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

How do food manufacturers calculate the calorie count of packaged foods? (2003)

scientificamerican.com

Post reply on HN
Print view View original

11–20 of 156 posts

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#11

Post by alexbock »

analog31 wrote: post #10
alexbock wrote: post #7

Earlier quoted context omitted.

Pizza in particular is an interesting example because I would expect leavened bread products to be one of the cases where summing the components would cause more problems than usual (because you're going to lose sugar and other nutrients to the yeast). I'm not sure how many calories worth of carbs yeast consume during normal bread making, but I would expect it to be measurable. When making wine, yeast can easily cons…

I suspect it's not a large effect for bread. My family bakes our own bread, and I would say without having measured it on a scale, that there's not a noticeable loss of weight from one end of the process to the other. Also, you'd notice the stuff bubbling if there were a significant production of CO2. I've also made beer, and it bubbles like crazy once it gets working. One possible difference is that bread is made fr…

If anyone can find a reference on the amount of CO2 produced during bread making it would be easy enough to calculate. I haven't been able to find any relevant sources except this old one [1] which mentions that the alcohol content in bread can be as high as 1.9 percent (which sounds quite high to me, but it's hard to critique when they don't explain their methodology).

[1] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1709087/?page=1

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#12

Post by lukasm »

In Poland it's regulated, you can be off by 15%. My mate did tests in a lab. All products he picked showed 15% lower kcal value on the packaging.

I pick reduced fat houmous from the fridge. The label says 100g has 256kcal (18g of fat, 13g of carbs, 8,5g of protein = 248kcal)

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#14

Post by alexbock »

lukasm wrote: post #12

In Poland it's regulated, you can be off by 15%. My mate did tests in a lab. All products he picked showed 15% lower kcal value on the packaging. I pick reduced fat houmous from the fridge. The label says 100g has 256kcal (18g of fat, 13g of carbs, 8,5g of protein = 248kcal)

You would have to be very confident in the accuracy of your measurements to throw away your margin of error like that. If you report values 13% lower to try to appeal to consumers and your lab numbers are off by 3%, now you're wrong by over 15% and you're going to have a problem (if anyone even notices, or cares).

Are there any Polish media reports about this practice? It sounds like the kind of thing the American media would love to report about if it was happening here. (Anything that fits the basic format of "Are companies misreporting arsenic levels in your child's grape juice? Story at eleven!" is going to make a journalist here salivate.)

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#15

Post by ourmandave »

Depends on what side of the package you're reading.

On the back label its says calories per serving. On the front they use a different serving size to make it seem like it's low-cal.

Like popcorn that's 140 cals/serving, but on the front it says 35 cals per cup.

http://www.foodfacts.com/ci/nutritionfacts/snack-foods/smart...

Or a pkg of 2 cupcakes is 380 cals, but they'll say 190 cals per cake (in small type). It makes you think it's 190 for the pkg of cakes.

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#16

Post by mysterypie »

cma wrote: post #4
alexbock wrote: post #2

The obvious next question then is how the protein and carbohydrate content is measured. That sounds like it would require more difficult analytical methods than bomb calorimetry and might not be worth the time and expense if it were not for the fact that those need to be reported separately on the label anyway. Brief searching suggests that proteins are measured with a nitrogen content test and fats are determined by…

The protein test was famously faked with melamine in the Chinese dog-food scandal that killed dogs and resulted in a recall: https://en.wikipedia.org/wiki/2007_pet_food_recalls

I was following that case closely back in 2007 because of my pet (thankfully not affected). An astonishing thing I learned was that practically every brand of cat food and dog food in North America is/was manufactured by a single company called Menu Foods in Streetsville, Ontario, Canada.

The pet food "brands" don't have their own factories. They all contract it out to the same place. A few journalists remarked on this secret fact that came out because of the recall.

We agonize over which brand is healthier for our cats and dogs, but it's all an illusion. The majority of US and Canadian pet food is made at one factory with the same basic animal ingredients, same quality control, same employees, and the same machines. Maybe they adjust the proportions slightly and tweak the flavors. Then they slap on different labels like Purina, Iams, Royal Canin, whatever.

It's saddening to learn that we really don't have as much choice as we think. I'm much more skeptical and questioning than the average person, and yet I assumed that the pet food's "brand name" somehow made a difference.

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#17

Post by Someone »

cma wrote: post #4
alexbock wrote: post #2

The obvious next question then is how the protein and carbohydrate content is measured. That sounds like it would require more difficult analytical methods than bomb calorimetry and might not be worth the time and expense if it were not for the fact that those need to be reported separately on the label anyway. Brief searching suggests that proteins are measured with a nitrogen content test and fats are determined by…

The protein test was famously faked with melamine in the Chinese dog-food scandal that killed dogs and resulted in a recall: https://en.wikipedia.org/wiki/2007_pet_food_recalls

The linked-to https://en.wikipedia.org/wiki/2008_Chinese_milk_scandal was way more scandalous.

When bought in bulk, milk isn’t priced by the liter because it is too easy to water it down. Instead, it is priced by kilograms of protein.

Measuring protein is hard, so they measure nitrogen instead (https://en.wikipedia.org/wiki/Kjeldahl_method#Conversion_fac..., https://en.wikipedia.org/wiki/Dumas_method#Advantages_and_li...). Melamine (https://en.wikipedia.org/wiki/Melamine) is cheap, rich in nitrogen, and, unfortunately, poisonous.

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#18

Post by chris_va »

alexbock wrote: post #14
lukasm wrote: post #12

In Poland it's regulated, you can be off by 15%. My mate did tests in a lab. All products he picked showed 15% lower kcal value on the packaging. I pick reduced fat houmous from the fridge. The label says 100g has 256kcal (18g of fat, 13g of carbs, 8,5g of protein = 248kcal)

You would have to be very confident in the accuracy of your measurements to throw away your margin of error like that. If you report values 13% lower to try to appeal to consumers and your lab numbers are off by 3%, now you're wrong by over 15% and you're going to have a problem (if anyone even notices, or cares). Are there any Polish media reports about this practice? It sounds like the kind of thing the American me…

Kind of reminds me of this story:

http://www.informatics.indiana.edu/donbyrd/Teach/Math//Gamow...

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#19

Post by chris_va »

jcfrei wrote: post #13

I have a related question: Can the calorie count reliably predict the effect on a persons weight? Or is the amount of carbohydrates and fats much more important?

Reliably? Short version is no, and it is even more complicated. Two individuals eating the same food will gain weight differently, depending on a large number of factors.

Re: How do food manufacturers calculate the calorie count of packaged foods? (2003)

#20

Post by tgb »

jcfrei wrote: post #13

I have a related question: Can the calorie count reliably predict the effect on a persons weight? Or is the amount of carbohydrates and fats much more important?

I think the answer is that it's really hard to study this, since food diets are hard to control over long periods of times but that current research is consistent with calorie counts being a good estimator of a person weight gain/loss. Here's a reasonable summary with some studies you can look at for the details. [1] But this is certainly an area of ongoing research and the devil is in the details: does it matter when you eat? [2] Does your gut microbiome affect it significantly? Genetics? Is there inter-person variation in how their body adapts to different food? Does fat/carbohydrates/protein/fiber content affect satiety in a manner that helps people lose weight in real life situations even if it doesn't matter in a clinically controlled setting where overeating is impossible?

We might never have full pictures of all these questions.

[1] http://examine.com/faq/what-should-i-eat-for-weight-loss/

[2] I know specifically that this is currently being tested in some large-scale, high-cost studies. Which tells me that researchers aren't as confident as [1] makes it sound like.

Post reply on HN