DanZ ([info]fclbrokle) wrote,
@ 2009-05-15 01:03:00
Previous Entry  Add to memories!  Tell a Friend  Next Entry
Studying Happiness
The Atlantic has a fantastic article out about a longitudinal study tracing the path of 200-odd graduates from Harvard in the 1930s. They've been studying them through the present-day, doing surveys and interviews, physical examinations, and more.

I've found that, over the years, I've gone through phases about what kinds of studies I consider important. Sometimes I believe passionately in the importance of a careful, controlled experiment, while at other times I become a fan of the holistic lens provided by in-depth case studies. I've come to the conclusion that both have their place, and the "Grant Study," as it's known, is an excellent example of why a holistic study can be important.

Let me be clear: the Grant study is an incredibly imperfect beast. It has no control group; it doesn't even have a hypothesis that it's trying to study. It's a study of a bunch of white men from Harvard in the 1930s, for crying out loud. It's incredibly colored by the views and perceptions of the principal investigator, who interprets the interviews, designs the questions, and theorizes about what's going on. By any scientific metric, this study has no right to be providing anything useful, and yet, it is deeply insightful.

After all, it's trying to answer the big questions. It's trying to say, "this is what makes us happy." It's trying to get an understanding of things so large that you could never design a control group, never state your hypothesis. That's the point. Happiness is such a huge beast that you can't study it all at once. What are you going to do, ask people to bubble in on a scale from 1 to 5 how happy they are? No, we're at a point where we need to get ideas, see lives from a distanced perspective, figure out what questions we should even be asking.

The Grant Study is a fabulous look at what makes us happy; it captures some amazing stories of evolving human lives. I wish I could see the data if only because the stories themselves capture my passion. It seems to me to be the kind of thing that can point the way to future resources that will tell us about ourselves. Here's to doing more research of this kind.



(2 comments) - (Post a new comment)


[info]easwaran
2009-05-15 08:21 am UTC (link)
"By any scientific metric, this study has no right to be providing anything useful, and yet, it is deeply insightful."

That just seems to come from a rather limited view of science, which scientists often perpetuate, I think. If one can do a carefully controlled study, double-blind randomized and everything else, then one should do it of course (assuming there aren't further ethical issues to consider), but one shouldn't pretend that anything other than this gives absolutely zero information. I think traditional statistical methods require that there be a hypothesis in advance, as well as all these other things, in order to live up to the advertised standards (i.e., 95% of all such properly done experiments should give confidence intervals that contain the true answer). But although straying from the prescribed methodology means that we can't give confidence intervals of this sort, we can clearly learn quite a lot from them.

As of course you're pointing out :-)

Do other people have access to the full data as well, or only the publications of the PI for the study? Because of course, building up data sets like this that other people can use can also be an incredibly productive thing.

(Reply to this) (Thread)


[info]fclbrokle
2009-05-15 07:59 pm UTC (link)
Heh, yup --- I've had this same discussion, in LJ comments at some previous point. We're entirely agreed; you just have to think carefully about how to interpret studies with different designs.

As far as I know, no one else has access to the full data, but I'm not sure. I think there are very very big confidentiality issues. I think the data is set to become public some time after the last person in the study dies.

(Reply to this) (Parent)


(2 comments) - (Post a new comment)

Create an Account
Forgot your login or password?
Login w/ OpenID
English • Español • Deutsch • Русский…