[roughly]

> I also don't see any controls for other circumstances, so it can also be explained that other factors could contribute both to eating processed foods and getting depressed (ex: working long hours, having low income, relationship stress)

Of course there were controls.

Directly from the article: “Adjusting for other health, lifestyle and socioeconomic risk factors for depression…”

As noted in another comment, the paper goes into further details on the controls they used.

Honestly, this kind of thing is insulting. If you’re going to levy that kind of accusation of incompetence against someone, you at least owe it to them to read the source material before doing so.

[fshbbdssbbgdd]

It’s good of you make this correction, we need to be accurate when discussing the research at hand.

That said, it’s worth pointing out that “adjusting for” a factor isn’t quite as effective as it sounds: https://dynomight.net/control/

Having built some regression models in my time, throwing a ton of extra variables in doesn’t fill me with confidence that I am “controlling” much of anything.

[roughly]

Following from the above, I'll presume you've got a specific methodological critique based on the process detailed in the paper?

[fshbbdssbbgdd]

The paper in question: https://jamanetwork.com/journals/jamanetworkopen/fullarticle...

It’s a research letter, which is peer-reviewed by JAMA but shorter than a traditional, full-fledged research article.

The full text of the letter’s description of the adjustments:

> with adjustment for known and suspected risk factors for depression, including

[followed by a list of over a dozen factors]

The data sharing supplement says that the code is on GitHub (hmm!), but doesn’t include a link (oh…).

That’s all I have to work with. I can’t make a specific criticism of the adjustments in this study. I’m left with my preconceived notions of how this kind of research is often done, even with peer-review.

When I read the words “we controlled for X in our population study of Y”, I usually take the lazy shortcut of not updating my internal model of the relationship between X and Y at all, because I’ve fooled myself too many times before. If they mention a natural experiment or instrumental variable I get excited.