I tried to use some dietary research as data examples in machine learning training courses. After running into self reporting (which I naively thought would be the exception, not the rule) I changed the my use of such sources to class discussions around reporting such as the following: https://www.vox.com/science-and-health/2018/9/19/17879102/br...
People are bad at reporting what they eat. That's a problem for dietary research
211–220 of 470 posts
Re: People are bad at reporting what they eat. That's a problem for dietary research
#212Earlier quoted context omitted.
This goes a long way to further convince me that it is portion sizes in the US. Having traveled, it is quite absurd to see the difference in standard order sizes. Even for zero calorie things like water and unsweetened teas/coffees. You just get smaller cups. I'm not even sure you can get a US large in Japan for most drinks?
Probably depends on where you go. I don't know about "the US", but as a "European" I thought serving sizes were comparable to what I get in restaurants at home. Drinks were an exception, since basically all restaurants had unlimited soda for next to nothing. This was actually great, since I was riding a motorbike in the desert in July. For reference, I live in France and visited LA and random towns in the western sta…
I'll be visiting France soon, so will be able to compare on that front. But I think it is an understatement to say that things are universally smaller.
And on the drinks, even places in Japan that had free refills still gave, at largest, an 8oz cup. Usually, I think they were even smaller. Even getting popcorn at Universal, the bags were large, but nothing compared to what I'd expect over here.
Some of this, I'm sure, is having gotten used to ordering the larges. For a time, it was not unheard of to get a 32oz soda at any given convenience store. May still be normal? I don't know.
(And, of course, this isn't getting in to the sizes of vehicles.)
Re: People are bad at reporting what they eat. That's a problem for dietary research
#213Earlier quoted context omitted.
I downloaded SnapCalorie to try it out on Android. I went all the way through the sign-up phase, only to discovery that I would need to activate subscription in order to have the 7-day trial. Ended up uninstalling the app :(
We're an early stage startup and the models are expensive, we're trying to get the price as low as possible, but yes we need to charge to cover costs right now. Sorry about that! You might get a yearly discount offer that is less than $2/month if you get lucky (A/B test split). But that's less than the cost of running the model for people, so hopefully others will consider paying full price.
I'm on an annual plan from another app (Calory, $30) otherwise I would have bit.
It gives some features with a totally free plan. That makes the IaP feel less like a bait and switch.
The proposition of SnapCalorie is compelling. Calory ui is decent and I use a scale so accuracy should be good but I think their database is shitty. Meatloaf will vary from 1.5 kcal/g to 3, steak will show as 1 kcal/g, stuff like that.
Re: People are bad at reporting what they eat. That's a problem for dietary research
#214Earlier quoted context omitted.
Portion sizes in the US are ridiculous... often 2-3x larger than here in Europe. When I regularly visited New York for work, and we'd get takeaway sandwiches, I'd have to open them and remove half the filling. I just couldn't physically eat that volume of meat, cheese or especially mayonnaise. For all drinks, I'd order small.
New York deli sandwiches are certainly not representative of what you get everywhere in the US. They are famously large.
They may be famously large, but I don't think they are abnormally large for most of the US nowadays? I certainly didn't think they were particularly big when I visited.
Re: People are bad at reporting what they eat. That's a problem for dietary research
#215Earlier quoted context omitted.
What happened to your app? I was on such a research team (Scripps) that used your app for the study (PROGRESS).
Unfortunately it was shut down after I sold the company to MyFitnessPal. I was a shitty business person who thought it made sense to try and build a free consumer product on a bootstrapped budget. We had some traction on the B2B side that paid the bills but COVID took a dent in it and it would have taken a long time to build back the revenue stream selling to healthcare companies (tip for others, it can take 6-18mont…
The study ended so no worries. In any case, congrats on the exit!
Re: People are bad at reporting what they eat. That's a problem for dietary research
#216Earlier quoted context omitted.
>and even trained professionals are still off by 40% I find this very hard to believe, unless the term “trained professional” is quite broad. When I was much more into fitness and weighed every meal to the gram, I could tell if a bowl of cereal was a serving to within a gram or two.
Isn’t that a bit of a special case because you know your cereal and you know your bowls? What about some cooked foods like meats which can vary in density and shape when raw, and also vary even further due to inconsistency in cooking, with more or less moisture cooked off? It’s possible to calibrate your estimates, but if you haven’t done that, it’s probably safe to assume you’re not particularly accurate.
Oh I would only weigh things raw - if we’re talking about guessing the portion sizes at a restaurant for example, you might say I’m cooked.
I wonder how good an ML model might be at that task. Maybe given a photograph of the plate and the menu description.
Re: People are bad at reporting what they eat. That's a problem for dietary research
#217Earlier quoted context omitted.
I am very diligent, and the truth is that it is hard and it changes how you eat to be more countable. On a cut, it matters more. On maintenance, it matters less. But most of it is a guessing game and making an assumption that it will all even out later. Ignore spices - you can assume 25 calories a day and it’ll still be too much. Be diligent about oils. 9 calories a gram bites you quickly. But ultimately, if you miss…
>But ultimately, if you miss 100 calories a day, and are in a 500 calorie deficit recorded, you are still going to lose .8 pounds a week. And if that is consistent, adjust your portions and be fine with how you record. And the thing is, you'll need to do this anyway - because you can't be sure in advance how many calories represents a "500 calorie deficit" for you, in your specific current conditions. I was quite und…
If you're willing to spend money, Macrofactor basically is an automated version of this with a bit more refinement.
Re: People are bad at reporting what they eat. That's a problem for dietary research
#218Does this actually pose an issue for most studies? This seems like it would be an issue for any studies relying on absolute food consumption being accurate. Most studies I come across frame their findings in relative terms (likely for this very reason): Individuals who engage in more of X compared to their peers show a correlation with outcome Y. For example, if you’re trying to determine whether morning coffee consu…
> This seems like it would be an issue for any studies relying on absolute food consumption being accurate.
Exactly. Those studies either don't get done, or when they're done, they produce garbage results that get ignored or get interpreted as diminishing the importance of absolute food consumption.
> it doesn’t seem particularly relevant if you believe everyone is underreporting their food intake
It says that virtually everyone underreports. It doesn't say that everyone underreports equally, and there are good reasons to expect this not to be the case. If embarrassment is a contributing factor, for example, you would expect people who are more embarrassed about how they eat to underreport more. If people remember meals better than they remember snacks, people who snack more will underreport more than people who snack less. If additional helpings are easier to forget than initial helpings, people will underreport moreish foods more than they underreport foods that are harder to binge on. With so many likely systematic distortions, it would be surprising if everyone underreported equally.
Re: People are bad at reporting what they eat. That's a problem for dietary research
#219Earlier quoted context omitted.
>people are on average off by 53% and even trained professionals are still off by 40%. Basically if you want to have a higher level of accuracy you need to use a food scale or something that measure the volume of food, people just can't estimate portion sizes visually. I can typically estimate them accurately without direct measurement, and with feedback that will tend to make errors cancel out over time. My trick is…
Yeah, dividing out a known portion size is a good hack that will probably help with accuracy. In our research most people's calories and error came from eating out where they didn't have these hints, but this is a good trick if you mostly cook for yourself!
Historically I would rely on the restaurant's printed nutrition info. But I don't really eat out often enough for this to matter.
Re: People are bad at reporting what they eat. That's a problem for dietary research
#220Have you ever wondered why it’s such a struggle for a diabetic to manage blood sugar levels in a sensible way? Here’s the answer. I assure you that anyone with diabetes is forced (and the word "forced" doesn’t fully convey the mental burden involved) to maintain an almost obsessive level of awareness about what they eat. There’s no comparison to someone simply “on a diet.”
I guarantee you, it’s an incredibly complex task. Unless one adopts a monastic approach of always eating exactly the same carefully measured meals at home, the challenge is constant.
If one day a system based on vision and AI could accomplish this task (and it can't, it’s impossible), it could charge any price and have millions of users.