annals of very large effect sizes

From Nicholas Kristof’s most recent column in the NYT:

Researchers once showed people sketches of a white man with a knife confronting an unarmed black man in the subway. In one version of the experiment, 59 percent of research subjects later reported that it had been the black man who held the knife.

Does anyone know the study on which this is based?

causes of differences

The other day I read a useful paper about causation, by Kenneth Waters, who is now at the University of Calgary. At least I found it useful. Every so often I’ll make a dive into philosophical literature on causation, and the recurrent problem that I find is that philosophers are generally preoccupied with causation for single events (examples of balls hitting other balls and convoluted assassination schemes) or for big generalizations about classes of events. Meanwhile, sociologists regularly form their causal questions as about differences between groups of persons, as in wanting to explain why whites score higher on standardized tests than blacks or why women live longer than men.

Waters explicitly tries to formulation insights about causation in population terms. As he puts it:

Focusing attention on singleton situations about a single lighting of a match, a single breaking of a vase, or the single catastrophic dropping of a boulder obscures important features of causation. Much light could be shed on causal reasoning by shifting attention to causes in populations.

To get there, he draws a distinction between potential difference makers and actual difference makers. The key move here is the study of actual differences, for which you have to be talking about the causes of multiple outcomes, that is, a population. While potential difference makers are the larger body of causes of an outcome, actual difference makers are the ones for which actual variation in the population in question explains actual variation in the outcome.

tummyache tips

If you find yourself sending the following e-mail to students:

[Student A], [Student B]:

I woke up this morning feeling really sick to my stomach. I thought it was something that would pass, but it hasn't passed yet. So, [A], I'm sorry for being so last minute about this, but the pain is so acute at the moment that I don't think I can walk up to campus to meet. We'll need to reschedule (and [B], if you don't hear more from me, us too). More when this subsides enough that I can think clearer.


You might consider the possibility that it is appendicitis, and that perhaps rather than staring at your monitor thinking “When is this going to pass? I’ve got writing to do!,” you should Uber yourself to the emergency room, stat. Because getting to the emergency room is just the start of a long waiting process to see the doctor, and you can’t get any pain medication before that happens, so by the time that happens you will have exhausted your mental inventory of theistic-like entities from known religions that you can beg and bargain with, should they exist, to make this horribly painful queasy painful pain in your stomach go away? Continue reading “tummyache tips”

academic name changes

OW’s post about her problems from not changing her name reminds me of a question that came up last week in Bloomington over lunch. The question concerns women scientists — well, let’s start out by restricting attention to women sociologists — who are placed in tenure-line academic jobs. Some women have publication records in which they publish under one surname, and then later on–say, no earlier than after completion of their PhD–switch and publish under an entirely different surname. Leave the matter of name hyphenation out of this: I’m talking about case where if my CV went from publications as “Jeremy Freese” to publications as, say, “Jeremy McDonald”.

The question: when this happens, how much of the time is it because the woman has gotten married, and how much of the time is it because the woman has gotten divorced? This is in principle an empicial question that has an answer, but I don’t know what the answer is, so I invite your speculation. (Or, if this actually a question with a known answer out there somewhere and somebody wants to offer that, all the better.)

name ghost

One way or another it is looking like it will cost me several hundred dollars and significant aggravation to deal with the fallout of US patriarchy. Back when I was married in 1970, the women’s movement was just kicking in and a summer employer insisted that they could not (would not) pay me unless I signed a form changing to my married name on my social security record. I never got a new card, however, and since that time, the only name I’ve used is my birth name.To do this, in the 1970s I had several times to verbally lie to self-appointed local government monitors of women’s names (marital status was never a question on the written document one was signing) who were insisting that married women must use their husband’s surnames on things like drivers licenses and employment records. Sometimes the courts upheld the patriarchists, sometimes the women. All this dust gradually settled around 1980 and since then married women have been left alone and allowed to use birth names in peace.(All you young-‘uns who are going about changing names willy-nilly for trivial reasons like marriage just make us older women sigh, given all the grief we incurred to avoid it.)

Because many people do change names at marriage, it is very easy to do so. You just drop by your local identity office with marriage papers and poof your name changes. This does not apply, however, if you are caught in the warp of the 1970s. If the SSA persuades itself that the name they have for you is your “legal name,” you must prove that there has been a legal name change. If you are a married women using your birth name you do not, of course, have such a court order, because you never changed your name. You are just dealing with the fallout of strong arm patriarchal bullying from the 1970s that gives many married women from that era an inconsistent set of names.

SSA knows who I am. I have a comprehensive identity record. They know my birth name, they can see my lifetime payroll records, they have my marriage certificate. They know what happened. There is no dispute about the facts. But they claim to be incapable of correcting their records to match reality without a court order. They say this is part of the heightened scrutiny on identity with e-verify. There are activists pointing out that this system disproportionately affects women. My lawyer says I should not have to pay her to do this for me, and I’m going to try one  more time on my own before handing it over to her.

I’m pretty mad but if I have to I can pay the money to get this straightened out. If I have to, I’ll get the court order. But as my friends say, “what are poor women supposed to do?”

transparency as a matter of research ethics

Apparently APSA has actually made a step toward open science as part of their ethical guidelines. According to a recent paper in Science:

The American Political Science Association in 2012 adopted guidelines that made it an ethical obligation for researchers to “facilitate the evaluation of their evidence-based knowledge claims through data access, production transparency, and analytic transparency.”

My understanding is that the ASA ethical guidelines are still stuck on the idea that people should get around to sharing data after they’ve finished all the articles they might want to write from a paper, which as a practical matter often means “never.”

Any plans afoot to modify the ASA code of ethics?