Happy happy happy

Some fun and games, and economics, for social theorists.

The economists at Marginal Revolution have picked up a paper on the happiness gap between men and women, here. The paper is downloadable as a PDF.

The basic story: women are not as happy as they used to be in, as measured subjectively (that is, by people’s self-reports of their happiness of wellbeing), in absolute terms, and relative to men. The paper suggests that it may be because of the women’s movement. (More on that later.)

There’s an article covering the issue in the New York Times, which also refers to another paper showing much the same result. Women aren’t as happy as they used to be, absolutely, and in comparison to men. The NYT article generated about 700 comments before comments were closed off.

Someone at Language Log had some issues with the number crunching, and then analysed some of the comments from the NY Times article. Steve Levitt, of Freakonomics fame, proffered some possible explanations for the gap.

The economists, a bit miffed that the linguists at Language Log – their phrase, not mine – had criticised their numbers defended their number crunching, and then added some more defence here. Language Log responded here.

The numbers made my head swim…. I’m no mathematician. Back and forth, he said / she said, angels on pinheads, storms in teacups. However what seems to be clear is that the journalist oversold the product, especially with respect to one paper. See the update from Language Log at the bottom of this post.

Go ahead and read all the numbers, if that’s your thing. I want to pick up on a couple of things, one minor, and one major. The minor one comes from the very first post on Marginal Revolution.

In the NYTimes article David Leonhardt correctly notes that “Although women have flooded into the work force, American society hasn’t fully come to grips with the change.” Alas, all he has to offer as solution is the usual platitudes about subsidized daycare and how men should do more of the housework – peculiar solutions to women’s unhappiness with increased opportunities. Leonhardt should instead have talked to my wife.

As I wrote this post, I asked my wife about her feeling guilty at home and at work but she told me she no longer feels this way. “Really?” I asked, “Why not?”

“I decided to act more like a man and get over it,” she responded.

It’s a quick one-liner, and it’s the kind of silly or ironic thing I say myself, to my husband, or to some friends. But really – should women have to be like men in order to be happy?

The major issue is much more important. Why do the economists immediately leap to blame the women’s movement?

They set up a syllogistic argument in the first section of the paper, and giving a couple of possible conclusions that could be drawn.

If: (1) The women’s movement increased the welfare of women more than men; (2) This was the most important gendered-biased change during our sample; and (3) Subjective well-being data provides policy-relevant welfare measurements; then we are led to infer that our data should show a rise in the subjective well-being of women relative to that of men since the 1970s. That we observed a decline in the relative well-being of women leads to questions about where the syllogism breaks down. Did men somehow garner a disproportionate share of the benefits of the women’s movement? Alternatively, perhaps the well-being data point to different, and largely undocumented, changes that have disadvantaged women

Then they crunch the numbers, and show that women’s subjective well-being had declined. So in their view, that shows that (1), (2), or (3) must be false. But which one?

in the discussion section of the paper, they tell us the answer. (1) must be false. In other words, having shown that the conclusion of the syllogism is wrong, they then conclude that some of the premises of the syllogism must be wrong. Perhaps the women’s movement wasn’t good for women after all.

Finally, the changes brought about through the women’s movement may have decreased women’s happiness. The increased opportunity to succeed in many dimensions may have lead to an increased likelihood of believing that one’s life is not measuring up. Similarly, women may now compare their lives to a broader group, including men, and find their lives more likely to come up short in this assessment. Or women may simply find the complexity and increased pressure in their modern lives to have come at the cost of happiness.

To be fair to the writers, they proffer various explanations for why women’s happiness may have declined. They are also very clear that the decline they have identified may have as much to say about subjective measures of happiness as they do about any actual gap between men’s and women’s relative and absolute happiness. That is, (3) could also be false.

But why ‘blame’ the women’s movement at all? Although the paper carefully sets out the possible explanations, the sexy one is, “It’s because of all those damned feminists!” Whatever the careful words, the underlying rhetoric is one of blaming women for not being happy. So quick, ladies! Race back to the kitchen, home decorating, and raising children. Clearly having choices to do something else is bad for you.

Like most thorough going feminists, I have no problems if what you decide to do is run a home and raise children. That’s a choice that men and women should be free to make. But the critical word is ‘choice’. If you choose that, then you go for it, girl! And I will be right there with you.

What I don’t like is the under current of the idea that the poor little dears (women) would really be much happier if only they gave up their ambitions. That’s all too similar to the idea that is a great idea for the lower classes to have religion, to comfort them in the day to day difficulties in their lives, to promise them a reward for subservience, and to keep them subservient. After all, they really aren’t smart enough to think for themselves. So best keep them quiet with religion.

The idea is incredibly patronising. And so is the idea that women would really be happier without the fruits of the women’s movement.

This paper is a draft paper. Maybe by the time it is a final version, the authors will have answered some of the many questions they ask.


2 responses to “Happy happy happy

  1. I thought there was reasonable research showing that there’s an optimal level of choice for maximum happiness – too much is as bad as too little. Possibly women have now got more choice than men so have become less happy. Especially since we’re talking barely detectable levels of difference as far as I can tell.

    Me? Not sad to miss out on the “have babies and stay home” option.

  2. Okay, some technical stuff, if you can stand it. At first I thought about various explanations for this along the lines of Freakonomics. But then, when I read the method I realised that there was no detectable difference in happiness AT ALL! Eh? How does this work?

    Okay, first point – the authors method finds a “significant” difference, but that only means statistically significant. In everyday terms the difference is Tiny and not Remotely Important.

    Second point. The total error in a survey is divided into sampling error and non-sampling error. In common with most statistical social scientists, the authors have ignored (abolished) non-sampling error. All they examine is differences that are significant from the point of view of sampling error. If you add in non-sampling error, none of the differences would be significant. This inconvenient truth is usually ignored because nobody knows how to deal with it.

    Non-sampling error is about how you find the respondents, who refuses to answer, what effect the question order has, and so forth. You might argue that as the same survey is used year after year, there won’t be any non-sampling error.

    Bzzt! Thank you for playing. WRONG. We know that response rates have been declining over the years in the USA (for example). What effect has this had on total survey estimates? NOBOBY KNOWS. So when you get a tiny difference that can only be detected with incredibly advanced statistical techniques, you know it will be swamped by all the differences they haven’t measured.

    To put it in laymans (sic) terms. The apparent reduction in happiness could be simply due to fewer happier women answering the survey, for example because they were at work and too busy to reply.

    So the authors have found something. But what they have found is that there is no detectable difference in happiness, unless you abolish non-sampling error. THERE IS NO DIFFERENCE.

    And to those who are wondering. Yes I do know what I’m talking about. Cheers.