Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

People are bad at reporting what they eat. That's a problem for dietary research

science.org

Post reply on HN
Print view View original

171–180 of 470 posts

Re: People are bad at reporting what they eat. That's a problem for dietary research

#171

Post by aziaziazi »

iwanttocomment wrote: post #119
aziaziazi wrote: post #29

For those that "track and weight everything" (how ?) do you manage ?: - sauces you make yourself? I often mix some different oils, mustard, seeds, miso, bit of leamon juice and spices… but weighting and logging everything will take 3x the time to do the sauce itself - different cooking time in one receive : oignons going first, tomato sauce in the middle and parsley at the end (but still cook a bit with residual heat…

As someone who has successfully tracked calories in the past with great effort, the trick is to be strict about measuring calorie-dense foods, but to be liberal with "lighter" foods where the calories are functionally de minimis. An ounce of olive oil has 250 kilocalories. An ounce of lean protein generally has 30-50 kilocalories. An ounce of green vegetables contains virtually no kilocalories. As such, things like o…

That may sense. Most of the folks here seems to track calories and other macro. In the meantime...

> micronutrients like vitamins are not functionally possible to be measured

... my concern is micro: I'm engaging on a full vegetable diet (+shrooms +minerals!) and am concerned about thinks like iron, selenium, calcium... I (got-used-to) love vegetable and eat a lot of them so I'm probably fine with most micros, however may miss some selenium for exemple. Some research seems to show that too much vitamins is usually ok but too much minerals may not be. The more I read the more I'm scared! What makes me feel safe is the three long-time vegan I know seems healthy and don't take any supplement appart obvious B12. Perhaps I should just focus on other thinks that doing mad about micros...

Re: People are bad at reporting what they eat. That's a problem for dietary research

#172

Post by wnorris510 »

m_ke wrote: post #139
wnorris510 wrote: post #130

I've done research in this space for many years at Google AI and now at SnapCalorie. The thing I find interesting is how confident people are in their ability to estimate portion size visually, and in truth how wrong they all are. We published in CVPR (top peer reviewed academic conference for computer vision) and people are on average off by 53% and even trained professionals are still off by 40%. Basically if you w…

At Bitesnap we were surprised at how much interest there was from researchers to use our app for diet tracking. It turns out giving people a piece of paper to write “grilled cheese sandwich for lunch” is not a scalable and reliable way to collect research quality data. We even worked with USDA on putting together a food logging dataset: https://agdatacommons.nal.usda.gov/articles/dataset/SNAPMe_A...

We've also been surprised at SnapCalorie how many researchers have approached us to use the app for more accurate diet tracking for medical study participants. The LiDAR based portion size has been a huge draw for them.

If anyone wants to check out our app or research its on our site: https://www.snapcalorie.com/

PS: Bitesnap was an awesome app!

Re: People are bad at reporting what they eat. That's a problem for dietary research

#173

Post by amanaplanacanal »

f1shy wrote: post #142
Sebb767 wrote: post #9

Earlier quoted context omitted.

> There is no way to keep the other variables like, sport, social interaction, stress and such out of the study. Sure, but with sufficiently many people, these effects will usually average out. I'm not saying the situation is great, but it's still an important field of study and we need to make progress in some way.

No if they correlate strongly: people eating more vegetables are more likely to do sport, and care about sleeping. Not to mention visiting a doctor much often. That is just one example.

Yeah this is a major issue. The first study that reports a link between some specific thing and health pollutes the data for all follow-up studies, because the folks that care the most about their health are going to change their behavior based on it. So after that you will always see a correlation with all the other things that have been reported to be healthy.

Re: People are bad at reporting what they eat. That's a problem for dietary research

#174

Post by zahlman »

While the article may have been published today, putting "self-reported food intake" into a search engine shows me that this is not at all a new finding. I would have considered it common knowledge, even. The entire reason people can conjure the mental image of a stereotypical obsessed dieter, weighing every morsel of food and looking it up in tables, is because everyone else has barely any idea what their intake looks like.

Re: People are bad at reporting what they eat. That's a problem for dietary research

#175

Post by marinmania »

wnorris510 wrote: post #130

I've done research in this space for many years at Google AI and now at SnapCalorie. The thing I find interesting is how confident people are in their ability to estimate portion size visually, and in truth how wrong they all are. We published in CVPR (top peer reviewed academic conference for computer vision) and people are on average off by 53% and even trained professionals are still off by 40%. Basically if you w…

I highly recommend people get a food scale/measuring cups and weighing everything single thing they eat (even small things like nuts and cooking oil) for at least two weeks. After that I think you have a much better appreciation for how many calories your regular meals and snacks have.

Re: People are bad at reporting what they eat. That's a problem for dietary research

#176

Post by mrgaro »

wnorris510 wrote: post #130

I've done research in this space for many years at Google AI and now at SnapCalorie. The thing I find interesting is how confident people are in their ability to estimate portion size visually, and in truth how wrong they all are. We published in CVPR (top peer reviewed academic conference for computer vision) and people are on average off by 53% and even trained professionals are still off by 40%. Basically if you w…

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 :(

Re: People are bad at reporting what they eat. That's a problem for dietary research

#177

Post by tejohnso »

wnorris510 wrote: post #130

I've done research in this space for many years at Google AI and now at SnapCalorie. The thing I find interesting is how confident people are in their ability to estimate portion size visually, and in truth how wrong they all are. We published in CVPR (top peer reviewed academic conference for computer vision) and people are on average off by 53% and even trained professionals are still off by 40%. Basically if you w…

> you need to use a food scale or something that measure the volume of food

Isn't that obvious? Basic high school science projects would have students using measuring devices. Are you saying that it's common for nutritional studies to tell people to eyeball their portions and that is then used as actual data?

I see from the article "Nutritional epidemiology studies typically ask people to keep a food diary or complete questionnaires about their intake over the past 24 hours, a week, or even several months." I find that hard to believe. How could any study like that be taken seriously? That's like having someone stand at a street corner for an hour and observe the population to then come up with an average BMI for the neighbourhood.

Re: People are bad at reporting what they eat. That's a problem for dietary research

#178

Post by bluGill »

chikere232 wrote: post #50
jjj123 wrote: post #33

Earlier quoted context omitted.

For me I mostly just try to log the high macro and/or calorie items. Like if I make a Caesar dressing I’m mostly counting the oil and if I’m being really meticulous I’ll measure the Parmesan and anchovy content. But I’ll ignore the 2tbsp lemon juice, garlic, mustard, etc. since it’s counting so little towards the totals I care about. If you’re trying to measure your vitamin intake this may not work for you, though.

For vitamins are probably easier to start in the other end and have a blood test to check how you're doing. I have no idea if that would involve selling your first born in the US though

Depends on the vitamin.

Many are water soluble and so any excess in the body is peed out by the end of the day and so all tests are useless. Fortunately you typically get more than enough as part of a typical balanced diet and so you shouldn't need to supplement in the first place if you are eating well. Though it is almost impossible to overdose so if it makes you feel good there is no harm in making the vitamin companies rich.

The rest you can get blood tests. In general it isn't worth testing unless your doctor suspects something is wrong though. Just eat a healthy diet and get plenty of exercise and you will mostly be fine. Maybe take some vitamin D in winter, but ask your doctor (my doctor told me vitamin d in winter so that is what I do)

Re: People are bad at reporting what they eat. That's a problem for dietary research

#179

Post by pavel_lishin »

tejohnso wrote: post #177
wnorris510 wrote: post #130

I've done research in this space for many years at Google AI and now at SnapCalorie. The thing I find interesting is how confident people are in their ability to estimate portion size visually, and in truth how wrong they all are. We published in CVPR (top peer reviewed academic conference for computer vision) and people are on average off by 53% and even trained professionals are still off by 40%. Basically if you w…

> you need to use a food scale or something that measure the volume of food Isn't that obvious? Basic high school science projects would have students using measuring devices. Are you saying that it's common for nutritional studies to tell people to eyeball their portions and that is then used as actual data? I see from the article "Nutritional epidemiology studies typically ask people to keep a food diary or complet…

I would wager that just paying attention to, and thinking about what someone eats has a decent impact on their health - so it feels like it's working, and like your estimates are accurate.

After all - once you started doing it, you started losing weight/building muscle/achieving whatever result.

Re: People are bad at reporting what they eat. That's a problem for dietary research

#180

Post by firesteelrain »

aziaziazi wrote: post #29

For those that "track and weight everything" (how ?) do you manage ?: - sauces you make yourself? I often mix some different oils, mustard, seeds, miso, bit of leamon juice and spices… but weighting and logging everything will take 3x the time to do the sauce itself - different cooking time in one receive : oignons going first, tomato sauce in the middle and parsley at the end (but still cook a bit with residual heat…

I overestimate on some things because it is safer than underestimating.
Post reply on HN