Lies, Damned Lies, and Lenore Weitzman
by Cristopher Rapp

Since its publication in 1985, sociologist Lenore Weitzman's The Divorce Revolution has had a critical role in shaping the national debate on divorce and its economic effects. In particular, the book's claim that in the year after divorce women's standard of living decreased by a whopping 73 percent while men enjoyed an increase of 43 percent caught the attention of pundits, legislators, and judges. This statistic has become one of the philosophical bases for deciding child custody and property division in divorce cases. It has also altered public perceptions of men, women, and divorce. It was cited hundreds of times in news stories, scholarly studies, and law review articles last year, and was regarded so clearly as holy writ that President Clinton cited it too in his budget proposal this year as part of his attack on deadbeat dads.

The only problem with this statistic, in fact, is that it turns out to be wrong.

In June of this year, Richard Peterson of the Social Science Research Council published a study of Weitzman's 73/42 statistic, which was arrived at using an "income/needs ratio." After precisely recreating Weitzman's study using the data sample and methods outlined in The Divorce Revolution, Peterson reported his findings: Weitzman's figures were actually the result of a computer transcription error and dramatically overstated the case. After correcting her errors, Peterson arrived at a 27 percent decrease in standard of living for women and a 10 percent increase for men in the first year after divorce-figures more in line with other studies dealing with this topic.

Because his corrected figures may actually still overstate the inequalities in the economics of divorce, Peterson's revision of Weitzman's numbers may ironically continue the distortion of the truth. As the media slowly begins to use Peterson's calculations to correct its uncritical acceptance of Weitzman's 73/42 statistic, it may lose in the shuffle the growing body of scientific and anecdotal evidence indicating that both women and men suffer economically after a divorce.

Amidst the hosannas for Weitzman's findings that echoed in the nation's courtrooms, lecture halls, and legislative chambers, a few researchers expressed some doubts about the accuracy of the 73/42 statistic. Some critics charged that her sample-228 people who had been divorced in 1977-78-was too small to be representative. Furthermore, the respondents were all from Los Angeles, an area which has its own unique culture of divorce and divorce laws. These concerns, however, received little play in the press, and Weitzman refused to let other researchers examine her data set, shielding her research from further scrutiny.

Atlee Stroup, professor emeritus of sociology at the College of Wooster (Ohio), remembers being surprised when the 73/42 statistic first hit the scene in 1985. He had been specializing in family sociology since the 1950's and felt that Weitzman's numbers were too extreme to be accepted at face value. "A general feeling that this ought to be evaluated very carefully led me to think about doing some kind of study," Stroup says.

The National Opinion Research Center, a sociological organization affiliated with the University of Chicago, provided Stroup with the data sample he needed to ensure that the study he undertook would avoid Weitzman's parochialism. Every year, the Center surveys roughly 1,500 adults, creating a national data sample representative of the major socioeconomic segments of American society. Along with Gene Pollock, a professor of economics at the College and an expert statistician, Stroup combined the surveys from the years 1983-1987, creating a cumulative data set of close to 7,500 respondents. Armed with this sizable, national data bank, Stroup and Pollock brought their considerable experience-a combined seven decades of research and teaching-to bear on the question of the economic repercussions of divorce.

They found that women and men, at every socioeconomic level, experience a decline in income after divorce. According to their data, women in the first year after divorce experience on average a 22 percent decline in family income, with professional women's family incomes declining the least (12 percent) and unskilled laborers declining the most (30 percent). These figures were far less dramatic than Weitzman's 73 percent, and comparable to, although still lower than, other studies methodologically similar to Weitzman's, which suggested an average 30 percent decrease for women.

When they looked at the status of men, however, Stroup and Pollock uncovered surprising information. "Keeping in mind the suggestion of Weitzman and others of economic gain by males with divorce," they wrote, "the results are sharply contrary to expectations." Instead of the 42 percent increase reported by Weitzman or the more common 10 percent figure, the data indicated an average 10 percent decrease in income, with professional men experiencing a decline of 8 percent and less-educated workers a drop of 19 percent. Stroup and Pollock wrote that Weitzman's "sharp generalization of a 42 percent rise in living standards for males certainly does not hold for our sample." More importantly, their findings presented an implicit challenge to the studies which reported lower figures than Weitzman but still agreed with the conventional wisdom that men benefit from divorce.

The findings of Stroup and Pollock, first presented at a conference of the National Council on Family Relations in 1992, and subsequently published in 1994 as "Economic Consequences of Marital Dissolution" in the Journal of Divorce and Remarriage, initially attracted only a modicum of attention, Weitzman's claims having by that time achieved a hammerlock on public opinion. But the results came as no surprise to men who had gone through a divorce; Pollock says that one editor at the Akron (OH) Beacon-Journal commented that he and Stroup had "discovered that water is wet." Unfortunately-and equally unsurprising given the politically correct zeitgeist concerning these issues-the observers in academia, government, and the media failed to acknowledge the implications of Stroup's and Pollock's research, and the promotion of Weitzman's 73/42 statistic proceeded apace. Her figures were cited in over 170 newspaper and magazine articles, 350 social science articles, 250 law review articles, 24 state appellate and supreme court cases, and one U.S. Supreme Court decision. The American Sociological Association awarded The Divorce Revolution its 1986 Book Award for "Distinguished Contribution to Scholarship." Weitzman repeated the statistic when she testified before the U.S. Congress, and legislatures across the nation revisited their divorce laws in response to her claims. (Weitzman herself takes credit for influencing 14 laws in California alone.) The attention culminated with the statistic's appearance in President Clinton's 1996 budget proposal.

Before long, the 73/42 statistic had for some observers attained the status of received truth. One enthusiast, for example, wrote in 1991 that the numbers "are far less surprising than is the fact that people have been so surprised by them."

Weitzman's numbers had been accepted so totally in media, legal, and social science circles, explains USC sociologist Angela James, that research which conflicted with them was for the most part disregarded. "It was taken as a given that [the numbers] were correct," she says. "I think there were some concerns raised on the part of scholars in that area of research, but they did not get nearly as much attention as the statistics themselves. Almost every article on the subject-and in many related subjects-cited that statistic...New scholars coming to the arena may not have read the skeptical reviews but they definitely knew of Lenore Weitzman's book."

How did this happen? How did the 73/42 statistic, so out of line with other scholarships as well as the real-life experiences of divorced men, gain such power in the media, government, and academy for more than a decade?

The answer was that in the increasingly radicalized atmosphere that characterized the debate over the economics of divorce, the 73/42 statistic had the force of an idea whose time had come. The disparity it pretended to uncover was so dramatic that it became the perfect media sound-bite, a shocking factoid which after many repetitions seemed to ring true. Perhaps more importantly, Weitzman's numbers could be used to promote just about any agenda, further ensuring the statistic's popularity. Weitzman herself, for example, said that her study indicated the need for changes within the no-fault system of divorce which by 1986 was in place in 48 states. Others saw the statistics and recommended scrapping no-fault entirely.

A large number of people, including most feminists, cited the statistic as proof that child support levels needed to be raised in order to equalize post-divorce standards of living, an argument which, according to Maggie Gallagher, author of The Abolition of Marriage, carried significant weight in the press. "I think the reason [the 73/42 statistic] played so broadly was that it was part of this big push to say that there is an easy solution: 'we should increase child support.'" Other, more radical feminists, most notably Susan Faludi, disputed Weitzman's findings, charging that the media's acceptance of them was evidence of a pervasive anti-feminist "backlash" which threatened to reverse the "gains" made by the women's movement in the area of divorce law.

Christina Hoff Sommers, professor of philosophy at Clark University and author of Who Stole Feminism?, explains that although some feminists valued the 73/42 statistic as evidence of patriarchal discrimination against women and others criticized it as a threat to the cause of feminine independence, Weitzman's findings actually served the same purpose for both camps. "Lenore Weitzman's research is used by many groups of feminists to trash matter what," she says. "It's either 'men are monsters,' or 'we don't need them.'"

Sociologist Atlee Stroup says that the 73/42 statistic also fit society's preconceptions about divorced men. "I think that, interestingly, we still have a concept of the carefree man floating around [after divorce]-now he doesn't have his 'burdens' and he's off again. It's the idea that men become part of that 'bachelor set' again, with no responsibilities and a lot of money."

Stroup's collaborator Gene Pollock has a simpler explanation: "I think it was a politically correct statistic."

As horror stories of a 115 percent disparity between men's and women's post-divorce standard of living made their way through the legislatures and editorial boards of the nation, Weitzman ensured the success of her 73/42 statistic by refusing to allow other researchers access to her data, claiming that she wanted to correct some errors in the master computer file before doing so. She had every right to do this, at least at first, explains Richard Peterson. "There are some norms that are generally accepted," he says. "If you collect data you have the right to keep it to yourself and not be required to share it with others until you publish from your data."

But when Weitzman's data files arrived at the archives of the Murray Research Center at Radcliffe College she had still not made the corrections, and what started as the exercise of her rights as a researcher began to look suspiciously like ten years of stonewalling. She reserved to herself the right to veto anyone from looking at the material and turned down Peterson's requests. It was not until a year and a half had gone by and the National Science Foundation, the organization which had funded Weitzman's research, threatened to declare her ineligible for federal grants in the future that she finally allowed Peterson to examine the data.

Once given access to the files, Peterson began to recreate Weitzman's study using exactly the same 228-person sample and the methods described in The Divorce Revolution. He found that the information in Weitzman's computer file in many cases did not match up with the paper records of the original respondent interviews. "The computer file was supposed to be coded from the paper records," he says, "but the computer file in fact did not reflect the paper records. For example, suppose the computer file said a person's income was $27,000 last year, but when I looked at the paper records it turned out that it was $37,000." After correcting the data file, Peterson arrived at figures more in line with national studies, with women's standard of living decreasing by 27 percent in the first year after divorce and men's standard of living increasing by 10 percent.

Peterson's findings are a welcome improvement over Weitzman's hyperbolic statistics; one feels a measure of relief knowing that the nation's elite will no longer be basing its decisions on a typo. But even Peterson's more moderate figures seem out of step with the research done by Stroup and Pollock, not to mention the real life experiences of divorced men across the country, who would likely be surprised to hear that getting divorced improved their standard of living.

The problem seems to lie in the "ratio of income to needs" used by Weitzman, Peterson, and others. Weitzman asked 114 women and 114 men to report pre- and post-divorce income for both themselves and their spouses, adjusting for alimony and child support payments where applicable. Then, she compared these figures to each person's "economic need," based on the Bureau of Labor Statistics' Lower Standard Budget for an urban family of four in 1977, to arrive at her ratio. Even as corrected by Peterson, her data was sketchy: income or needs data was missing for 134 of the respondents, and, according to Peterson, the data for family size, age of the household head, oldest child, and other relevant variables was "problematic" with notable "inconsistencies." From this admittedly flawed data set, an income/needs ratio was derived, and the results labeled "standard of living."

Gene Pollock feels that this measure of needs is unreliable. "I think you can talk about standard of living based on income," he says, "but when you start talking about income as a ratio to needs, that is a very slippery concept."

In the real world, standard of living is determined not simply by a ratio of income to needs, but also by a lifetime of economic choices, earnings, investments, and purchases. An accurate measure of standard of living would necessarily include property owned, savings, houses, furniture, automobiles, the neighborhood one lives in, clothing, and the like, most of which is typically divided between spouses in a divorce settlement. It seems unlikely that the average man would have a 10 percent higher standard of living after losing half of his marital property, much less a 42 percent increase. Pollock points out that Weitzman's income/needs ratio "omits completely any prior assets that people have. In divorce settlements, so much of the issue is who gets what, and it's not just alimony or child support but also the division of the other assets. And most of the studies that I've read have omitted that part."

Warren Farrell, the San Diego-based author of The Myth of Male Power, agrees with Pollock and adds that Weitzman's and other studies usually omit several expenses which men face. "There are five expenses that men have after divorce, typically speaking, that Weitzman just didn't measure," he says. These include: mortgage payments on a home they no longer live in, rent on a home or apartment they do live in, child support payments, alimony, and higher percentages of dating expenses. Says Farrell, "No one that I know has controlled for all five of these variables."

Pollock feels that using this "measure of needs" as a proxy for standard of living is misleading. "Years ago people used to say about computers, 'garbage in, garbage out.'" he says. "The thing is, you can make [the income/needs ratio] sound quantitative by calling it 'standard of living,' and it will seem like hard and fast numbers you are using because somehow you adjust income for needs. But you make a strong implicit assumption that your measure of needs is correct. Even though the bulk of people write it that way, that still doesn't make it the right methodology as far as I'm concerned. Maybe I'm just being stubborn, but as a statistician, you don't quantify something by just arbitrarily deciding how you are going to measure one of the variables."

For over a decade statistics like Weitzman's-which suggest that women suffer tragically from divorce while men blithely benefit-have been at the center of the national discussion of divorce and its economic effects. They have served as what Thomas Sowell calls "aha!" statistics, appearing to confirm in dramatic fashion the worst expectations about the American social system. In Weitzman's case, the 73/42 statistic suggests at bare minimum a jarring unfairness, perhaps even an outright misogyny. It is easy to see how for radical feminists it would become part of a wish-fulfilling fantasy.

The real power of statistics is cultural, their ability to shape the way society thinks, discusses, and acts. Peterson has corrected the errors in Weitzman's study, but has given a veneer of legitimacy to the conclusions which for eleven years decision-makers in the media, academy, and government have drawn from it. While the 73/42 statistic may have reached the end of its shelf-life, the idea that divorce catapults men into some sort of financial Elysian fields will likely continue to influence public policy and discussion, at least until researchers like Stroup and Pollock-as well as the stories of thousands of divorced men-get the attention they deserve.