How to Lie with Sexual Statistics

In 1954, a freelance American writer named Darrell Huff published How to Lie with Statistics, a book which sold 1.5 million copies, introduced thousands of people to the abuse of numbers, and, if it achieved its purpose, prevented many conmen from fooling the gullible. I was reminded of this work recently when considerable publicity was given to two pieces of apparent survey research, “‘Submit to Your Husbands’: Women Told to Endure Domestic Violence in the Name of God” by Julia Baird and Hayley Gleeson of the ABC, and Change the Course: National Report on Sexual Assault and Sexual Harassment at Australian Universities, drafted by eight members of the Australian Human Rights Commission in Canberra. While both have already been widely and rightly criticised on a variety of grounds, I doubt that even the critics realise just how dubious the two studies actually are.

With “Submit to Your Husbands”, the main problem for a critic is where to start. The conclusion of this report is that “the men most likely to abuse their wives are evangelical Christians who attend church sporadically”, and that, more generally, fundamentalist Christians in Australia have a great problem with domestic violence, which church leaders are “both enabling and concealing”. In fact the report shows nothing of the sort. As difficult as this may be to believe, the two authors who compiled the study have, in fact, done no research themselves of any kind on the Australian situation. The few Australian studies they cite show, if anything, the very opposite of what they claim.

The authors depend almost entirely on a handful of American studies, which are themselves of dubious quality. But the religious situation in the United States is categorically different from the religious situation in Australia. Church-going is vastly greater there than here, among both mainstream and fringe denominations. There are, for instance, 2.1 million Jehovah’s Witnesses in America, compared with 54,000 here; 2.7 million Seventh Day Adventists there compared with 67,000 here; and 71 million Catholics in America, compared with 5.4 million here. Adherents of fundamentalist Christian churches are actually an insignificant component of the Australian population. According to the 2016 Census, only 1.1 per cent of the population describe themselves as “Pentecostals”, with only 0.7 per cent members of minor fundamentalist denominations like the Jehovah’s Witnesses. The two authors have not studied these fundamentalist denominations here, and have provided no evidence at all as to how common domestic violence is among these groups. Nor have they provided any comparative data on how this compares with mainstream Christians, with adherents of non-Christian religions, or with non-believers.

This essay appears in the October edition of Quadrant.
Click here to subscribe

Perhaps only three pieces of actual Australian research relevant to their report are cited by them. One is a 2010 book by Dr Lynne Baker, Counselling Christian Women on How to Deal with Domestic Violence, which found that 22 per cent of the perpetrators of domestic violence and abuse go to church regularly. In other words, 78 per cent of the perpetrators of domestic violence did not go to church regularly—the great majority—which hardly bears out the authors’ claims. Moreover, this was not a survey of all Australian church-goers, but “a study of Anglican, Catholic and Uniting churches in Brisbane”, a single city. The authors also cite a 1992 study of 1704 Anglican and Uniting Church women undertaken by Glenys Conrade of the University of Queensland, which “found that 3.3 per cent reported they had been abused in the past 12 months”, while “22 per cent of the perpetrators went to church regularly”. Thus, 96.7 per cent of these women had not been abused in the past twelve months, and 78 per cent of the perpetrators did not attend church regularly—surely the more salient way of viewing the evidence. If anything, by their figures church-going acts as a deterrent to domestic violence.

Finally, the authors cite “an anonymous [2016] survey of 148 Sydney Anglican rectors”, “which found that, on average, each rector had seen 2.25” cases of domestic violence “in the past five years”, or one case every two years—surely not evidence of an epidemic, regardless of how deplorable each case was. The same page on the ABC News site on which they reported this also stated that “police in Australia” deal on average “with 433 cases of domestic violence each day”. Of these, presumably 100 or so take place in Sydney each day, or 73,000 in two years, compared with the one case known to these Anglican rectors (or 148 among all of them) in the same two-year period. How on earth can it be said that church-going increases the likelihood of domestic violence, as the two authors claim? On the ABC site, there is also repeated innuendo about Australia’s churches, such as “Australian Churches Risk Becoming a ‘Haven’ for Abuses, US Seminary Professor Says”, which quotes Ruth Tucker, “a seminary professor living in Michigan”, without stating whether she has ever set foot in this country or knows anything whatever about the situation here.

The researchers also apparently do not understand the difference between a cause and an association, a distinction fundamental to any social research. It may well be the case that married men who eat two or more meat pies at AFL matches beat their wives more often than married men who eat one meat pie, or none. Eating several meat pies may well cause these men to reach for their Zantac, but it does not, so far as anyone knows, cause them to reach for a cricket bat to hit their wives. Much the same is almost certainly the case with evangelical Christians, even assuming that they are more likely than others to beat their wives.

Who are these Australian evangelical Christians? This basic question is not addressed by the authors. It is almost certainly the case that these evangelical Christians are not drawn randomly, but disproportionately from specific demographic groups. The majority, or a significant minority, are almost certainly of Eastern European, Pacific Islander or African background, ethnic groups among whom domestic violence may well be more common and more entrenched than among Anglo-Celtic Australians. They are very likely to be less well educated than the population as a whole and, in particular, poorer, with more children’s mouths to feed, and more financially insecure than the general population. If they also engage in domestic violence at a higher rate than the average, these factors are very likely more important than what their minister says. In considering the astronomical rates of domestic violence among Aborigines, their poverty and marginality would assuredly be posited by Left-liberal commentators as the main cause of this violence. But in the case of evangelical Christians, the two authors fail to consider any causal factor apart from their supposed religious beliefs.

In a later publication on ABC News (“What Does the Australian Research Say?”, July 24), the two authors do make a claim with which one can agree: “There is very little research on the nature and prevalence of domestic abuse in church communities, unlike other countries, and most is dated … Comprehensive, independent Australian data regarding domestic violence within the churches are long overdue.” This highly original way of approaching the subject makes extravagant, perhaps defamatory claims about domestic violence in the churches, and then hopes that someone will do the research which will prove that their claims are true.

This examination of domestic violence and Australian churches appeared in instalments on the ABC News website. Anyone familiar with this site will know that, apart from straightforward news reports, a common hallmark is to carry planted stories, a propos of nothing in the news, on issues favoured by the Left, and invariably from a Left-liberal perspective. A gratuitous, unsubstantiated attack on evangelical Christians in Australia, needless to say, fits right in with the agenda of “your” ABC, and should be seen in this light.

The second survey, about sexual assault and sexual harassment at Australian universities, is no less dubious in its methodology and findings. Sponsored by the Human Rights Commission (HRC), it found that 51 per cent of all university students were sexually harassed on at least one occasion in 2016, and 6.9 per cent were sexually assaulted on at least one occasion in 2015 or 2016. To produce these findings, an invitation was sent to no fewer than 319,959 students at thirty-nine Australian universities via e-mail to participate in a survey about sexual harassment. A total of 30,930 students (9.7 per cent of those invited to participate) replied. The conclusions of the HRC were thus based entirely upon this sample of students who voluntarily participated. It then weighted this sample by gender, domestic versus international students, undergraduates versus postgraduates, and claimed that their findings were subject to a standard error of only plus-or-minus 0.4 per cent.

This is nonsense. The entire design of the survey is wholly flawed, and its findings are virtually useless. The proper way to conduct a survey of this type is to take a random sample of the 319,959 students sent e-mails, based on a computer-generated table of random numbers, and survey say 5000 of them. If a student so selected does not participate, one then finds another randomly selected student to survey. This is the methodology employed in all election polls, and usually produces numbers which are within 3 to 5 per cent of the actual election results.

The methodology employed by the HRC, consisting entirely of voluntary respondents, omits over 90 per cent of all students from the survey, and is virtually certain to overstate, perhaps wildly, the percentage of students who claim to have been sexually harassed, since it will be they in all likelihood who reply. Their procedure would be like asking 300,000 Americans by e-mail what they think of Donald Trump: 20,000 reply, of whom 14,000 think he is a monster, 3000 the greatest thing since sliced bread, and 3000 have no opinion. Such results are obviously invalid as a random sample, which, if conducted in a valid way, is likely to find a more even division of opinion. The methodology employed by the HRC in their survey would not be recognised as valid by any professional polling agency.

Change the Course then compounded their mistake by—believe it or not—including in their findings alleged sexual harassment which occurred on public transport to and from a university, a fact buried on pages 224 to 226 of the report. In other words, if some ratbag in Preston, three kilometres from Melbourne University, makes a lewd remark to a female student travelling to the university by tram, this is included as an example of sexual harassment at Melbourne University. This is just preposterous. The ratbag in question has no connection with the university, and Melbourne University has no control over what occurs on trams in Preston. It is obvious that harassment on public transport has been included simply to pad out the negative figures. In fact 22 per cent of all students who claimed to be sexually harassed said it had occurred on public transport. And if sexual harassment on public transport is included, why not the harassment of university students in pubs, off-campus parties, or on the beach?

Without minimising the impact of sexual harassment on young and vulnerable students, the impression gained from carefully considering these findings, as exaggerated as they probably are, is that there are very low levels of alleged harassment and still less assault at our universities. Ninety-four per cent of students who claimed to have been sexually harassed, and 87 per cent of those who claimed to have been sexually assaulted, did not make a formal complaint or report it to their university. Among students who witnessed another student being sexually harassed and did nothing about it, the most important reason was because they did not think it serious enough. Among witnesses even to a sexual assault—133 persons—only eight called the police. Among students who claimed to be sexually assaulted—a serious crime which should be reported to the police—only twenty-six women and two men actually called the police.

The most common forms of “sexual harassment” were “inappropriate staring or leering” (14 per cent), “sexually suggestive comments or jokes” (11 per cent), and “intrusive questions about an individual’s private life or physical appearance” (9 per cent). The only reasonable conclusion from this is that sexual harassment, let alone criminal assaults, are extremely rare. Many of these take place in halls of residence, or at parties where the alcohol flows, just as they would among non-university youth, and what does one expect from normal eighteen-to-twenty-two-year-olds? Expecting there to be no sexual undercurrent among them is like hoping that water will run uphill.

Moreover, the claims reported in the survey are, without exception, unsubstantiated accounts set out in e-mail responses to the HRC, and are all made without evidence. The HRC apparently did not investigate a single claim made to it, let alone contact the alleged perpetrators for their versions of these events, which might be very different.

One might expect the Australian Human Rights Commission, of all institutions, to be cognisant of the rights of the accused; it would be the first body to scream if, say, Aborigines or asylum seekers were not allowed to defend themselves from serious accusations. But as with the ABC and its report on domestic violence, the HRC has a track record of partisan advocacy, exemplified by the pronouncements of its former head, Gillian Triggs. In this case, it appears to have resorted to highly dubious, if not overtly fraudulent means to find what it wanted to find.