Not all trigger warnings are the same


I follow the debate over trigger warnings only loosely. Please feel free to add information in the comments.

In what I see, the debate over trigger warnings is hampered by ill-defined terms and unhelpful hyperbole. I want to give a very basic description of what I think should be a relatively simple approach to the issue, call out a gender problem, and then offer my own example.

To show you where I’m coming from: What prompted me finally to write this was the combination of this popular op-ed by Judith Shulevitz, this essay about the problem of teaching about rape in law school, and the flap over Christina Hoff Sommers’s anti-anti-rape-culture campus tour. I noted that a letter to the editor in the Oberlin Review about her upcoming talk began with this: “Content Warning: This letter contains discussion of rape culture, online harassment, victim blaming and rape apologism/denialism.”

Impending discourse

There are three kinds of relevant warnings that I would group together under the category of “impending discourse notification.” That is, warnings that take the form: something is about to be discussed or displayed. Keeping these three things straight would be really helpful.

1. Warnings of content likely to be disturbing to many people in the audience.

For example, graphic images of violence during a regular TV news program, descriptions of rape on NPR’s Morning Edition, or sociology classroom lectures that contain images of Blacks being lynched. In these cases, a warning of the impending discourse is something like common courtesy. It says, “we are about to see or hear something important enough to risk disturbing the audience, and potentially disturbing enough that you should gird yourself.” In these discrete cases warnings are not controversial in principle, though of course individual applications may be off target or offensive. Many settings carry an implied warning: A horror film can be expected to surprise you with specific acts of violence, but you know something bad is coming; a sociology class on racial inequality should be expected to include discussions of lynching, though some students have no idea about lynching; a history documentary on war is expected to show people being killed. Warnings in these cases seem optional.

2. Warnings of content that may trigger post-traumatic stress responses.

I am not expert on this, obviously, but my understanding (from, e.g., here) is that post-traumatic stress disorder (PTSD) was included in the DSM III in 1980, partly based on the experience of Vietnam War veterans. The condition was understood to involve reliving memories of trauma, avoiding reminders of trauma, and hyperarousal that can lead to high levels of distress. There are many kinds of traumas that can lead to PTSD, but some are much more common than others, especially violence, sexual abuse, and existential threats. You can’t expect to prevent all triggering events, but you can take steps to avoid common ones, or warn people when you are going to show or discuss something to an audience likely to be include people with PTSD. Again, war movies are expected to show graphically violent war scenes, but lectures to audiences of combat veterans about disability benefits should not. This is a question of sensitivity and awareness, not blanket prohibitions and censoring. And this is about shocking or graphic imagery, not mere mention of a topic. We just can’t have a democratic discourse without mentioning bad things, sometimes spontaneously. The Oberlin newspaper warning above is wrong. And I don’t agree with another Oberlin essayist who says trigger warnings should be treated as disability accommodations, “as common as wheelchair ramps.” (Of course, I would make an accommodation for a specific student — and I have — who asks to opt out of a specific class session based on the topic).

3. Warnings of obnoxious, offensive, disagreeable, or dangerous ideas.

These warnings are unnecessary and wrong. If someone wants to say the problem of campus rape is exaggerated, that Black men are genetically aggressive, the Holocaust is a myth, or Creationists are stupid — let them. Hand out flyers or picket at their talk, discredit them in the Q&A, denounce them on Twitter, or ignore them. If they are receiving honorary degrees or other accolades (or money) from governments or universities, that’s political fair game to protest. But protecting people from hearing bad ideas is a bad idea (outside of incitement to violence). On campus or in the classroom, exposure to bad ideas is essential to critical intellectual development. If you’re never offended in college you aren’t learning enough.

A gender problem

I have complained elsewhere that the non-criminal procedure for responding to campus rape “downgrades sexual violence from a real crime to a women’s issue.” Something similar is going on with trigger warnings. Although PTSD-type responses can be triggered by many kinds of experiences, it looks like sexual violence is the main arena of debate over campus trigger warnings. Why? This should not be reduced to a “women’s issue.” My admittedly limited exposure to this debate often makes me cringe at what seems like a demand for special protection — from discourse — for women. Women are in fact more likely to experience PTSD than men, but that’s only partly because they are more likely to be sexually assaulted. Men are more likely to experience other potentially traumatic events, including accidents, nonsexual assaults, combat, or witnessing violence, all of which can lead to PTSD. People with sensitivity to trauma-related triggering deserve respect and sensitivity. But women — like any subordinate group — need to exert leadership in the discourse surrounding that inequality, and that doesn’t come from avoiding the topic or silencing their opponents. If the only people discussing rape are people who have never been raped, the dialogue is likely to be male-dominated. We have to work on maintaining the line between offensive and unpleasant on the one hand and truly trauma-inducing on the other. If it’s necessary to avoid the latter, it’s all the more important for those who are able to engage the former.

How did I do addendum

I think we can learn a lot from these discussions. They have raised the question, “What if we acted like sexual assault is actually common?” That reality is hard to grasp — for people who are victims or not — because the experience is so often private.* In the chapter in my book about family violence and abuse, I didn’t include an impending discourse notification, but — after opening with a detailed story of violent abuse — I raised the issue of how discussing the topic might affect students:

The subject of family violence and abuse is personal and painful. Instructors and students should pause at this point to consider the possible effects of discussing these topics, especially for those who have experienced abuse in their own lives. Because this kind of victimization still is so common in the United States, most of us will know someone who has been touched by it in one way or another. However, because families often are protected by a cultural—and sometimes legal—expectation of privacy and a shroud of secrecy, those who suffer usually do so in isolation. That leaves us with the complexity of a problem that is widespread but experienced alone and often invisibly. Such isolation can make the experience of abuse even worse. One benefit of addressing the issue in this book is that we can help pierce that isolation and encourage victims to realize that they are not alone.

I think advising people in the classroom to “pause to consider” before launching into the topic is reasonable — it’s a common experience with a known risk of traumatic effects. But I didn’t write that just to protect people who might have a traumatic reaction to the topic, I did it because it’s a learning opportunity for everyone.

* In the book I tried to put rape in normal-experience terms: experiencing rape (18% of women by one reasonable estimate) is more common than using the Pill for contraception (17% of women currently), but less common than smoking cigarettes for young-adult women (22%, ages 25-34). Does that help?


Filed under In the news, Me @ work

Regnerus responds

Photo by carnagenyc from Creative Commons.

Photo by carnagenyc from Creative Commons.

The news, reported in The Daily Texan, with documents retrieved via public records request, is that, in the face of conflicting views about Mark Regnerus’s promotion to the rank of full professor,  UT’s Dean of Liberal Arts, Randy Diehl, commissioned a report on the scandal by sociologist Marc Musick. The report is an excellent review and summary of the affair, and provides ample evidence for declining the promotion. And for the rest of us, it had the beneficial effect of flushing out Regnerus, who wrote his most detailed response yet to the accusations against him — a response he may or may not have realized would become public. (The new documents are linked in the Texan article; for my coverage, you can start here for a review with links.)

It’s difficult to try to draw a line, as Musick does, apparently at the dean’s request, between ethical misconduct and bad research. It’s really where the two are combined that Regnerus causes trouble. More on the promotion issue later.

Musick was entirely correct when he wrote:

Based on these [media] appearances and his [court] testimony, it is self-evident that Professor Regnerus has used his research in the debate over same-sex marriage in direct contradiction to the statements he made in the NFSS article and response to commentaries. When combined with clear evidence that he colluded with politically-motivated organizations prior to the publication of the study, it leads to the appearance that the post-study behavior was an extension of the political work that was happening prior to the study. In light of all of this activity, it appears that the statements he made in the article could certainly be seen as misleading at best and an outright fabrication of his intentions at worst.

This is the heart of the ethics side of the complaint: his bad research was part of a covertly-organized political effort, and he lied about it to cover that up. Regnerus simply asserts this isn’t true, but to believe his self-serving description of his own intentions is to be made a fool of. It’s just not plausible that,

I did not intend to utilize the results for any political or legal purpose, and stated so when I completed work on the manuscript in late February 2012. My interests, from the outset of participation in this project up through December 2012, lay squarely in the social science question that gave rise to the study.

Only God can truly see into the unlit depths of Regnerus’s heart — but the rest of us can be pretty sure he’s lying based on his actions.

Regnerus claims that as he became immersed in the subject he grew convinced that same-sex marriage is a bad policy, and began “to worry about esteeming the systematic severance of children from their biological origins.” But he was part of the “coalition” (his word!) against gay marriage from before the study was even fielded. His email to Brad Wilcox, prior to conducting the study:

I would like, at some point, to get more feedback from Luis [Tellez] and Maggie [Gallagher] about the ‘boundaries’ around this project, not just costs but also their optimal timelines (for the coalition meeting, the data collection, etc.), and their hopes for what emerges from this project, including the early report we discussed in DC.

What pure interest in “the social science question” involves planning an “early report” with the leading activist against gay marriage, Maggie Gallagher?

Lots of research is as poor quality as Regnerus’s. It’s in combination with the rotten ethics that we see the more serious problem — it’s how the research fits in with his diabolical political plans and his reprehensible moral views. That is, the research was not just bad, it was bad in a purposeful direction. That’s not discernible from a reading of the single, (not really) peer-reviewed article.

Cause and effect

The issue of causality is described in the report as one of methods, but I think it’s really an ethical issue.

Regnerus has been having this both ways from the beginning, and it highlights the challenge of (and for) public intellectuals who speak to multiple audiences. In the original paper he wrote, “I would be remiss to claim causation here.” So that is his cover (and he quotes again here). But in presentations to friendly audiences he is much less guarded. As I reported earlier, in a talk he gave at Catholic University:

He first described in some detail the “standard set of controls” he used to test the relationship between having a father or mother who ever (reportedly) had a same-sex romantic relationship and his many negative outcome variables. And then he proceeded to present bivariate relationships as if they were the results of those tests. He didn’t say they were adjusted [for the controls], but everyone thought the results he showed were controlling for everything. For example, to gasps from the crowd, he revealed that 17 percent of “intact bio family” kids had ever received welfare growing up, compared with 70 percent for those whose mother (reportedly) ever had a same-sex romantic relationship. If you don’t realize that this is mostly just a comparison between stable married-couple families and single-mother families, that might seem like a shockingly large effect.

The causal story at that talk was hammered home in two other ways. First, he presented the results as evidence of a “reduced kinship theory,” under which parents care less about their children the less biologically related they are. Second, he said his “best guess” about why he found worse outcomes for children of women who ever had a lesbian relationship than for those whose fathers ever had a gay relationship was that the former group spent more time with their mothers’ lesbian partners. Both of these descriptions are based on a causal interpretation of his findings.

Anyway, on to the political machinations.

Regnerus lies about Brad Wilcox’s lies

Regnerus complains that Musick brings up the “tired ethical complaint” about Brad Wilcox, who, Regnerus claims, “held an honorific position with the Witherspoon Institute.” And he offers this: “In my interactions with him, he never acted with authority, only advice suggestive of his own opinion.” Regnerus no-doubt thinks he is using a clever legalism, as if Wilcox did not have literal signing authority for dispersing Witherspoon funds and therefore did not offer anything beyond “his own opinion.” But it’s clearly wrong.

Just to be clear how ridiculous this hair-splitting is, here is the email exchange that they no-doubt both now regret (which Musick quoted as well). Regnerus writing, Wilcox answering in bold caps:

Tell me if any of these aren’t correct.

  1. We want to run this project through UT’s PRC. I’m presuming 10% overhead is acceptable to Witherspoon. YES
  2. We want a broad coalition comprising several scholars from across the spectrum of opinions… [goes on to discuss individuals]. YES
  3. We want to “repeat” in some ways the DC consultation with the group outlined in #2. … [details of how the planning document will be crafted] YES
  4. This document would in turn be used to approach several research organizations for the purpose of acquiring bids for the data collection project. YES

Did I understand that correctly?

And per your instruction, I should think of this as a planning grant, with somewhere on par of $30-$40k if needed. YES

Regnerus may now say, indignantly, “Professor Wilcox did not — and does not — speak on behalf of Mr. Tellez,” the Witherspoon president, but he certainly understood Wilcox as speaking for Witherspoon in that exchange. Otherwise, why wouldn’t he ask Tellez these organizational questions directly?

In a 2012 blog post on the now-defunct (and deleted) Family Scholars blog hosted by the Institute for American Values, Wilcox wrote that he never served as an “officer” of Witherspoon. He was, on the Witherspoon website as preserved by the Internet Archive, listed as “director” of the institute’s Program on Marriage, Family, and Democracy from late 2008 to mid-2010. That program still exists on the website, incidentally, but it no longer mentions any director — Wilcox is the only director ever listed in the Internet Archive pages. As of last month, Wilcox’s CV doesn’t mention this position.* (I don’t understand the purpose of an honorific position if you’re not proud of it.)

And then there’s the Wilcox email where he refers to the study as “our dataset.”

Campaign, or coincidence?

Regnerus tells a story of coincidences. For example, Tellez may have (in his words) wanted the research done “before major decisions of the Supreme Court,” but that had nothing to do with Regnerus’s goals, which were to finish his report by January 2012 “for no other reason than I wished to finish it and move on to other projects.” At the time the research was funded, Regnerus says, he did not share Tellez’s political goals. Coincidentally, they both happened to want to project completed in the same time frame. And then, in another coincidence, Regnerus later came around to joining in Tellez’s opinion that same-sex marriage must be stopped. Is this a more plausible story than the simpler one in which Tellez, Wilcox, and Regnerus were all on the same page all along? The evidence for the conspiracy is pretty robust, considering Regnerus, Wilcox, David Blankenhorn, Maggie Gallagher and other anti-gay marriage activists planned the research at a meeting in Washington hosted and paid for by the Heritage Foundation. On the other hand, the evidence for the coincidence is Regnerus’s solemn word. This conspiracy is a theory kind of like evolution is a theory — it’s the only plausible explanation for a known series of events.

In the coincidence story, the survey was delayed, so Regnerus would have to keep working on it beyond January 2012. However, he nevertheless just “decided to give a journal submission a shot” in November 2011 anyway. Not that he was aiming for Tellez’s Supreme Court deadline. Just because. So he “contacted [Social Science Research editor] Professor James Wright to ask if he’d consider reviewing a manuscript on a study like this one,” before the data were even collected. You social scientists out there — have you ever asked a peer-reviewed journal editor if they would consider publishing something “like” what you were working on before you even had the data collected?

In fact, this “give a journal submission a shot” idea came from Wilcox, who in the email mentioned above suggested sending it to SSR because Wright was (the late) “Steve Nock’s good friend” and “also likes Paul Amato,” whom they had secured as a consultant. In the end, Wright would use both Wilcox and Amato as reviewers.

The coincidences Regnerus speaks of also include the meeting he had in August 2011 in Denver with Wilcox, Glenn Stanton from Focus on the Family, and Scott Stanley, after which (he wrote to Tellez at the time), “we feel like we have a decent plan moving forward” for “public/media relations for the NFSS project.” In his response to Mucisk, Regnerus now writes, “Denver was a convenient stop on the way back to Austin from the American Sociological Association annual meeting in Las Vegas, and I took the opportunity to meet socially with a few peers.” That includes Stanton, who “lived about an hour’s drive of where we met.” (I’m not sure why you need a “convenient stop” from Las Vegas to Austin, which is a short nonstop flight.)

See, no campaign. Sure, he also arranged for the study to be shared with “some conservative outlets” before publication, attended a “short function hosted by the Heritage Foundation” about the study just before it was published, and “another such function … at the offices of the Institute for American Values.” But he doesn’t even know, “frankly,” “how such groups came to be apprised of the impending study release.”

Then, after describing, literally, how he colluded with politically-motivated organizations prior to the publication of the study, Regnerus concludes, “This hardly merits the accusation that I ‘colluded with politically-motivated organizations prior to the publication of the study.'”

And oh, sure, on closer inspection (he actually says, “I see now…”) he did use the “media training” document that Heritage provided, which he has falsely testified he “largely ignored,” in his own promotion of the study. In his post on (here), he wrote:

Q: So are gay parents worse than traditional parents?

A: The study is not about parenting per se. There are no doubt excellent gay parents and terrible straight parents. The study is, among other things, about outcome differences between young adults raised in households in which a parent had a same-sex relationship and those raised by their own parents in intact families.

The Heritage talking points (from Musick’s report) included this:

Whether gay parents are worse than traditional parents.The study is not about parenting. There are no doubt excellent gay parents and terrible traditional parents. The study is about outcome difference between young adults raised in a same-sex household and those raised by their own parent in intact families.

Well, he says now, “I very likely did use a few lines” from the document. “So be it.” Nevertheless, “to suggest I received extensive media training — and leaned on it in a comprehensive campaign — is out of touch with my lived reality.” (Who’s a phenomenologist now?) It’s tempting, after reading his response, to assume that whatever Regnerus specifically denies is exactly true.

On the labeling issue

I noticed something new in reviewing material for this post. If you’ve made it this far, bear with me here on this detail.

The infamous Regnerus article was published with the title, “How different are the adult children of parents who have same-sex relationships?” In the article he referred to the adult children who reported that their mother ever had a same-sex romantic relationship as “LM” for “lesbian mother,” along with “GF” for “gay father.” In the rebuttal to his critics, published later in 2012, he acknowledged these were the wrong terms:

Concern about the use of the acronyms LM (lesbian mother) and GF (gay father) in the original study is arguably the most reasonable criticism. In hindsight, I wish I would have labeled LMs and GFs as MLRs and FGRs, that is, respondents who report a maternal (or mother’s) lesbian relationship, and respondents who report a paternal (or father’s) gay relationship. While in the original study’s description of the LM and GF categories I carefully and accurately detailed what respondents fit the LM and GF categories, I recognize that the acronyms LM and GF are prone to conflate sexual orientation, which the NFSS did not measure, with same-sex relationship behavior, which it did measure.

But he insisted this was just a question of confusing terms, not an attempt to actually label these parents according to their sexual orientation. He added:

The original study, indeed the entire data collection effort, was always focused on the respondents’ awareness of parental same-sex relationship behavior rather than their own assessment of parental sexual orientation, which may have differed from how their parent would describe it.

This came up in the Mucisk report, and Regnerus responded:

As noted in Professor Musick’s assessment, the problem of locating an optimal acronym here is something to which I have already confessed … It remains a significant regret. And yet the distinction between a woman’s same-sex relationship (to use Professor Musick’s acronym) and a woman’s “lesbian” relationship (as I assert by using the MLR acronym) is no doubt a narrow one. As ought to be obvious, I use the term “lesbian” as an adjective here, not a noun [emphasis added]. It describes a relationship, not a self-identity.

But did Regnerus really intend to use “lesbian” as an adjective? No, he did not. I know this because, in the email exchange between Social Science Research editor James Wright and Brad Wilcox (in which Wilcox lied by omission and which Wright later misrepresented), we can see the original title of the article Regnerus submitted, which is not the title subsequently published. The original title was, “How different are the adult children of lesbian mothers and gay fathers? Findings from the New Family Structures Study.” Clearly, Regnerus’s original intention was to describe the parents of the people he surveyed as “lesbian mothers” and “gay fathers” — using nouns referring to the people, not adjectives referring to their romantic relationships. It was not a matter of confusion; it was an attempt to create a false impression of the study’s implications.

Promoting Regnerus

In our department, promotion to full professor requires “an exemplary record in research, teaching, and service” which has made the candidate “widely regarded as a scholar.” However, these terms are not defined, and no quantities of research or citations are included. These things are left vague, and much rides on the interpretation of the experts consulted, who are considered the best judges of academic merit. So, what if a professor brings scandal and disrepute to himself and the institution? What if he expresses views that are morally reprehensible? What if he lies about his work, including in his work?

I don’t envy my colleagues in the excellent department of sociology at the University of Texas-Austin (about this case — I do envy them in other ways). Their directory lists 39 professors, only one of whom is disgraceful in those ways. It’s not a simple matter, denying a tenured professor a promotion. It’s a personnel decision governed by laws, and it’s wrapped up in the tenure system, which is important for academic freedom.

In the case of Regnerus I’ve already expressed my opinion.

Honest social scientists do not combine these activities: (1) secret meetings with partisan activist groups to raise money and set political agendas for their research; and, (2) omitting mention of those associations later. If Regnerus, Wilcox, Allen, and Price, had included acknowledgements in their publications that described these associations, then they would be just like anyone else who does research on subjects on which they have expressed opinions publicly: potentially legitimate but subject to closer scrutiny (which should include editors not including people from the same group as reviewers). Failure to disclose this in the publication process is dishonesty.

Based on that — more than based on the morally reprehensible views — I would vote against Regnerus’s promotion. But I am not privy to the process at UT, to their reviews and other materials, and I haven’t been asked for my opinion or advice.

* Is it unethical to take academic activities off our academic CV if they make you look bad? It emerged in one of the gay marriage trials that Brigham Young economist Joseph Price, testifying as an expert against gay marriage, took a grant from the Witherspoon Institute off his CV. In this case I don’t see that Wilcox ever had Witherspoon on his CV, but he was listed as a director on their website.


Filed under In the news

7 facts about the gender gap, for #EqualPayDay

Well, actually, 7 fact-filled posts culled from the many I’ve written on gender inequality, so just call it lots of facts.

1. The gender gap is just one number.

But when you break it out into hundreds of numbers, it’s variations on a theme. This post shows the gender gap by education, kids, marital status, and hours worked. And then by college major. And then I show the distribution of women across 484 occupations, according to the gender gap within each:


2. Occupations matter.

But treating occupational differences as “choices” is at least half ridiculous, so controlling for occupation to get the “real” gender gap is at least half wrong. Of course people choose jobs, but they also take what they can get. So why call it “occupational choice”? For example, one of the most common “choices” to make before “choosing” to be a “retail sales supervisor” is “cashier.” Isn’t choice just a big ball of magical idiosynchronicity?

3. When you assume everyone “chose” their jobs, you miss this woman who was fired for being pregnant.

But the “occupational choice” people don’t notice this, because they’re too busy discussing the “professions” women “choose” and the subjects they major in for their advanced degrees.

4. It’s not just occupations, but hours worked.

And men work more hours (for pay). That’s true, and it’s part of why men earn more (see this paper). But I showed here that, in the occupations with the most overwork (people working 50+ hours) men earn more in almost every one — among those working 50+ hours:

5. Nursing assistants earn less than light truck drivers do.

Because gender. Or maybe there’s some other reason, but I couldn’t find it in a long list of job abilities and working conditions. Among those in these two jobs who: are ages 20-29; are high school graduates only; worked exactly 50+ weeks and 40 hours per week last year; and were never married with no children; the light truck drivers earn 13% more.

6. Don’t just compare then and now.

Way too many people compare “then” and “now” without realizing that gender progress (on many indicators) stalled or slowed two decades ago. For example, as I described here, the percentage of Americans who “prefer a male boss” is lower now (33%) than it was in the 1950s (66%). Wow! But it’s barely lower than it was 20 years ago. Here’s the latest figure from Gallup:


7. It’s complicated.

And, at the risk of jargoning you: intersectional. White women earn more than Black men. But at each educational level Black men earn more.

 Happy Equal Pay Day!


Filed under In the news

Who’s your marriage market?

Richard Florida and the CityLab crew have produced some maps showing the relative size of the single male and female populations in metro areas across the country. They run the maps by age group — here’s the one for all single men and women ages 18-64:

180278ee4It mostly shows larger single female surpluses in the South and Northeast, and male surpluses in the West and upper Midwest.

The maps are interesting, but marriage markets aren’t as simple as gender. For example, among White, Black, and Hispanic newlyweds, 87% married someone in the same race/ethnic group, and 77% married someone on the same side of the BA/no-BA education divide. (I previously showed some figures on the relative number of “marriageable” Black and White men, by education, here.)

Just to underscore that point, here are the match rates in more detail. To make this I counted the matches by race/ethnicity (Black, White, Hispanic), education (BA/no-BA), and age (within 5 years) of people who were married in the previous 12 months, in the American Community Survey 2010-2012 (from

Here are the match rates, broken down by sex and race/ethnicity:

who they married 2010-2012.xlsxThis shows:

  • Altogether, half the newlyweds match their spouses on all three characteristics, and Whites are most likely to match.
  • Blacks are least likely to match on age.
  • Black women are more likely to match on race than Black men.
  • Hispanics are most likely to match on education (mostly without BAs).

Of course, lots of people don’t match on these traits — maybe even especially those adventurous types who pick up and move when they see a map like this. But whether you’re a matcher or not, before you plan your marriage-seeking move you need to know what you’re looking for (and what’s looking for you).


Filed under Uncategorized

Research on teen crashes confirms that reporters selling books on phone risks hype phone risks

Using phones while driving is dangerous and should stop. But the focus on this issue distracts us from other dangers in driving (which have — you’d never believe from the news — declined rapidly in recent decades). And it distracts us from the broader danger of relying on motor vehicle transportation.

I dwell on this subject because it offers lessons beyond its substantive importance (see all the posts under the texting tag). Today’s lesson is about conflicts of interest in the news media.

The AAA Foundation for Traffic Safety published a study of about 1,700 moderate or severe car crashes in which people ages 16-19 were driving. To identify possible causes of the crashes, they used cameras and motion sensors in the cars, and analyzed the seconds before each crash. The headline result probably should have been that 79% of the crashes occurred when teens were driving too fast. But that’s apparently not news, so the AAAFTS and all the news media reporting the story focused on the fact that 59% of the crashes showed distraction as a likely cause.

The report website highlighted the data on distractions, and that’s reasonable. One of their findings is that distractions in their survey account for a greater proportion of accidents than are reported officially — something we’ve assumed but have had trouble establishing empirically. So that’s useful. They used this graphic:


By this accounting, phones were involved in 12% of crashes, second only to interacting with passengers. But this is an artifact of the way the categories are binned. And a lot of smaller categories are left out of the figure, such as eating and drinking (2%), operating vehicle controls (3%), looking at another vehicle (4%), or smoking-related distractions (1%). (Note also that one crash can have multiple related distractions, but they don’t report the overlaps so you can’t do anything about them.)

So I redid the categories. I don’t see why eating and drinking should be a separate category from grooming, or why singing/dancing should be separate from adjusting the radio. So I made a new category called “physically doing something besides driving,” which includes eating or drinking, using an electronic device (besides phone), grooming, reaching for an object, smoking-related activity, operating vehicle controls, and singing/dancing to music. Also, for some reason their figure lists “looking at something outside the vehicle” but only includes “attending to unknown outside vehicle” in that category. I added two other types of distraction to that category, “attending to another vehicle” and “attending to person outside” — bumping up the outside distraction substantially.

Here’s my new version of their figure based on the same data (from table 13 in the report). It’s more comprehensive but uses fewer categories:


Now cellphones are fourth. So that’s a lesson about using arbitrary category collapsing and then ranking the categories. (This happens all the time with occupations, for example, where people say, “The top X occupations…” but the occupations reflect different levels of granularity.)

Anyway, back to cellphones

The New York Times reporter Matt Richtel won a Pulitzer Prize for his reporting on distracted driving. And he published a book — A Deadly Wandering — that tells the tragic story of a driver who killed someone while he was texting. Unfortunately, he is prone to hyping the problem of texting, which his audience is unfortunately prone to fixating on. I previously pointed out that, on the website promoting his book, his publisher uses an extremely wrong statistic, claiming that texting “continues to claim 11 teen lives per day.” He has mentioned this statistic (or its variant, that texting kills more teens than drunk driving) on Twitter, and also in media appearances. I pointed out that this number is more than all the teens killed in motor vehicle accidents, so it’s obviously baloney. I emailed Richtel about this, and he told me he would “get it fixed.” I emailed the publisher. I emailed the Times. I emailed the Diane Rehm show. No one changed anything. Cellphone crashes are like child abuse: people will believe any statistic about how bad it is and attack anyone who’s skeptical.

Of course I don’t want to minimize the problem of distracted driving, and there’s nothing wrong with telling people it’s dangerous. And it’s not my area of expertise. So I’ve only given the issue a few hours. But playing into a public hysteria about a very narrow, behaviorally-driven problem, rather than exposing the systemic problem that it reflects, is not good.

And now that Richtel is selling a book about texting, he’s got a conflict of interest — if he hypes the problem in his NYT stories, he makes more money. So here’s the NYT headline on his story:


And this is his lead paragraph:

Memo to parents: Distracted driving by teenagers is riskier than previously thought, particularly when it comes to multitasking with a cellphone.

Again, it is true the report finds cellphone distraction causes more accidents than police reports have shown — so this is not irrelevant — though, of course, even with the new accounting they still cause orders of magnitude less than Richtel’s own promotional site claims. But mentioning phones in the headline sets the NYT apart from most of the coverage of this report:

  • Washington Post: “AAA: 58 percent of teens involved in traffic crashes are distracted”
  • ABC News: “Distractions a Problem for Teen Drivers, AAA Study Finds”
  • Houston Chronicle: “Distraction a factor in 6 in 10 teen driver crashes”
  • Chicago Tribune: “Distracted driving a key contributor to teen crashes, study shows”

On my first page of Google News searches, only the LA Times also mentioned phones: “Teen drivers distracted by cellphones, talking in most crashes.”

Who cares?

Some people who are tired of me complaining about this think you can’t have too much hype about safe driving, so who cares? But the distraction matters. The evidence that phones are a fundamental cause — a social cause — of accidents and deaths is very weak, although they are certainly the proximate cause in many cases. But we don’t have randomized controlled trials to test the effects of phones. I suspect the people crashing while futzing with their phones are mostly the same people who would be crashing for some other reason if cellphones didn’t exist.

When I look at the video compilation the AAA put out to accompany their report — which mostly shows teens crashing while using their phones — I am struck by what terrible drivers they are. They look down for three seconds and drive straight off the road without noticing. In contrast, I routinely see people driving on the freeway completely absorbed in their phones — driving obnoxiously slowly but using their peripheral vision to keep going straight. They are at grave risk of an accident if something crosses their path or traffic stops, but they’re not veering all over the road. Their slow speed probably mitigates their risk of crashing. I AM NOT RECOMMENDING THIS, I’m just saying: bad drivers cause accidents, and if you give them a phone they’ll use it to cause an accident.

Did you know teen driving fatalities have fallen by more than half in the last decade? (During that time incidentally, teen suicides have risen 45%.) Did you know that, from 1994 to 2011, mobile phone subscriptions increased more than 1200% while the number of traffic fatalities per mile driven fell 36% (and property-damage-only accidents per mile fell 31%)? Don’t count on Matt Richtel to tell you about this.

And yet, of course, thousands of people die in car accidents every year in the U.S. — at rates higher than the vast majority of other rich countries. But as long as people drive, there will be bad drivers. If we really cared, we would replace individual cars with mass transit (or self-driving cars) — putting transport in the hands of computers and professionals. Nothing’s perfect, of course, but preventing car accidents isn’t rocket science, and blaming a systemic problem on the individual behavior of predictably error-prone drivers doesn’t seem likely to help.


Filed under In the news

Vox interview on the Moynihan chilling effect

Jenée Desmond-Harris from interviewed me about the Moynihan backlash post. The piece is here. In it she links to this blog, but not to the specific post. If you’re looking for that, it’s here.

Leave a comment

Filed under Me @ work

Getting serious about promoting marriage to end poverty

This expands on some practice-what-you-preach criticism of conservative marriage promotion, with some numbers. I’m not endorsing the approach described here — I’m saying marriage promoters should adopt this if they are serious about promoting marriage to reduce poverty.

At Demos, Matt Bruenig wrote:

After rigging the institutions to capture the majority of the national income and basically all of the national wealth, segregating themselves residentially, intermarrying almost solely in their rich enclaves, and even sealing off their schools from being accessed by the unwashed masses, these rich social conservatives turn around and implore others to marry people that they wouldn’t touch with a ten foot pole, people they can’t even bring themselves to make even the most minimal of community with.

In response, Sandy Darity tweeted: “I proposed that a marriage antipoverty strategy should have rich white men marry poor black women.” I don’t want to put the onus for ending poverty just on pro-marriage pundits. Instead — as Darity suggests — we should think in terms of broader policy.

Whose norms?

Marriage promotion is mostly about convincing (educating, coaching) poor people to marry other poor people. That follows from the “culture matters” perspective on marriage decline advocated by some social scientists as an explanation for declining marriage rates. For example, in a New Yorker profile of Orlando Patterson, Kelefa Sanneh writes:

[William Julius] Wilson argued that declining professional prospects made some black men less marriageable. Patterson thinks that declining marriage rates had more to do with the increased availability of contraception and abortion, which eroded cultural norms that had once compelled men to marry the women they impregnated.

Whether the proximate cause is men’s reduced economic prospects or changing norms, the fact is that if poor people changed their attitudes (norms, culture) about marriage — if they put more priority on the importance of marriage and worried less about the economic qualities of the match — there would be more marriage and, they say they believe, less poverty, inequality, violence, and abuse).

An obvious problem with this whole enterprise is that the marriage boosters assume the next marriage they generate through marriage promotion will be as economically beneficial to the participants as the average existing marriage observed in the population. But if one of the reasons for non-marriage is poor economic status, then it follows that the next marriage generated will on average be much less beneficial economically than the average marriage (I expanded on this here). So the plan to reduce poverty by promoting marriage among the poor is running uphill. Or, it would be running uphill if it was running at all, but of course (ridiculous research shenanigans notwithstanding) their billion dollars spent has yet to generate a marriage, so this is really all very generous speculation.

If they really wanted to change “the culture”

For several decades, marriage promoters have been complaining that “the culture” isn’t pro-marriage enough. The latest version of this, from David Blankenhorn and colleagues, seeks to “restore a marriage culture among the less privileged.” But, although it’s true that poor people (especially poor Black people) have seen a faster drop in marriage rates, that’s not where the biggest anti-poverty gains are to be had. If you really want marriage to reduce poverty, and you really think policy can change “the culture” to make more marriages, then what you really need is (as Darity said) some rich (mostly) White men to marry some poor (disproportionately) Black women.

Why not? Is it really more far-fetched to imagine you could change rich White men’s attitudes toward poor Black women than it is to suppose you could “restore a marriage culture” among the poor? Why? Maybe one reason policies to increase marriage among the poor haven’t work is because the economic benefits aren’t great enough. If you were the kind of person that goes in for this sort of policy (which, again I am not), you’d have to assume poor people would be more receptive to the idea of marrying rich people — that’s one important premise of Wilson and Patterson’s perspective. So the problem is rich people don’t want to marry them.

How difficult can this be? Just to put some numbers to the idea, I did the following simple exercise. Take all the poor single mothers — specifically, non-married women living in their own households with their own children, with family incomes that put them below the federal poverty line — and match them up with rich single men.

How many rich single men do you need? With this definition, I get 3.5 million poor single mothers. I started with the richest single man, and went down the income ladder till I had enough to solve the single-mother poverty problem. It turns out you only have to go down to $80,000 per year in income. Here’s the matching, with the race/ethnicity of the two groups shown:

rich men marry poor women.xlsxIf the problem is that poor women are too economically choosy to marry the poor men in their lives, then we could easily lift these 3.5 million single mothers — and the 7.1 million children in their families — out of poverty simply by changing the anti-marriage views of these selfish, rich, single men. Of course, we’d have to reduce racist attitudes also, but not entirely — only a third of the non-Black rich single men would need to open their minds to the possibility of marrying a Black woman. You would have to be creative with the incentives for these men, including consciousness-raising and parenting classes, as well as, for example, Starbucks gift cards and subscriptions to the Economist.

Now, no one thinks you can socially engineer — through shame or tax incentives — the marital behavior of entire populations, so this strategy couldn’t be expected to completely eliminate the problem of single mothers and their children living in poverty. But it couldn’t be less effective than the marriage promoters have achieved with the last billion dollars they spent.


Filed under Uncategorized