Monday Miscellany: Allomancy, Bargh, Cherrypicking, Divorce

1. This week in results I can barely believe, gentrification appears to decrease the chances that longtime residents will leave. I’m slightly less skeptical given that the researcher himself didn’t anticipate getting these results–he expected the opposite. (h/t Scott)

Lance Freeman, the director of the Urban Planning program at Columbia University, says that’s what he believed was happening, too. He launched a study, first in Harlem and then nationally, calculating how many people were pushed out of their homes when wealthy people moved in.

“My intuition would be that people were being displaced,” Freeman explains, “so they’re going to be moving more quickly. I was really aiming to quantify how much displacement was occurring.”

Except that’s not what he found.

“To my surprise,” Freeman says, “it seemed to suggest that people in neighborhoods classified as gentrifying were moving less frequently.”

Freeman’s work found that low-income residents were no more likely to move out of their homes when a neighborhood gentrifies than when it doesn’t.

2. I used to think Allomancy was my favorite system of magic. I was so, so wrong. (Also, credit to Scott)

3. Mental health issues in hospitals and emergency rooms are a growing problem.

Nationally, more than 6.4 million visits to emergency rooms in 2010, or about 5 percent of total visits, involved patients whose primary diagnosis was a mental health condition or substance abuse. That is up 28 percent from just four years earlier, according to the latest figures available from the Agency for Healthcare Research and Quality in Rockville, Md.

By one federal estimate, spending by general hospitals to care for these patients is expected to nearly double to $38.5 billion in 2014, from $20.3 billion in 2003.

The problem has been building for decades as mental health systems have been largely decentralized, pushing oversight and responsibility for psychiatric care into overwhelmed communities and, often, to hospitals, like WakeMed.

In North Carolina, the problem is becoming particularly acute. A recent study said that the number of mental patients entering emergency rooms in the state was double the nation’s average in 2010.

More than 10 years after overhauling its own state mental health system, North Carolina is grappling with the consequences of a lost number of beds and a reduction in funding amid a growing outcry that the state’s mentally ill need more help.

4. So, uh, how does being a hitman work? Oh good, someone did quantitative research on that.

The results of their detailed search of British cases that matched this description in the period between 1974 and 2013 only turned up 27 contracted hits or attempted hits “committed by a total of 36 hitmen” (there was only a single “hitwoman”), but the researchers used the sample to tease out the details and profiles of typical killers-for-hire.

The main thrust of the paper, which will be published in the Howard Journal of Criminal Justice, is that hitmen do not operate with the drama, professionalism, or glamour that mob films and spy novels afford them. In actuality, the majority of killers select jejune settings for their crimes, have occasionally bumbling performances, and are often hired by contractors with lame motivations.

5. I once had a professor who offhandedly said “hah, he probably just cherrypicked whichever results looked pretty” when a student expressed shock at the results of some priming research from John Bargh.

(The example of priming I hear most often cited in popular conversation is that giving a stranger a cup of hot coffee, rather than iced, will make them think more positively of you.)

Not five minutes later, the professor praised Bargh for being the leading researcher in this growing new field. The ridiculousness of being known for squashing data into conclusions and being one of the most well known social psychologists didn’t seem to register.

That story makes this article investigating Bargh’s work even more compelling. (And because aaaah, what, Donellan has a blog?! </psychnerdery> I’m going to particularly note this in-blog link on Bargh as well.)

ETA: as I was scheduling this post, there was an update. It’s like a soap opera!*

6. Hunh. Discussing relationship movies decreased divorce rates. In retrospect, I suppose this makes sense. Instead of some awkward feeling counseling session, the movie couple can serve as a proxy. As long as everyone pretends they’re talking about Jack and Rose you can debate issues like differing family backgrounds or drawing one another naked. (link via Julia)

*Really, who can resist making bad shower puns? Not I. 


  1. HFM says

    For the gentrification bit, I wonder how much of that effect is due to rent control? In my city, anyone who’s been fortunate enough to get a rent-controlled place (because they’ve lived here for a while, they got lucky, or both) doesn’t dare move, because they couldn’t afford anything nearly as nice if they had to pay market rates.

    • M can help you with that. says

      HFM @ 1 —

      I haven’t read the study yet, so I don’t know if rent control was, um, controlled for; it wouldn’t surprise me if it accounted for some of the effect. In other places high rates of home ownership pre-gentrification might also contribute to a smaller displacement effect. (I’m thinking in particular of a neighborhood near me; solidly lower-middle-income, but mostly families that owned their own homes. The local community association basically started a pro-gentrification campaign; bringing in higher-income owners led to better city maintenance of streets, etc., and boosted the home equity of the people who already lived there. I suspect that this isn’t something that would be a dynamic in most gentrifying areas, though.)

  2. says

    There was a great discussion about the relationship movie study on Reddit’s r/science. I think the best comment was the one by djimbob.

    If I’m understanding it right, he was pointing out that the group assignment wasn’t random – they asked people whether they wanted to participate in marital therapy, and the people who refused therapy or couldn’t find the time to make it there formed the no treatment control group. But it seems possible that the sort of people who refuse therapy are generally disagreeable or not that committed to their marriage, and that the sort of people who are too busy to make a therapy appointment are too busy to spend time with their spouse. And in fact, the no treatment control group had a three-year divorce rate of 24%. There aren’t obvious clear statistics on three-year divorce rates in the general population, but if you extrapolate back from other statistics it looks like the average three-year divorce rate is 12%. But 12% is what the three treatment groups – including the watching-a-movie treatment group – got. So it could be that none of the treatments had any effect, but the no treatment control group was selected for particularly high likelihood of divorce. And this made the three treatment groups all look superior.

    • Robert B. says

      They let people pick what group they would be in? Geez, the things researchers will do to get their N up…

      But wait a minute. If people more likely to get divorced were getting sorted into the control group, then they were getting sorted out of the experimental group. Assuming the four groups were of equal size, and the general population is getting divorced at 12% per three years, selecting in enough likely-divorcers to get the control group up to 24% would have lowered the expected divorce rate in the other groups, according to the null hypothesis, to 8%. Every option in this study, including no treatment, is now worse than the general population.

      At this point we would have to conclude that being in the study increased the divorce rate by a quarter… or that your 12% over three years figure for the general population is off. How did you get it?

Leave a Reply to M can help you with that. Cancel reply

Your email address will not be published. Required fields are marked *