reporting and under-reporting is a significant shortcoming of archival data, and the actual incidence of female violence against males could well be 9 times greater than what the archival data appears to indicate. Critics of the CTS have raised the same kind of complaint about findings based on CTS surveys, arguing that some people might deliberately lie, or might experience memory problems, when they take the survey, which could result in either over-reporting or under-reporting on the part of survey participants. 33
Unfortunately, the problems of lying and mental incapacity are inherent in all data. As was discussed at length in Chapters 10 and 11, there are many reasons why people can be expected to have strong motivations to lie to police, prosecutors, medical personnel and shelter operators. It is also a well-documented fact that most acts of violence never even get reported to these archival sources of data in the first place.
To some extent, survey data, although intended to fill in the gaps left by archival data (by attempting to control for strong motivators like the desire to avoid going to jail, etc.) may themselves be vulnerable to participant dishonesty and faulty memory. Generally, there should be fewer motivations to lie, however, where responses are confidential and anonymous and cannot be used either for or against the respondent in any way, as compared to data gathered in contexts where strong motivations to lie are present, such as where one's responses may determine whether one will be taken to jail or not, or whether one will get custody of her children or not, or whether one will be able to obtain benefits like housing, support and money.
Nevertheless, it certainly is conceivable that even in the context of a confidential, anonymous survey, a male survey participant might be resistant to the idea of acknowledging that he has abused a woman. On the other hand, it is equally conceivable that a male participant may have difficulty admitting that he “let himself get beat up by a woman” or that he's “afraid” of a woman. Since men are trained to “blow off” the injuries they receive, a woman might be expected to have a better recollection of the times when she either victimized someone or was victimized by someone. At the same time, since women are trained to be on the lookout for signs of even potential violence, a woman might perceive a wider range of conduct as having violent connotations than may be objectively reasonable under the circumstances. 34 The combination of these factors could result in a tendency among some women to over-report, and a tendency among some men to under-report.
In light of these facts, the manner in which the CTS is administered is very important. If it is only administered to one gender, but not the other, the potential for either over-reporting or under-reporting is greater. For this reason, participation by both genders can be expected to heighten the accuracy of survey data, since the male tendency to under-report would then be counter-balanced by the female tendency to over-report.
CTS survey data attains the greatest degree of reliability when members of both genders participate and the male and female responses corroborate one another. In this respect, it can be said that some CTS-based studies may be more reliable than others, as there are higher levels of corroboration in some CTS-based studies than there are in others.
Some critics of the CTS have attempted to create the impression that there is never any agreement between male and female responses to CTS survey questions. This is not true. A few of the earlier CTS-based studies did have relatively low levels of agreement, but in subsequent CTS-based studies utilizing more refined administration methods, high rates of agreement were noted. For example, the Dunedin study found an extremely high rate of agreement between the male and female responses. The Dunedin study also happened to be one of the studies that found that females perpetrate significantly more violence — both mild and severe — against their intimate partners than males do. 35
As a matter of fact, levels of corroborative agreement, and therefore reliability, seem to be higher among the studies that have found significantly higher rates of perpetration by females against males than the studies that merely show “equal” or “nearly equal” rates for both genders. 36
Straus and Gelles have noted that even when male responses to the CTS survey were completely disbelieved, and reliance was placed entirely on female versions of events, the results do not change much. Even female-only responses to the CTS confirm that women are at least as violent toward men as men are toward women. 37
There is one respect in which the CTS can legitimately be criticized for lack of reliability: the attempt to measure how an individual experiences a victimization; that is to say, the psychological consequences of victimization. By necessity, one can only obtain survey information about an individual's internal experience of pain by asking him or her about it. Yet, this is almost guaranteed to result in a false response. For example, asking a perpetrator, “Did it hurt your partner when you hit him/her?” invites the perpetrator to alleviate personal guilt feelings by minimizing the consequences of his/her actions: “Sure I punched him, but he didn't cry about it, so it must not have hurt him much.” Since the perpetrator cannot be expected to provide reliable information about the victim's internal experience, it would be necessary to rely entirely on the victim's responses. Yet, the victim may have his/her own motives for either exaggerating or concealing the extent of the pain he/she suffered. A victim who has learned through experience that her expression of pain is likely to be rewarded with sympathy or other benefits, or who has learned that guilt feelings can be exploited for personal gain, can be expected to over-report the extent of negative consequences she has experienced. 38 On the other hand, a victim who has learned through experience that his expression of pain is likely to be punished, and/or that his suppression of pain is likely to be rewarded, can be expected to under¬report the extent of negative consequences he experiences. Without any way to cross-corroborate responses, there is no way to control for over-reporting on the part of one gender and under-reporting on the part of the other gender.
For this reason, CTS survey data about the nature and extent of the psychological harm suffered as a result of violence, as distinguished from the actual incidence of violent acts, is inherently unreliable. Based on years of sex role-typing, the psychological consequences of domestic abuse for women will generally be over-reported, while the psychological consequences of domestic abuse for men will generally be under-reported.
Ironically, the same people who attack the methodological soundness of CTS-based research in connection with measuring the incidence of abuse nevertheless cite it as reliable authority for the proposition that domestic abuse supposedly has “more negative outcomes” for women than it does for men. 39 Yet, the part of the CTS that supposedly supports the latter conclusion is the only aspect of CTS-based research that employs inherently unsound methodology.
The CTS only measures violence between intimates. It does not measure violence between people who are no longer involved in an intimate relationship with each other. This is understandable, given that the purpose of the CTS is to study the tactics people use to deal with conflicts with their intimate partners. By definition, a person someone used to know but with whom one is not involved in an intimate relationship is not a person with whom one is involved in an intimate relationship.
On the other hand, it is true that by limiting its focus to intimate relationships, the CTS misses some of the violence that occurs in our society. The archival data does suggest that a woman is more likely to be a victim of lethal violence after the termination of a relationship with a man than during one. A married woman does appear to be less likely to experience lethal forms of violence than a separated or divorced woman is. On the other hand, men may also experience a heightened risk of domestic violence during or after the breakup of a relationship. We cannot really know for sure, because what research has been done on the subject of domestic violence against men always seems to exclude inquiries into the subject of violence by ex-girlfriends.
The same reasons that justify the exclusion of information about violence by former girlfriends from U.S. government-backed studies of violence by intimates can also be cited as good reasons for the exclusion of information about violence by former intimates in the CTS. The best justification for excluding such information is that its inclusion would tend to undermine the reliability of CTS research. The very factors that lead to the breakup of a relationship typically cause a considerable amount of ill-will toward one's former partners, ill-will that can be expected to cloud both one's memory of past events and one's interpretation of current events. For example, there are documented cases in which former boyfriends and ex-husbands have had protection orders issued against them, and in some cases have been arrested, for things as simple as sending birthday cards to their children, simply because the former girlfriend or ex-wife interpreted the act as an effort to “harass” the mother by reminding her of the father's continued existence.
It's not just the women, either. In my practice, as well as among people I have known on a personal level, I have seen some pretty bizarre twists of logic and psychological self-protective devices being used by both genders when it comes to their perceptions about their former partners. Specifically, I have observed a strong tendency among members of both genders to demonize their former partners, to interpret everything the other person said or did in the past as an “abusive” act, and to interpret everything the other person now says or does as a “vindictive” act.
The termination of an intimate relationship is a deeply painful experience. The grief one experiences during a divorce has been compared to the grief one experiences upon the death of a family member. Some scholars have even gone so far as to say that a divorce is the most stressful event a person can experience in life, outranking all other setbacks life can throw one's way — being fired or demoted, changing jobs, moving, the death of a loved one, etc. It almost always involves a profound blow to one's self-esteem; and the risk of mental health problems is extremely high during and immediately following the breakup of a relationship. It is not a time when a person can be expected to give accurate, unbiased responses to survey questions about the relationship.
Including survey questions about former intimates would also compromise the reliability of CTS surveys in another way. As we have seen, CTS surveys depend for their reliability on cross-corroboration. Relying exclusively on the responses from only one-half of the couple is not likely to generate very reliable information. It is only when we are able to analyze the rate of agreement between the male and female halves that we can be certain of a high level of reliability in the results obtained from the survey. Data about former partners, which by definition must be based entirely on only one of the parties' recollection and interpretation of events, are not only inherently suspect but are not even capable of being tested for reliability through analysis of the level of agreement among responses given.
Those who would fault surveys on the grounds of insufficient measures of corroborative reliability should not fault the same surveys on the grounds that they do fail to include inherently unreliable, unverifiable data.
Despite the many valiant efforts to discredit the CTS, it is recognized by responsible scholars as the single best instrument available for accurately measuring the actual incidence, frequency and severity of violence among intimate partners, as distinguished from the societal response to domestic violence. 40 It also bears pointing out that findings based on CTS survey data have been replicated and corroborated by studies that do not use the CTS. 41
There is a considerable amount of hypocrisy among critics of the CTS. Specifically, the same people who say it is “unreliable” and therefore should not be believed when it tells us that women commit as much or more violence than men, often rely on the same CTS studies for their figures about the rate of domestic violence against women! The statement that “2,000,000 wives are abused by their husbands every year” was derived from Murray Straus' 1977 CTS-based survey, as was the factoid that “a woman is abused by a man every 15 seconds.” It is only when someone points out that the same study from which these figures were derived also shows that a man is abused by a woman every 14 seconds and at least 2,000,000 husbands are abused by their wives every year, that people then turn around and say the CTS is not reliable. You can't have it both ways.
Because it is well-documented that most domestic violence never gets recorded in the archival data, the CTS is the only reliable instrument that has been devised, to date, for determining the true extent of domestic violence. Archival data must still be relied upon for the collection of data concerning fatalities, but with respect to domestic violence in general it really only tells about the societal response to domestic violence, not its actual prevalence. 42 For the latter purpose, data obtained from surveys, such as the CTS, are essential.
The common practice among educators, women's organizations, and members of the media, to cite survey data when talking about violence against women, but to cite archival data when talking about violence against men, is deceptive, manipulative and, frankly, irresponsible. It perpetuates myths and stereotypes about men and women that serve only to impede inquiry into and the development of knowledge about the phenomenon of domestic violence. Those who impede knowledge impede solutions.
1 Walker, The Battered Woman, supra, p. xiii.
2 See, e.g., Dobash, R.E. and R. Dobash, Violence Against Wives: A Case Against Patriarchy, New York: Free Press (1979)
3 Fiebert and Gonzales, supra.
4 See, e.g., McCarthy, supra; Gonzales (1997), supra; and Stets and Pirog-Good, supra.
5 Contrary to popular belief, statistics reporting that “20%” (or whatever the statistic du jour happens to be at the time) of women seen in emergency rooms are victims of domestic abuse are not about the number of women who are at the emergency room to receive treatment for injuries received as a result of domestic violence. Rather, they reflect the number of women who answered “yes” to the survey question, “Do you believe you have experienced domestic abuse at some time in your life?” A woman who believes she has been verbally or emotionally abused could answer such a question affirmatively even though the abuse had nothing whatsoever to do with her reason for being in the emergency room on the particular day in question. This is why there is a huge discrepancy between the findings based on archival data concerning women who have actually sought treatment for injuries caused by domestic violence (a comparatively small number), and findings based on survey questions about domestic violence in general that are asked of women who are in the emergency room for any reason. Unfortunately, many unscrupulous politicians, women's organizations and journalists have used data from surveys conducted at emergency rooms as a basis for making statements like, “Emergency room records show that more women are treated in emergency rooms for domestic violence than for all other kinds of injuries combined.” The campaign to characterize the male gender as a bunch of vicious misogynists has been so successful that such misrepresentations are rarely, if ever, questioned, even when they are patently absurd.
6 Straus, et. al. (1996), supra. Of course, if a man accused of abuse were to try to defend himself on the ground that it wasn't “severe” or that it only happens “occasionally” rather than “chronically,” he would most likely be chastised for trying to minimize the evil he has wrought. Domestic abuse literature almost always asserts that even a single act of domestic abuse, no matter how severe, is sufficient to earn a person the “batterer” label, since a single act of violence, even if mild, creates a perpetual implied threat of additional, potentially more severe, violence in the future. Yet, when research data shows that women commit as many acts of violence as men do, the data is attacked because it might be classifying as “batterers” those women who only “occasionally” batter men or who only do it “mildly.” No logical explanation has ever been given for why a single act of violence, no matter how severe, makes a person a “batterer” when the person is male, but does not make a person a “batterer,” even if the act is severe, when the person is female.
7 While attending court in Hennepin County, Minnesota a few years ago, I observed a pro se defendant enter a “guilty” plea to battery for the singular act of attempting to take his wife's car keys away in order to prevent her from driving drunk. A neighbor called police when she heard the drunk woman yelling profanities at the man. When the police arrived, the woman begged them not to take the man to jail, but the police said they had a “mandatory arrest” policy. Under the law, “battery” is defined as any unwanted touching, no matter how slight, so technically the man was guilty. The prosecutor's office pursued the charge because it had a “no-drop” policy in domestic cases. The end result is that the archival data, which consists simply of a tabulation of arrests and convictions without providing information about the nature or circumstances of the behavior upon which arrests and convictions are based, now reflects that yet another man was arrested and convicted of “battering” his wife in Hennepin County, Minnesota.
8 Persecution of the male gender has reached a feverish pitch in recent times. For example, male infants and toddlers as young as two years old have been cited, charged, publicly disciplined and sued for such “outrageous” conduct as giving a friendly hug to a female classmate in the day-care center. See Sommers, Christina Hoff, The War Against Boys: How Misguided Feminism Is Harming Our Young Men, New York: Simon and Schuster (1994)
10 See, e.g., Flood, Michael, “Responding to men's rights”, XY: men, sex, politics, 7(2), Spring 1997; Flood, Michael, “Contemporary Issues For Men,” XY: men, sex, politics, 5(1), Autumn 1995; Flood, Michael, “Claims About Husband Battering,” DVIRC Newsletter (Summer, 1999), pp. 3-8; and Orman, Kate, The Battered Husband Controversy, web document (2001)
11 Kennedy, L.S. and D. G. Dutton, “The Incidence of Wife Assault in Alberta,” University of Alberta, Canada: Population Research Laboratory (1987).
13 Straus and Gelles (1990), supra; see also Straus, et. al., “The Revised Conflict Tactics Scale (CTS2): Development and Preliminary Psychometric Data,” Journal of Family Issues, vol. 17, no. 3 (May, 1996), pp. 283-316; and Minnesota Center Against Violence and Abuse, Incidence Rates of Violence Against Women: A Comparison of the Redesigned National Crime Victimization Survey and the 1985 National Family Violence Survey, Washington, D.C.: U.S. Violence Against Women Office (1998)
14 As a matter of fact, the CTS has also been criticized for including “choking” as an item in the “severe” category. See Bagshaw, Dale and Donna Chung, Women, Men and Domestic Violence, University of South Australia: Pirie (2000)
15 Ibid; see also Flood (1995), supra; Flood (1999), supra; Orman (2001), supra; Smith, M. D., “The Incidence and Prevalence of Woman Abuse in Toronto,” Violence and Victims, 2, p. 173-187 (1987); and DeKeseredy, Walter S., Four Variations of Family Violence: A Review of Sociological Research, Ottawa, Ont., Canada: National Clearinghouse on Family Violence (1993) The most recent Statistics Profile issued by Statistics Canada includes choking, and finds that men are twice as likely as women to be victims of this kind of violence. Statistics Canada, supra.
16 Mouradian, Vera, Abuse in Intimate Relationships: Defining the Multiple Dimensions and Terms, Wellesley College: National Violence Against Women Prevention Research Center(2000.) Mouradian cites Gondolf, E. W., “Who are those guys? Toward a behavioral typology of batterers, Violence and Victims, 3, 187-203 (1988); Gray, H. M. and V. Foshee, “Adolescent dating violence: Differences between one-sided and mutually violent profiles,” Journal of Interpersonal Violence, 12, 126-141 (1997); Hudson, W. W. and S. R. McIntosh, “The assessment of spouse abuse: two quantifiable dimensions,” Journal of Marriage and the Family, 43, 873-886 (1981); Makepeace, supra; Marshall, L.L., “Development of the Severity of Violence Against Women Scales,” Journal of Family Violence, 7, 103-121 (1992); Marshall, L.L., “The Severity of Violence Against Men Scales,” Journal of Family Violence, 7, 189-203 (1992); Pan, H. S., P.H. Neidig and K.D. O'Leary, “Male-female and aggressor-victim differences in the factor structure of the modified Conflict Tactics scale,” Journal of Interpersonal Violence, 9, 366-382 (1994); Sheppard, M. F. and J.A. Campbell, “The Abusive Behavior Inventory: a measure of psychological and physical abuse,” Journal of Interpersonal Violence, 7, 291-305 (1992); Straus and Gelles (1986), supra; Straus, et. al., (1996), supra; and Tjaden, P. and N. Thoennes, “Prevalence and consequences of male-to-female and female-to-male intimate partner violence as measured by the National Violence Against Women Survey,” Violence Against Women, 6, 142-161 (2000). Headey, Scott and de Vaus specifically measured scratching, and found that women perpetrate this kind of violence more frequently than men do. Headey, Scott and de Vaus, supra.
17 Straus, et. al. (1996), supra.
18 See Dobash, Russell P., et. al., “The myth of sexual symmetry in marital violence,” Social Problems, 39 (1992), pp 71-91.
19 Straus, et. al. (1996), supra.
20 See, e.g., the discussion in Chapter 11 of the Violence Against Women Survey. According to the Violence Against Women Survey, an assault on a vagina or an anus is a “sexual assault,” but an assault on a penis or testicle is not.
21 Straus, et. al. (1996), supra.
22 U.S. Bureau of Justice Statistics, Sex Offenses and Offenders: An Analysis of Data on Rape and Sexual Assault, NCJ-163392, U.S. Department of Justice: Washington, D.C. (1997)
23 Claims about the incidence of domestic violence against women are all over the map. The Violence Against Women Office claims that 1 in 4 (25%) of women are abused. Tjaden and Thoennes (1999), supra. UNICEF puts the figure at between 20% and 50%. Kapoor, Sushma, Domestic Violence Against Women and Girls, UNICEF: Innocenti Research Centre (June, 2000.) Some sources estimate it's around 5%, while some feminist scholars have maintained that 100% is the correct figure.
24 The proportion is even lower when the fact that not all rapes and sexual assaults are perpetrated by husbands and boyfriends is taken into consideration. A fair number of rapes and sexual assaults are perpetrated by strangers and acquaintances, not husbands and boyfriends.
25 See, e.g., Flood (1995), supra.
26 See, e.g., Orman, supra; and Flood (1995), supra.
28 Most stalking cases do not involve intimate or romantic partners. Zona, M. A., R.E. Palarea and J.C. Lane Jr., “Psychiatric diagnosis of the offender-victim typology of stalking,” in J. R. Meloy, ed., The Psychology of Stalking: Clinical and Forensic Perspectives, San Diego, Calif.: Academic Press (1998)
29 Harmon, R. B., R. Rosner and H. Owens, “Sex and violence in a forensic population of obsessional harassers,” Psychology, Public Policy, and Law, 4, 236-249 (1998)
30 Tjaden and Thoennes use a slightly narrower definition, under which some showing of a reasonable possibility of fear on the part of another person is required. The test is still a subjective one, however, in the sense that it is not necessary to actually intend to cause fear, so long as someone else might experience fear, as a result of your visual or audible proximity to them. Tjaden and Thoennes (2000), supra.
31 For reasons outlined in Chapter 11, the Violence Against Women Survey funded by the Violence Against Women Grants Office is an example of a study that was conducted in allegiance to this principle.
32 Statistics Canada (2000), supra. Some of the studies that have attempted to measure the incidence of stalking include: Fremouw, W. J., D. Westrup and J. Pennypacker, “Stalking on campus: The prevalence and strategies for coping with stalking,” Journal of Forensic Science, 42, 666-669 (1997); Harmon, Rosner and Owens, supra; Tjaden and Thoennes (2000), supra; Walker, L. E. and J. R. Meloy, “Stalking and domestic violence,” in J. R. Meloy, ed., The Psychology of Stalking: Clinical and Forensic Perspectives, supra, pp. 140-161; Meloy, J. R. and S. Gothard, “A demographic and clinical comparison of obsessional followers and offenders with mental disorders,” American Journal of Psychiatry, 152, 258-263 (1995)
33 Cf. Jouriles and O'Leary, supra; and Szinovacz, supra.
34 There have been feminist scholars, for example, who have published books and articles detailing how Beethoven's symphonies are “all about rape.” Women's magazines are replete with articles such as “The Ten Signs Your Boyfriend Might Be A Batterer,” and the like.
35 Moffitt and Caspi, supra. See also Morse, supra; National Institute of Justice, Partner Violence Among Young Adults, supra; and Magdol, et. al., supra.
36 Compare Moffitt and Caspi, supra; Morse, supra; National Institute of Justice, Partner Violence Among Young Adults, supra; Magdol, et. al., supra; Burke, Stets and Pirog-Good, supra; Minneapolis Star-Tribune/WCCO, supra; Foshee, supra; O'Leary, supra; Vivian and Langhinrichsen-Rohling; and Straus and Gelles (1986), supra.
37 Straus and Gelles (1986), supra.
38 The Betty Friedan story is instructive in this respect. In one of her feminist tracts, she went into elaborate detail about how she supposedly had been severely abused during her marriage to what she portrayed as a patriarchal monster. Her story had the intended effect: she won immediate favor, sympathy and attention from the media (not to mention increased book sales.) Eventually, however, evidence came out that her ex-husband, far from being a patriarchal monster, had actually been a strong supporter both of Ms. Friedan personally and of the feminist cause in general. Apparently feeling pangs of conscience about smearing an innocent person's reputation for her own personal gain, Ms. Friedan started urging the media to stop talking so much about the nature of her relationship to her ex-husband. Ultimately, she acknowledged that it was she who had been abusive during the relationship.
39 See, e.g., Giles-Sims, Jean, “The Psychological and Social Impact of Partner Violence,” Domestic Violence Literature Review, Synthesis, and Implications for Practice, United States Air Force and the National Network for Family Resiliency (1997). In fairness, it should be pointed out that although Ms. Giles-Sims seems to take the position that domestic violence is worse for female victims than for male victims, she does acknowledge that “research on psychological consequences of partner violence has varied in quality of measurement,” that “there is a limited amount of information...on males as victims,” and that “more research is needed on consequences for male victims. Ibid.
40 see, e.g., Grotevant, H.D. and C.I. Carlson, Family Assessment: A Guide to Methods and Measures, New York: Guilford (1989); Hertzberger, S.D., “The Conflict Tactics Scales,” in D. J. Keyser and R. C. Sweetland, eds., Test Critiques, 8, Kansas City Test Corporation of America (1991); Plichta, Stacey B., “Violence and abuse: Implications for women's health,” Women's Health: The Commonwealth Fund Survey (1996), p. 237-270, at p. 240 (“calling the CTS “a widely used instrument with good reliability and validity.”)
41 Walsh, M.R., ed., Women, Men and Gender: Ongoing Debates, New Haven: Yale University Press (1997), p. 218.
42 Reliance on archival data can also be dangerous when it is used in combination with faulty logic. Consider this passage from Hague and Malos' treatise on domestic violence: “activists [in] the domestic violence field know that it is simply not true that as many men are abused by women as vice versa. There are no refuges for abused men....One phone line exists in the country [for] men....” Hague, G. and E. Malos, Domestic Violence: Action for Change, New Clarion Press: Great Britain (1993). Notice how the authors cite evidence of the societal response to male victims as if it were evidence of the incidence of violence against men. This is like citing the unavailability of colleges for black people in the early twentieth century as if it were evidence of lower levels of intelligence among black people in general.
| EJF Home | Join the EJF | Comments? | Get EJF newsletter | Newsletters |
|| DV Home | Abstract | Contents | Tables | Index | Bibliography |
| Chapter 4 — Domestic Violence Statistics |
| Next — Battered women and battered statistics |
| Back — NIMH studies of domestic violence |
This site is supported and maintained by the Equal Justice Foundation.