How's Your Experience with Body Fat Scales?

Triathlon Forum
1

I’ve been a long time user of the Withings body fat scale - specifically the latest model, the WS-50.

For those not familiar with this scale, it (and others like it e.g Tanita) use a method called body impedance analysis to determine body composition i.e fat vs lean mass (muscles, bone etc). You stand on it in bare feet, it sends a weak electric current up one leg and down the other and based on the time from A to B, it can determine your percentage body fat.

I’ve always wondered how accurate this is. And in particular, which of the two measurement models it offer - “normal” or “athletic” - was the right one for me.

Withings’ own guidance is as follows: if you work out more than 8 hours a week and have a resting HR of below 60, then you’re athletic. Otherwise, you’re normal. Well, I work out for about 8 hours a week and my resting HR is about 60 so I’m right on the fence.

A little while ago, I came across this: some research into the accuracy of these scales by the revered gadget guru DC Rainmaker.

http://www.dcrainmaker.com/2011/08/bod-pod-consumer-scale-comparison-tests_25.html

His conclusion is that they are on the whole pretty accurate - to within 5% or so.

So I set out yesterday to repeat DC’s tests for myself. Like him, I used the BodPod.

Results as follows:

Withings - athletic mode 8.2%
Withings - normal mode 18.5%
True percentage (BodPod) 14.7%

Pretty disappointing to be honest. Not sure how to interpret this. Maybe it would be more accurate if I was more obviously either “athletic” or “normal”.

By the way, it’s worth stating that from a weight measurement point of view, the Withings was spot on - agreed with the (highly accurate) BodPod scales down to the decimal place - 171.7 pounds.

In the end, I do agree with DC’s conclusion - absolute accuracy is not as important as consistency since the day to day trend is the important thing. So no plans to give up my scale.

But still a little disappointing.

Anybody got any other experiences to report?

2

I’ve been looking at getting some of the Withings scales and started a thread a little while ago to canvas some opinions and experiences.

The general opinion regarding the body fat analysis of the WS-50 seemed to be that it was better at detecting changes in hydration than body fat, but I guess if you can control for that reasonably well then you might be able to pick up trends over time.

I can’t offer any personal experiences, but my instinct would be to just rely on the related metrics that you can trust, i.e. weight.

Thanks for sharing your experiences though.

3

One consideration is that Bodpod (air displacement plethysmography) isn’t the best in terms of accuracy. It is third behind hydrodensitometry (under water weighing) and DEXA. Often, Bodpod over reports by a few percent compared to HD. There’s a chance you’re lower than 14% (and frankly a chance that you’re higher).

4

Where does the good old tape measure and body fat callipers rate on that scale?

I remember seeing a behind the scenes documentary of one of the World Tour team’s training camps recently and this was still the method of choice, presumably primarily for logistical reasons.

5

Withings - athletic mode 8.2%
Withings - normal mode 18.5%
True percentage (BodPod) 14.7%

Pretty disappointing to be honest. Not sure how to interpret this. Maybe it would be more accurate if I was more obviously either “athletic” or “normal”.

That seems inline with what DC Rainmaker found. Normal Mode is only 3.8 % Body Fat away from BodPod, which is actually better than his average. . I would guess that you’re somewhere between the Athletic Mode and the Normal Mode, so maybe just assume your body fat is somewhere in the middle of the 2 results. Averaging them gets you 13.35%.

6

His conclusion is that they are on the whole pretty accurate - to within 5% or so.

I think most of us can eyeball body fat% within 5% .

I’ve found it helpful to measure short-term issues like hydration more than for tracking body fat.

7

If the article below is to be believed, the various body fat measuring alternatives can be ranked as following - from least to most accurate:

#7 BIA scales
#6 handheld BIA
#5 Calipers
#4 BodPod
#3 DEXA
#2 Hydrostatic (“dunk tank”)
#1 InBody

(I do wonder about the impartiality of this ranking - specifically, I’ve never heard of the “InBody” - smells like paid product placement to me…)

http://www.shape.com/weight-loss/weight-loss-strategies/best-and-worst-ways-measure-body-fat

But as I mentioned in my previous post, I’m not too concerned about the inaccuracy of my Withings - as long as it is “consistently inaccurate”. As long as you are careful in your measuring, you should get valid trend information: ultimately, the absolute measurement is not critical, you really just want to see if you are getting better or worse.

For example, it’s well known that your hydration levels will skew the results of a BIA scale. So my method is really simple: I jump on the scale first thing every day, after taking a pee but before eating/drinking anything. My assumption is that my hydration level at that time will be pretty consistent from day to day so hopefully my measurements are valid.

And at the end of the day, you can’t beat the cost or ease of use of BIA scales - and that’s probably the most important thing.

8

If you are familiar with how the math for body fat measurement (even underwater weighing) is derived and the assumptions taken to derive the equations for calculating body fat percentage, there are so many limitations and flaws in the methodology that it boggles the mind. Accurate underwater weighing is somewhat accurate for determining the body fat percentage for populations of average males. That is about it. Trying to go beyond that, like trying to accurately determine the body fat number for a single highly athletic single male individual is essentially a hopeless task. It gets even more crazy if your subject is a highly athletic female.

So, unfortunately, all of these home body fat measurement devices are just attractive paperweights. They generate all kinds of interesting numbers, but the numbers mean nearly nothing. The best way to gauge your body fat is to have your skinfolds taken over a period of time (all measurements to taken by the same pro). The body fat percentage you get will still be garbage, but the sum of your skinfolds will show a clear progression over time. However, even that method completely ignores intra-abdominal fat which varies hugely between individuals, even those with very similar skinfold numbers.

9

So my method is really simple: I jump on the scale first thing every day, after taking a pee but before eating/drinking anything. My assumption is that my hydration level at that time will be pretty consistent from day to day so hopefully my measurements are valid.

Unfortunately, this is not a valid assumption, and not for an athletic individual in training for sure.

10

I bought a Omron HBF-514C Full Body Composition Sensing Monitor and Scale for $60. It seems pretty good. After a few months, I paid to weigh myself in a mobile dunk tank. The scale was within a tenth of a pound & degree! I was shocked.

My biggest gripe about body fat stats is that the recommended medical ranges are usually off for tall people.

11

I bought a Omron HBF-514C Full Body Composition Sensing Monitor and Scale for $60. It seems pretty good. After a few months, I paid to weigh myself in a mobile dunk tank. The scale was within a tenth of a pound & degree! I was shocked.
My biggest gripe about body fat stats is that the recommended medical ranges are usually off for tall people.

Can you expand on what you mean here??? Too high or too low??? How tall is “tall” from your perspective???

12

https://www.nhlbi.nih.gov/health/educational/lose_wt/BMI/bmicalc.htm

**BMI Categories: **
Underweight = <18.5
Normal weight = 18.5–24.9
Overweight = 25–29.9
Obesity = BMI of 30 or greater

I find these ranges untenable.
I’m 6’ 187, and the omron gives me a BMI of 24.8, nih above yields: 25.4.
And yet I’ve slimmed down from 219, which felt overweight but not obese as nih proclaims.

Other food for thought:
http://www.webmd.com/diet/bmi-drawbacks-and-other-measurements
http://blogs.plos.org/obesitypanacea/2012/02/10/why-the-body-mass-index-bmi-is-a-poor-measure-of-your-health/

13

https://www.nhlbi.nih.gov/health/educational/lose_wt/BMI/bmicalc.htm

BMI Categories:
Underweight = <18.5
Normal weight = 18.5–24.9
Overweight = 25–29.9
Obesity = BMI of 30 or greater

I find these ranges untenable.
I’m 6’ 187, and the omron gives me a BMI of 24.8, nih above yields: 25.4.
And yet I’ve slimmed down from 219, which felt overweight but not obese as nih proclaims.
Other food for thought:
http://www.webmd.com/...d-other-measurements
http://blogs.plos.org/...sure-of-your-health/

Well, really all BMI calculators should give same answer. BMI is pretty easy to calculate if you just use the original metric formula and convert your height and weight to meters and kg, which is really not that hard:

BMI = (Weight in kg)/(Height in meters, squared). H = 6 ft (12in/ft) = 72 in/39.37 m/in = 1.83 m, then H squared = (1.83)(1.83) = 3.34. W = 187 lb/2.2 lb/kg = 85 kg. So, BMI = 85/3.34 = 25.4. Not sure why your Omron says 24.8 but it should be 25.4.

The other thing to remember about BMI is that it is really just for “average” people; a well-muscled person can easily exceed the BMI standards but still be relatively lean.

14

I’m curious would you adjust your nutrition/training based on body fat number? Do you need it to be so accurate? I think it’s fun to talk about being under 10% but from a training perspective and nutrition one doesn’t matter. We know if we are eating the “right” stuff and we know at what weight we perform the best based on the course and exercise type. For example, I know I need more muscle to really kill it on a hilly bike course but can get away with less if the run is also super hilly as my run time will offset my slower bike time.

I have good luck getting a caliper test by the same person at regular time intervals to really get a good handle on tracking. Probably not percentage accurate but the trend was what mattered and the overall body type for the activity.

15

My overall experience has been pretty negative.

I’ve owned two different sets of scales, and have found similar problems with both:
They both overestimated my BF%, probably significantly - They measure me generally in excess of 20%, if you took one look at me then you’d see that this is BS. If I had 12kg of bodyfat that I could shed then I’d have the potential to be a seriously competitive pro cyclist (which I don’t).The trends were inconsistent - This is a much bigger issue for me. I don’t mind it reading 25%, and then if I lose a little bit it reads 24.5%. If that’s the case then I can focus on the movement rather than the number. What I found was that it was confounded by how hydrated I was. I could get on it and see one number, drink a litre of water, wait an hour, then get a significantly different number. I can’t trust something like that.

16

I’ve used a Tanita model for about 10 years. I’m sure the technology has improved a bit, but it has tracked some minor variations over time. I can’t really tell if it is accurate versus another method but I get a strong sense that it is consistent if you measure yourself under the same conditions regularly. The athletic mode on older unit seems to just subtract 4% from the number as far as I can tell and I believe their interpretation of athletic leans more toward body building than aerobic fitness. I think it is worthwhile if you want to get a better picture of the composition of your weight change because raw weight numbers are a pretty imprecise measure of progress.

17

It reminds me of my Stages power meter. It seems to be consistent, but I have no reasonable way of knowing if it’s accurate. You do need to be at the same hydration and level of glycogen stores (since it’s stored with water) to get descent data.

I think putting on a pair of recently washed jeans is equally accurate.

18

I’ve had a Tanita for almost 20 years. The BF reading does change as my weight changes over time, the numbers go in the expected direction (up if I’m gaining weight, down if I’m losing) and does track with the basic look in the mirror test so it clearly is measuring body fat. But I highly doubt the number on my scale is accurate. It is really good at tracking my hydration level. If I do a long hot ride my BF% will swing 4 or even 5% until I rehydrate.

All and all, its not a bad feature on a scale and, frankly, its hard to find a nice scale without it. But, at least for my scale, I take the numbers it spits out with a grain of salt.

19

I think putting on a pair of recently washed jeans is equally accurate.

That is the gold standard home body fat test!

20

BMI is simply a function of your weight (assuming your height doesn’t change!) and so IMHO is not a particularly useful indication of your physical condition: muscle weighs more than fat.

Here’s the best example I can give: earlier this year, I was in the best shape of my life. Qualified for US Nationals for the first time and was turning in PRs pretty much every time I went out to bike, swim or run. Then I spent two months working overseas during which time I’m sad to say both my workout schedule and my diet completely fell apart.

Getting back home, I nervously approached the (Withings) scale for the first time. Just looking in the mirror, I could tell I was out of shape. To my amazement, my weight had only increased by two pounds! But my BF had gone from 14% to 20% (per discussion above, these numbers might be high but the trend is clear).

It’s not about weight, it’s about body composition.