5 Comments

This doesn't affect the main point of your post, but I think the height example is potentially misleading because the SD for height is low and humans are in a pretty narrow band of absolute heights. But that doesn't seem to be true for happiness or other psychological traits.

Expand full comment
author

This is an interesting argument. Certainly it's true that with a larger standard deviation the same effect size corresponds to a large absolute difference. But I find it very hard to compare the standard deviation of height to that of happiness... I actually find it pretty plausible that one might be much larger than another (and I consider this a real weakness of the general focus on effect sizes) but not sure how to make this argument with any confidence.

Expand full comment

I wonder how much of the happiness-gratitude correlation is just having more stuff to be grateful for.

Expand full comment

Funny, I received this piece at the same time I received my beta access to https://consensus.app/search/ , a project I'm very interested in. And my first criticism of it is quite well exemplified by your post:

Although it's great to get an overview and a bibliography on a subject, A whole bunch of scientific studies seeming to support a point does not necessarily close the very specific question we are asking. And it feels a bit naive to think otherwise. Had I tried to answer your question with this tool, I also would have come up with a monumental "yes, of course": https://consensus.app/results/?q=Does+gratitude+increase+happiness%3F

Expand full comment
author

I agree that when there's a large body of work, you can usually cherry-pick studies to get basically whatever solution you want. The standard solution to that is to look at meta-analyses instead, which is what I did here. (I don't know much about consensus, but most of those quotes don't seem to directly answer the question, or are kind of ambiguous. A meta analysis is basically a person summarizing the literature in a proven structured way, so for now at least, I certainly trust meta-analyses more than consensus.)

Expand full comment