Friday, October 26, 2012

Women Vote Their Hormones: The Study Itself

This relates to the recent fuss at which resulted in the withdrawal of a post about the study I will discuss here, "The Fluctuating Female Vote:  Politics, Religion and the Ovulatory Cycle" by Kristina  M. Durante,  Ashley R. Arsena and Vladas Griskevicius.

I obtained the manuscript from Durante's website.  It may not be in its final form.

The justification of this study is pretty tough to understand for someone who is not steeped in the holy juices of evolutionary psychology.  For instance, the authors study both political values AND religiosity because, to quote from the study*:

Building on the idea that reproductive goals might drive political and religious attitudes (Kurzban, Dukes, & Weeden, 2010; Li, Cohen, Weeden, & Kenrick, 2009; Weeden, Cohen, & Kenrick, 2008), we examine whether hormonal fluctuations associated with fertility influence women’s politics, religion, and voting.

Reproductive goals.  How do these enter into the scenarios?

Political ideology is believed to serve deeper functions (e.g., Jost et al., 2003; Graham, Haidt, & Nosek, 2009). Several theorists, for instance, have proposed that political and religious ideology are related to reproductive goals, arguing that an individual’s current mating strategy drives that person’s political and religious attitudes (Kurzban, Dukes, & Weeden, 2010; Li, Cohen, Weeden, & Kenrick, 2009; Weeden, Cohen, & Kenrick, 2008). Specifically, lower levels of religiosity and more liberal political attitudes may facilitate a short-term mating strategy associated with more permissive and promiscuous sexual behaviors.
Consistent with this idea, studies find that mating concerns are a strong predictor of religious attendance (Weeden et al., 2008) and social political attitudes on legalizing marijuana (Kurzban et al., 2010). Experimental evidence also finds that the local mating ecology influences women’s religiosity, with the presence of many desirable single females leading women to become more religious (Li et al., 2009). Because a glut of single females might pose a threat to a woman’s own romantic relationship, women are believed to become more religious and espouse the sanctity of commitment to protect their relationships. Taken together, these findings suggest that religiosity and political attitudes are somewhat flexible, with people adjusting their orientations to serve their current reproductive goals.

Fascinating stuff!  I still don't quite get what "current mating concerns" might be.  Is this the interpretation given to the statistical correlation between more permissive sexual norms and voting liberal?  That lower "religiosity and more liberal political attitudes may facilitate a short-term mating strategy associated with more permissive and promiscuous sexual behavior? "  That's a weird twist on the usual take on these matters which would probably reverse the argument.  Besides, the question is surely empirical.  Find out if conservatives are more or less likely to carry out adulterous affairs etcetera than liberals.  What the study seems to believe is that people become less religious and more liberal when they want a one-night stand or two.

I haven't read the Li et al. article from 2009, about the glut of many desirable single females leading women to become more religious.  It's supposedly experimental, but obviously the researchers couldn't place such "single females" into some area to wait until they could measure the religiosity of other women's views there.  Some sort of an experiment with undergraduates, I presume.

You probably get the point.  Everything, my dears, is about mating strategies, and all those strategies are deeply hardwired in our tiny noggins.  We carry Stone Age minds, even though nobody knows what those minds looked like or whether we still have them.

If you have a hammer, all you see are nails.  If you study mating strategies, they apply to voting, too.

But I digress.  What's the female Stone Age mind supposed to want to do when it happens to be attached to ovulating ovaries?  This is the main thesis of Durante and her co-authors:

The driving theory behind this research is that ovulation should lead women to prioritize securing genetic benefits from a mate possessing indicators of genetic fitness (Thornhill & Gangestad, 2008). Accordingly, ovulating women have an increased desire specifically for short- term sexual relationships with men possessing purported markers of genetic fitness, such as symmetry, masculinity, and social dominance (Durante et al., 2012; Gangestad, Thornhill & Garver, 2002; Gangestad, Thornhill, & Garver-Apgar, 2005; Garver-Apgar et al. 2006; Pillsworth & Haselton, 2006). In fact, in the 2008 U.S. presidential election, ovulation boosted women’s preference for the more attractive and symmetrical candidate (Barack Obama) over the less attractive and less symmetrical candidate (John McCain) (Navarrete, McDonald, Mott, Cesario, & Sapolsky, 2010).
Given that ovulation leads women to be more open to short-term sexual relationships, ovulation might alter women’s religious and political attitudes to facilitate such relationships. Because openness to short-term sexual relationships is associated with lower religiosity (Weeden et al., 2008) and more liberal political attitudes (Kurzban et al., 2010), ovulation may lead women to become less religious and more liberal.

Wow.  What fun!   Note the word I have bolded, that "accordingly."  We leap from the idea that ovulating women want to have sex to the idea that ovulating women want to have sex in the form of one-night stands with large-eared men (sorry, Obama) as long as those ears hang symmetrically.  Symmetry, EP folks tell us,  signals good health.

From one-night stands we quickly move to the idea that ovulation may lead women to become less religious and more liberal.  Why?  Because ovulation leads women to be more open to short-term sexual relationships and because such openness is associated with lower religiosity and more liberal political attitudes.

Watch the powerful ovulatory machine!  It turns everything on its head.  More seriously, I think that chain of arguments consist of mostly weak links and questionable assumptions about causality.

But more importantly, IF it truly is the case that ovulating women have a hardwired evolutionary tendency to go for short-term sex when they ovulate, then we would expect that form of a relationship to have become dominant in the human societies over time, not the kind of monogamy or serial polygamy we actually observe.  Over time most children would have been born from such short-term relationships, and most children of any one woman would have different fathers.  This is not what we observe in this world. **

We could argue that cultural arrangements have stopped this from happening, allocating women to individual men as their property, say.  But EP folks never pay much attention to culture and in any case women usually are not supposed to want short-term sexual relationships.  That's what men want, we are told, over and over again.  Women want long-term providers.

I've confused myself here.  That's not my fault but the fault of the patchwork that stands for the rigid and misogynistic type of evolutionary psychology, the type I call EP.  The basic ideas keep changing, slippery as eels, and what women's sexuality might be becomes a kaleidoscope.

I've been told we women never competed in the reproductive markets, I've been told that men want many, many women,  and that women want only one man.  I've been told that men are therefore by nature adulterous and women are not.

When the logical impossibility of that was pointed out (as adulterous heterosexual men need some women, at least, to also be willing to have short-term sex), the EP canon decided that men and women are both adulterous but for different reasons:  Men to sow the maximal seed, women to get the highest quality seed possible.  And so on and so on.

It never ends.  But the latest story is that women are not quite without libido.  It rears its tiny head a little bit, especially around the time of ovulation!  And what delicious stories can then be told about women's reproductive strategies:
Ovulating women, for example, experience increased libido (Bullivant et al., 2004), have greater interest in attending social gatherings (Haselton & Gangestad, 2006), pay more attention to men (Anderson et al., 2010), and enhance their appearance (Durante, Li & Haselton, 2008; Durante et al., 2011; Haselton et al., 2007).

How does all this relate to voting behavior?  I'm not quite sure how seriously the authors take the argument that women vote for male politicians as if the latter were that mysterious strange lover they desire when ovulating.  That is mentioned in the article, but most emphasis appears to be on the thesis that being horny causes women to give more support to policies such as marriage equality and abortion access.  Sounds pretty weak to me.

What about the empirical data and analyses in the study?  Here I met with immediate difficulties.  The manuscript I read is not transparent.  It gives insufficient descriptive statistics on the samples the authors used.

This lack of descriptive statistics matters enormously, and this is why:

The authors discuss their results from the beginning to the end as applying to fluctuations in women's voting behavior and as applying to differences in the women's behavior between the fertile times and non-fertile times.

But, and this is a huge, huge but, the authors did not, in fact, ask the SAME women about the views at different times of their menstrual cycles.  They compared two different samples of women.  One consists of women who were assumed to be in the ovulatory stages of their cycles (based on a calculation formula), the other consists of women who were assumed to be neither in the ovulatory stage nor the pre-menstrual or menstrual stages of their cycles. 

Now, it's OK to use a cross-sectional study to draw inferences about something like ovulation and its many awful consequences.  What is NOT OK is to fail to give the descriptive statistics about the two samples.  We need to know how similar the two groups of women are, before we can use results from them to infer something about the effects of ovulation on any one woman.   Ideally, the two groups should be identical in all other aspects except for whether the women are ovulating or not.

And that data is not given.  The study mentions variables such as age, ethnicity and income, and discusses how they vary between the single women and the women in a committed relationship.  But the comparable discussion on the most important two samples in the study is missing.

Given that omission, I cannot really judge the findings***.  Whenever the authors find a difference between the two samples it could be because the women in the samples differ in more ways than whether they are ovulating or not.

The manuscript tells us nothing about the sampling process but mentions that the participants were obtained through the Internet.  They seem to have self-selected into the study (which paid a small financial compensation).  Given that possible self-selection, looking at the overall statistics on the ovulatory and non-ovulatory samples is crucial, to at least guarantee comparability of the two groups within a study.  Possible self-selection would also mean that the results cannot be statistically generalized to the overall population.

To reiterate:  The message of this study is in its title: "The Fluctuating Female Vote."  We need very strong evidence that data from two different groups of women can be used to draw that conclusion.  In concrete terms, the results of the study tell us nothing about that fluctuation because they compare Ann's views when she was ovulating to Betty's views when she was not.  The Anns in one sample must be like the Bettys in the other sample for the title of the study to apply.

But purely intuitively, many of the findings seem pretty weird.   For instance,  one finding is that single women are MORE religious than married women when in the non-ovulatory stages of their cycles but LESS religious when in the ovulatory stages.  What chameleons these women are!

Finally, if I could have one present for Christmas (or the equinox or whatever), it would be that some researchers outside EP (the narrow kind) went and replicated a bunch of these studies, possibly using the same data.  I really really want to see the results verified or falsified by good statisticians who have not drunk the KoolAid.
*All direct quotes in this post are from the study.

**In fact, the EP studies argue that women whose permanent mates are  less satisfactory (by failing symmetry or sexiness tests and so on, all assumed to measure reproductive fitness) have an increased amount of daydreaming about other men.  In questionnaire studies with, say, 50 pairs of dating  American college student couples, average ages around 20 to 21.  From this the studies conclude that ancestral women would have acted on those urges if they were able to get away with such behavior.   But the Durante et al. study doesn't make this distinction at all.

There's a deeper problem in all these ideas about evolutionary adaptations:  Humans probably lived in groups even in the prehistoric past and the outcomes of all sorts of mating strategies depended on more than the simple theories based on abundant sperm and choosy eggs:  The games people play.

Women cannot create children alone and neither can men.  The overall outcomes were probably based on many different variables.   Hence, it's pretty simplistic to assume that simple mating strategies would be the obvious evolutionary adaptations.  Remember that what genes are getting passed on is the pathway here, and those genes obviously depend on who finally mated with whom and which children were cared for to become fertile adults, in turn.   And so on, generation from generation.

***I could, of course, but there wouldn't be much point in it.  That must wait until the required descriptive statistics are available.