When Study Results Have Morally Uncomfortable Implications

February 5, 2009

Now and then I’ll read a news story that discusses the results of a new study on some topic or other, usually in the areas of psychology, sociology, or medicine. Usually the findings are interesting and the news reports of those findings discuss the implcations of the news (such as that a new drug may lead to better cancer therapies). But sometimes the results of the study have personally, socially, or morally uncomfortable implications.

  Example 1: Today’s   Albuquerque Journal (Feb. 5, 2009) newspaper reported the New Mexico Domestic Violence Homicide Review team’s annual report from 2005. It dealt with issues about alcohol use, domestic violence, stalking, and homicide. The third item jumped out at me: "The victim and the alleged killer were separated or separating when the killing took place in 71 percent of the deaths." That is, nearly three-quarters of the domestic violence related homicides happened during or after a separation.

  The Uncomfortable Implication : Statistically, people who are abused by a spouse or partner are actually much safer staying in that situation than getting out of it. If you’re getting beaten and smacked around, leaving your abuser is more likely to end in your death than staying with him or her and putting up with the abuse.

  Example 2: Recent studies published in peer-reviewed medical journals found that fat can be spread from one person to another. One CBS News story (http://www.cbsnews.com/stories/2007/07/25/health/webmd/main3097001.shtml) reported:

The findings, published in

  The New England Journal of Medicine

, show that obesity is "socially contagious." That means that people tend to follow suit when their friends and family become obese or lose weight to ditch obesity. "We find that a person’s chances of becoming obese increase by 57% if they have a friend who becomes obese, 40% if they have a sibling who becomes obese, and 37% if a spouse becomes obese," say researchers Nicholas Christakis, MD, PhD, and James Fowler, PhD.

  The Uncomfortable Implication : If you want to lose weight (or not gain any more), stay away from fat people, because you may get fatter just being around them.

Both of these stories were reported in the mainstream news media, but the journalists were careful not to discuss any implications of the studies and findings. Why? As a society, of course,  we don’t want people to remain in domestic abuse situations, and we don’t want overweight people shunned by others because they are afraid of "catching fat." Of course, there is a danger in simply pulling statistics and factoids out of context. But the above statistics and results are accurate (as far as I know) and the logical implications are accurate (as far as I know). So what should we do, as journalists or as a society, when scientific research produces information whose implications we find uncomfortable?

Should these sorts of findings not be reported, for fear of people acting on their implications?
Should the findings be reported but carefully phrased to avoid explicitly discussing the implications?

What do you think?



#1 Kevin (Guest) on Friday February 06, 2009 at 6:15am

I agree with your assessment of Example 1, and the logical implication you point out. Furthermore, I think this is exactly the sort of implication of this study we _should_ be talking about. If we, as a society, want the “correct outcome” to be woman leaves (abusive) man, then we must step in and do something to correctly weight the scenario. As we can see from this study, absent outside influence, the correct response (empirically determined) would seem to currently be woman stays with man.

Per Example 2, I think that staying away from fat people or “catching fat” is a false choice. I would argue that this calls for more social education about the problems of peer pressure and the instinct of humans to follow suit without even realizing it. Concluding that we should socially ostracize or avoid overweight people is a bit like cutting your head off to get rid of a zit. It causes far more problems than it alleviates. Your conclusion here might be equivalent to concluding based on Stanley Milgram’s famous obedience studies that we should avoid everyone with an authoritative manner or an assertive personality, lest we fall victim to their domination. [The more interesting implication of the Milgram study is the challenge it presented to the finding at the Nuremburg Trials that “just following orders” is really no excuse. Milgram turned assumptions about human response to authority on their heads, showing time and again that large majorities of people would comply with his “teachers” despite expert predictions that only 1-2% would.]

#2 Sharon (Guest) on Friday February 06, 2009 at 6:42am

I am reminded of comments in The Honest Broker by Roger Pielke. Scientists serve society best by presenting the information and expanding the options and clarifying the scope of choice for the decision maker. We can only view these options as “best” or “worst” in context of the social and political views of the time. So, there is a role for others here besides research scientists - the implications of the results make this obvious. This information is important because it shows that one of the courses of action to respond to a situation is not working well. This information can provide the impetus to fix it. In that respect, it’s so valuable.

#3 Emily (Guest) on Friday February 06, 2009 at 9:07am

I see that your advice to abused women is empirically grounded, however, statistics can also be used to take action.  If there was a safe place for women to go; if abusers could be locked up and taught to change their behavior; if women could carry firearms and be trained in self-defense… all of these reactions use the same data. 

Besides, how many women who leave their abusive spouses are NOT murdered?

Implications are not facts.  Otherwise, I should be a Christian because statistically, Christians are happier and healthier than non-Christians.

#4 Ben Radford on Friday February 06, 2009 at 10:35am

Kevin, I agree that the study in Example 1 calls for more social education about the problems of peer pressure and the instinct of humans to follow suit. Perhaps journalists should take that advice, though they may feel uncomfortable in going beyond the study like that.

Emily, I’m not sure that the statistic necessarily says anything about whether women have a safe place to go (virtually all cities and towns have a “safe shelter” for victims of domestic abuse—even here in New Mexico), or the other notes. The statistic, from how I read it, takes those factors into account: with or without safe shelters, with or without women carrying guns, etc, victims are much more likely to be killed if they separate from their abuser.

I agree that, to give context, the journalist might have included the number of women (and men—men are also victims of domestic violence) who are NOT killed by those who leave, that would be good information to add, and it might lessen the impact the Uncomfortable Implication.

#5 AtheistRickB on Friday February 06, 2009 at 4:19pm

I don’t think your implication in example 1 is correct or else I am misreading the example.  The text says that 71% of the homicides occurred when victims were separated/separating.  So for example, if there were 10,000 cases of abusive relationships and 100 of those ended in homicide, 71 of those 100 would have occurred during the period they were separating.  That means that 9,929 abusive couples could have separated without getting killed during separation (although an additional 29 would be killed at some other time). 

Hence, the statement that “Statistically, people who are abused by a spouse or partner are actually much safer staying in that situation than getting out of it.” does not appear to be correct.  I know I just made up the numbers above but the point is that until one knows the full extent of what the statistics are trying to tell you, you need to be very careful about drawing conclusions.  Now, if there was only 100 abusive couples and all of them decided to separate and 71 got killed, then what you say would be correct.  But I did not see the total number of abusive couples separating listed anywhere in the note you posted.

Am I missing something here?

#6 Ben Radford on Friday February 06, 2009 at 7:00pm

Rick B, hm. The news piece did not mention the total number of couples, I’d probably have to check the original report. I agree that the statistic doesn’t suggest that 71% of abusive relationships end in homicide, just that the status of the victims at the time of death is separated (or separating) in 71% of the homicide cases. At the very least it shows how a statistic like that can be read in different ways!

#7 Emily (Guest) on Friday February 06, 2009 at 9:08pm

I guess I wasn’t clear.  You can interpret the data to shrug and say, “Women would rather be beaten than dead, right?” or you can look at the data and say these women need a safe place and combat training so that they don’t get dead when they leave their spouse.  You can act or do nothing.  I say acting is better.  And actions are based on the statistics and a desire to improve them.

#8 Ben Radford on Saturday February 07, 2009 at 10:27am

Emily, I agree with you. Everyone agrees that abuse victims need a safe place, support, protection, etc. and no one in their right mind would encourage any person to stay with an abuser. I don’t see it as a choice between acting or doing nothing—of course action is needed, but the question I was bringing up is, “What happens when the research suggests that acting (i.e. separating from the abuser) may be more dangerous than staying?”

To me it’s an interesting moral dilemma, and it’s one that has occurred in real life: Friends encourage a battered woman to leave her husband, who then ends up killing her because of the separation. Even though the friends know they did the right thing by trying to get her out of the abuse, they may be filled with regret that, if they hadn’t done that, she would still be in the abusive situation—but at least she would be alive. Of course the murder is completely the abuser’s responsibility, but that doesn’t bring her back.

The issue becomes putting someone at much greater temporary risk in order to get them out of a less dangerous long-term, lower-risk situation.

#9 Larry Clapp (Guest) on Sunday February 08, 2009 at 8:31am


Agree it’s an interesting question.  Also agree with Rick that your Example 1 is bogus.  The study (but it’s not really a *study*, is it, just the results of a single state)—anyway, the results don’t say “If you leave, you stand a 71% chance of getting killed”, they say “If you get killed, there’s a 71% chance that you left.”

Two variables: survive or murdered, leave or stay.

You have “murdered & stay” and “murdered & leave”.  You need the other two: “survive & stay” vs “survive & leave”.

Just a guess, but suspect that numbers for “survive & leave” are higher than all the others.

#10 Ben Radford on Sunday February 08, 2009 at 10:58am


I agree… to really make sense of this factoid we’d need the other two conditions: “survive & stay” vs “survive & leave”. I also suspect that numbers for “survive & leave” are higher than all the others. Good comments!

#11 Annec (Guest) on Monday February 09, 2009 at 10:25am

( When study results are uncomfortable)

Not since the days, when in Egypt,
That the head of the Sphinx was heightened
From the flood,
Had such asperity betided her.

But from these times had come the daughters of Aaron.

About Egypt: The Gospel of Bartholomew… free with SSAE

After studying a very old Bible, copyright MCMXLVIII, the preface explains; ‘Careful investigation of the results of archaeological research and a more thorough examination of the Bible itself, have made it clear that some of what once seemed false is true while some of those things that once appeared true must be rejected.’ Now those that have studied and have ministered these interpretations over the years, and do not understand, are the innocent ones. It has been 150 years since the publication of ‘On the Origin of Species’ and now we have a new recording, ‘An Analog from Adam to Nero’

Annec has the power to vision a new history, but how can this be? Do we not presently have one? We have any number of histories, recordings of various weights and virtue; biased scribbles of past events. Annec shapes the images that loom behind us by clearing away the fog of forgotten antiquity. Denial and refutation is not practical, we guarantee, the earth is not flat.

#12 Sunny-Avandia lawyer (Guest) on Friday February 13, 2009 at 1:46am

Thanks again for your wonderful information.

Commenting is not available in this weblog entry.