The Debate on Female Happiness Heats Up

I blogged a few days back about the interesting new paper by Betsey Stevenson and Justin Wolfers analyzing trends in happiness by gender, and finding statistically significant reductions in how happy women are relative to men.

Elsewhere on the Internet, the paper has drawn the ire of a number of bloggers.

Stevenson and Wolfers have fired back on Marginal Revolution, where Wolfers is guest blogging this week.

In the end, I think the critics score some minor points, but Stevenson and Wolfers are the ultimate winners. The results they find are compelling and statistically significant, but of a moderate magnitude in terms of “economic” significance. By “economic significance,” I mean “how important the effects are in terms of real world impacts.” Women in 1972 were at the 53.3 percentile of the male distribution; they are now at the 48.8 percentile. Is this a monumental shift? Maybe not. But compared to how much other factors move happiness metrics, it is pretty large. They are quite honest about the magnitudes in the paper. To the extent their results are being exaggerated, it is by people like me who write blog posts about their paper without being explicit about the size of the effect. The authors can’t reasonably be blamed for that.


MamaKangaroo

Excuse my ignorance, but please explain what is meant by "percentile of the male distribution." Thanks

JPWinnon

I think their results demonstrate good news.

All of the media out there is constantly trying to tell us that we are not happy (without their , due to ).

With such an overwhelming barrage of negativity, they were only influenced as much as indicated in that report?

That's absolutely amazing! Most advertisement and media would indicate that we are the most unhappy people in the world. The fact that anti-depressants are now the number one prescription drug might have actualy skewed their data and kept the results from showing more unhappiness.

It does offer the question, if men and women are so happy, why do we need to take so many drugs to make us that way?

Josh

MamaKangaroo - line up 100 women in order of their happiness and then line up 100 men in order of their happiness and then shift the two lines of people so that each woman is across from the man who is as happy as her. In 1972, the woman in the middle (the 50th) would have been across from the 53rd man (who is slightly happier than the average (median) man). Today that 50th woman is across from the 49th man (who is slightly less happy than the average (median) man).

dissappointed

What is drawing the biggest ire (and most deservedly so) is not the reports themselves on happiness, but the coverage of those reports. The 700 or so comments on the original NYTimes story all leaped on the conclusion as meaning that women and more unhappy than men, and then extrapolated negative conditions from their own lives and relationships out to the woes all of humanity.

The point is not just how small of an effect is found, but how overwrought and unscientific the commentary on it is. Basically a molehill of a result being made into a mountain of rhetorical anecdote-based proselytizing.

Hence, why I was annoyed when it came up on this blog, as I had been impressed by the level of rigor in the analysis you presented in your book, and have increasingly been dissappointed in the sensationalism scattered into the blog.

Although, to be fair the comments to your blog posts were generally of higher caliber than those in the wider newspaper. And I am glad to see that you reposted the most interesting link brought up in the comments onto the main page.

Read more...

ZBicyclist

"Stevenson and Wolfers are the ultimate winners" in the sense that they are getting a lot of publicity for their research, and publicity is currency.

But, in the end, the public is not well served by the PR machine having exaggerated the effect so greatly.

It also shows how the blogosphere can provide some control over this. I was impressed by the questions that were raised on the "Language Log" blog http://itre.cis.upenn.edu/~myl/languagelog/archives/004965.html

misterb

Dr. Levitt,
You vote for Stevenson and Wolfers; I vote for Liberman. Liberman's main point, and the point that S&W don't address at all is that the variation between groups is much smaller than the variation within groups. S says that the variation between groups is 1/8 of a standard deviation; assuming a normal distribution of happiness (a completely unwarranted assumption), if we want to maximize happiness, we shouldn't focus on men v. women, but on why the people at the top of the distribution are so much happier than the people at the bottom. Unfortunately, i have a feeling that the data would ultimately show that the happy people are richer, healthier and better-looking which wouldn't get many headlines, would it?

MCP

I think it's got to do with two basic things :
1) hormones, and
2) manipulation of the male.
When the first shows up as a chemical fact, men naturally assume superiority.
When the second occurs, men ascribe it to the first, plus the physical fact that women have more delicate and complicated systems, especially mental ones. This reflects in inter-sexual conflict, giving a strong inpression to men that women are control freaks (true, IMHO).
The end result is, either way, women end up feeling inadequate when confronted (directly or indirectly, as in imagined), and therefore feel unhappy.
I speak through experience.
Cheers, mate!

Steve

Dr. Levitt - I am addressing this point here even though it is off-topic because I do not know of a way to write you otherwise. I just want to point out that I have no real problem with the partial RSS feed. However, after a month or so of use after the change, I can tell you that while I used to read 6 of every 10 articles posted, I now read roughly 1-2. I do not know why this is, exactly. Perhaps it's because the partial feed doesn't show me how long an article is, how much I'm about to commit to. But I thought you'd like to know how the change has affected the reading habits of someone who is neither for nor against partial feeds.

Jonathan

I agree exactly with what misterb says in post 6 of this thread. Since the variation within groups is much larger than the variation between groups, it would seem inadvisable to use such data for good policy making (perhaps economically insignificant, although I am not an economist).

ori

Lots on articles and studies on being happy.

An article from the Post today:
"Americans report being generally happier than people from, say, Japan or Korea, but it turns out that, partly as a result, they are less likely to feel good when positive things happen and more likely to feel bad when negative things befall them. Put another way, a hidden price of being happier on average is that you put your short-term contentment at risk, because being happy raises your expectations about being happy. When good things happen, they don't count for much because they are what you expect. When bad things happen, you temporarily feel terrible, because you've gotten used to being happy. "

http://www.washingtonpost.com/wp-dyn/content/article/2007/09/30/AR2007093000632.html?nav=hcmodule

EB

When the second occurs, men ascribe it to the first, plus the physical fact that women have more delicate and complicated systems, especially mental ones. This reflects in inter-sexual conflict, giving a strong inpression to men that women are control freaks (true, IMHO).
The end result is, either way, women end up feeling inadequate when confronted (directly or indirectly, as in imagined), and therefore feel unhappy.

This is a joke, right?

madcap

All those complaints are clearly the result of unhappy women.

november

Josh,

Your explanation was excellent. You should be teaching (if you aren't already).

Rita: Lovely Meter Maid

Well, at the risk of getting a "happiness" lecture from "EB", I must say that "happniess" is not a fixed state for anyone, of Either gender. It's a wisp of a thing, often enough, and the very stability that people hunger for to achieve happiness (good job, home, career and money) can and Does also bring with it, often enough, a staid complacency and a feeling of being cemented into place by rivets of dreary obligation. At least some of the time. "Happiness" fluctuates for people, as an elusive but potent idea.

conchis

misterb says:

"Liberman's main point, and the point that S&W don't address at all is that the variation between groups is much smaller than the variation within groups."

To the extent that they don't address it, I think it's largely because it isn't relevant. The key issue isn't the size of the between group variation relative to the within group variation, it's the size of the between group variation relative to the potential variation amenable to change.

The best guess at the moment appears to be that this may only about 20% of the total variation in happiness, because the most differences in how happy people say they are appear to be depend on relatively fixed personality traits etc.

Moreover, Justin Wolfers addressed the potential variation issue directly over at Marginal Revolution: relating the male-female differential to changes in income and employment.

MCP

11. EB, it's not a joke at all! If you are a woman, you'll understand female insecurity, which makes womyn want to stay on top of things.
The only way, because they're soft-wired, is to subtly 'control' their men.
As a reaction, any sensible man (hard-wired) will quickly perceive this pattern, and usually react in a physical manner (cliches : she drove me to drink; she drove me up the wall; she tortured me mentally ... all true according to the male psyche, and justified, in my personal experience, IMHO!)
Women, and men, can be really bad! These are the majority (true of ANY sample, including both sexes) of the species, and the few remaining ones have the guts to not let insecurity in the way, and get on with life.
Fear (insecurity) is the obstacle, and the conquering of it is the key.
No joke!
Cheers!
MCP

david

the book you wrote - "freakonomics" - was very interesting in that it challenged basic assumptions with new data techniques and analysis. i would say that men are happier because of women - 1) working more - they have careers with more pay than in the 70's; therefore, they can afford to pay half (for a vacation, dinner out, etc...), thereby reducing the man's monetary obligations and 2) ordinary women can and are becoming more attractive - either through reconstructive surgery - see breast implants; therefore, men are happier showing their girlfriends off at parties and more importantly, men are more aroused by the new body, face, etc...of their girlfriend/wife.

Energy Guy

It appears that Wolfers may have used the wrong survey weight from GSS in the analysis and so all of the numbers are wrong. His posting of Stata code at Marginal Revolution shows that they used OVERSAMP as the weight when they should have used WTSSALL. WTSSALL corrects for how GSS survey just one adult per household so single people are over-represented. I don't have access to the full dataset, but my initial checks using the Berkeley SDA tool for analyzing GSS shows substantial differences between their numbers and the correct numbers when you change to the proper weighting variable.

Alex C

My thesis advisor always used the term "statistically *reliable*" instead of "statistically significant" since the latter term (unfortunately now ingrained in our discourse) is deeply misleading wrt. the common meaning of "significant". An insignificant difference can be statistically reliable.

And as for your (very refreshing) mea culpa ("To the extent their results are being exaggerated, it is by people like me who write blog posts about their paper without being explicit about the size of the effect. The authors can't reasonably be blamed for that.") -- thank you very much, and that's exactly what the Language Loggers were complaining about. Their only complaint with the original study was that the raw data weren't available. Theirs was a comment on media exaggeration of reliable yet insignificant scientific findings into a "Happiness Gap" (ugh).

EB

MCP, I really have to disagree with you that women are "the weaker sex," which is the only thing you're claiming.

1 - I'm a woman and have survived things, and even grown from and thrived in spite of things that have broken men.

2 - I don't feel a need to manipulate a man. Men who can be manipulated dropped off my radar when I was dating because I wanted nothing to do with them. The problem with men is, there are too few of them, and far too many boys who are just aging faster than they like. Real men can't be manipulated, which makes your entire argument based on fallacy, not fact.