Bodyfat % testing, which is accurate which is not

[KoolaidMannn]

So my exercise physiology class at school did bodyfat testing with 9 different devices. Our sample group was 68 people, 39 male, and 29 females. We also prefromed rank order correlations, comparing each test to hydro static weighing.

Group data:



Rank Order Correlation:
Hydro vs Skinfold. R=0.749
Hydro vs Circumfrences. R=0.2366
Hydro vs Valhalla (a bio-electrical impedance test). R=0.6142
Hydro vs Omron (also a BIA). R=0.6566)
Hydro vs Tanita (BIA). R=0.5572
Hydro vs Bod Pod. R=0.778
Hydro vs Infrared. R=0.6366
Hydro vs DXA. R=0.786
Hydro vs BMI. R= -0.07882

The circumference test was a 3 site cm measurement that measured the abdomen, upper arm and forearm (all relaxed). Then the formula is %fat= A+B-C - 19.6(for females) and -10.2(for males)

Skindfold testing was a four site test with calipers that were performed at the bicep, tricep, subscapula, and suprailiac. Basically pinch loose skin and measure with caliper

So we see that the strongest correlation is between hydrostatic and Dexa. With the next strongest being BODPOD and skinfold. With the weakest being circumferences (navy body fat and 3 site) and BMI.


Thought this was relevant and cool info for the site to have.

[quark]

Interesting.

I'm disappointed Navy did so badly. More evidence that BMI is not really useful for individuals.

Selection bias - are exercise phys students in better (or different) shape than most?

[Murelli]

Skinfold = single fold (abdominal)?

[KoolaidMannn]

Interesting.

I'm disappointed Navy did so badly. More evidence that BMI is not really useful for individuals.

Selection bias - are exercise phys students in better (or different) shape than most?

That’s kind of a hard question to answer since I don’t know anyone’s fitness levels other than people I’m friends with. Based on just looks, to me 90% do not look any different than normal college student.

Could we look at fat free weight and compare that to a general population average (not sure if there is one out there)

However we are completing a one rep max squat, bench and leg curl this week, so that could give some context to answer your question.

[KoolaidMannn]

Skinfold = single fold (abdominal)?

It was a four site test with calipers that were performed at the bicep, tricep, subscapula, and suprailiac. Basically pinch loose skin and measure with caliper

[MattNeilsen]

This is interesting - thanks for posting it. @timelinex was just cautioning me in another thread about using the Navy Test...this seems to nicely back up his advice. Appreciate it.

[cwd]

I'm surprised the electrical tests were so good, and disappointed that the circumference estimate was so bad. I like my Navy test.

Were any of these impedance tests consumer devices, like the common bathroom scale with fat-tests, or are these more complicated systems?

[Wilhelm]

He also reviews his scale in this.

[Mahendra]

Thanks for posting this!

I did a BodPod test in November, and found that the Navy method underestimated my BF% by about ~3-4%.

[PatrickDB]

Thanks for posting this!

I did a BodPod test in November, and found that the Navy method underestimated my BF% by about ~3-4%.

Someday you need to write your experiences up as a case study in how to get shredded.

[KoolaidMannn]

I'm surprised the electrical tests were so good, and disappointed that the circumference estimate was so bad. I like my Navy test.

Were any of these impedance tests consumer devices, like the common bathroom scale with fat-tests, or are these more complicated systems?

Yeah I was surprised the bia’s did so well too.

Tanita was a scale where you just entered bodyweight and age I believe. Omron is available for purchase for like $35 I think it’s a device you hold in your hands straight out in front of you. and the skin fold was done with calipers which you can also buy for pretty cheap

[mbasic]

The Valhalla thing is the brand name, but I think they have a model you stand on with a pedestal out in front you, and you grab some handles with sensors. So there's a sensor for each hand, and the foot sensor is also split in two.

Don't quote me on this, but it was explained to me by globo-gym-trainer-guy-dude . . .
the device takes several different readings LH to RF, LH to LF, LH to RH, etc and so on every which way and average them or runs them thru a formula or whatever.

My wife did it as part of a fitness pre-test/post-test thing at the globo.
IMO, it was way off (reading "leaner" than reality).

[cwd]

After shopping for an impedance-measuring scale, I decided not to buy one.

I know how my composition is changing, because I track weight, e1rms, and waist size. The scale's BF% could tell me the same thing, with added noise due to varying hydration.

I already know how not-pretty my belly is, from looking in the mirror. Putting a number on it (15% or 25% or whatever) doesn't seem to add any useful information.

[platypus]

Interesting.

I'm disappointed Navy did so badly. More evidence that BMI is not really useful for individuals.

Selection bias - are exercise phys students in better (or different) shape than most?

Ya, the tape test is an unbelievably bad measure of bodyfat percentage. I have seen people with six-pack abs fail, and really fat people pass.

The military could weed out most of the overweight people by just requiring everyone be able to do 5 dead hang chinups. Although that might weed out more people than intended.

[Woody]

After shopping for an impedance-measuring scale, I decided not to buy one.

I know how my composition is changing, because I track weight, e1rms, and waist size. The scale's BF% could tell me the same thing, with added noise due to varying hydration.

I already know how not-pretty my belly is, from looking in the mirror. Putting a number on it (15% or 25% or whatever) doesn't seem to add any useful information.

I reckon this is the most sensible approach. Short of being dissected you can never really know your BF% and weight and waist measurements are perfectly able to tell you whether its going up or down.

[ch]

Thanks for posting this. Even the best body fat measurement tools are estimates, and it appears that most of them undershoot compared to hydrostatic weighing.

I feel like for the average person, the eye test and waist circumference are more useful (and cheaper) metrics than any of these commercial devices.

[cwd]

I know how my composition is changing, because I track weight, e1rms, and waist size. The scale's BF% could tell me the same thing, with added noise due to varying hydration.

I reckon this is the most sensible approach. Short of being dissected you can never really know your BF% and weight and waist measurements are perfectly able to tell you whether its going up or down.

The impedance scales *do* make sense for a large class of people -- dieters who are not training for max strength.

Waist measurements are very coarse and fluctuate a lot. So an exercise/training program that doesn't spit out an e1rm or an actual RM test regularly, like cardio or bodybuilding, or even crossfit, doesn't give you any proxy for muscle gain/loss that you can combine with weight to guess composition changes.

This is another good reason that dieters should be training for strength, I guess. But an impedance scale is a lot easier.

[mgil]

Based on those R values and sample size, skin fold method is essentially equal to DEXA and BodPod using hydrostatic as truth.

I've heard this spoken about anecdotally and maybe seen a study on it.

[KoolaidMannn]

Based on those R values and sample size, skin fold method is essentially equal to DEXA and BodPod using hydrostatic as truth.

I've heard this spoken about anecdotally and maybe seen a study on it.

Yeah I’m like 99% sure I did the rank orders correctly, I can also post my work if anyone wants to verify it. But yeah I agree with skinfold the only bias I see is if the person taking the caliper measure does so incorrectly, we tried to account for this by taking each Measurement 3 times then taking the average.

The hydrostatic bias would be if the subjects weren’t able to blow all their air out underwater. Also we tried to account for this and had people repeat the test 3-5 times as needed.

[bobmen10000]

Based on those R values and sample size, skin fold method is essentially equal to DEXA and BodPod using hydrostatic as truth.

I've heard this spoken about anecdotally and maybe seen a study on it.

Speaking only anecdotally, the skin fold method was nearly identical to a couple readings using BodPod within a 7 day period; however, I experienced fluctuations of nearly 10% with the skin fold method during a short interval, less than 2 weeks, at the same facility but with different clinicians. Its efficacy seems to be mitigated by the individual practitioner much more so than the other methods used in this study, thus making it less than ideal as a means to measure bodyfat % in my mind.