The Pitfalls of Body Fat “Measurement”, The Final Chapter

If you've been reading my series on body fat testing, you have now learned that even the best techniques have larger error rates than most people realize.

Does this mean body fat testing is useless for individuals? No, but it does mean you should exercise caution when using these techniques, and be aware of the large error rates. Here are some important tips to keep in mind:

  1. Always remember that any number you get from a body fat test is a VERY rough estimate, and could be significantly off.  Thus, don't put too much faith in the specific number.
  2. Point #1 holds true even when measuring change over time.  I've heard many people say things like, "I lost 4 pounds of fat and gained 1 pound of lean muscle."  The numbers are never even close to being this precise.  Given the error rate for measuring change over time, there's a good chance those numbers are way off.
  3. Even the best techniques have a 4-5% error rate when measuring change over time.  This means, to accurately detect a decrease in body fat in most people, the body fat percentage needs to drop by a minimum of 4-5%.  This means you should take long periods of time between measurements.  I would say a bare minimum of 3 months, but 6 months is probably better.  I see too many people taking measurements as often as every 4 weeks or so.  That's too frequent and unreliable.
  4. Remember that fat-free mass and muscle are not the same thing.  So just because your fat-free mass increases, doesn't mean you had an increase in muscle.
  5. You don't need to have your body fat tested.  A combination of body weight and circumference measurements (like waist circumference) will give you a very good gauge of whether you're losing body fat.  If your circumference measurements are decreasing, you are likely losing fat.
  6. If you are going to try to track change in body fat over time, then I recommend hydrostatic weighing or skinfolds.  These methods have shown the best accuracy rates for measuring change over time.  However, these techniques are difficult to do with extremely obese people.  For extremely obese people, I recommend simple body weight and circumference measurements.
  7. When using skinfolds to track change over time, you don't even need to bother calculating a body fat percentage.  If the sum of your total skinfolds is decreasing, then you are likely losing body fat.
  8. Whatever technique you choose, keep the conditions as identical as possible between measurements.  That means having the same technician perform the measurement on you, using the same equipment, and at the same time of day.

I hope you enjoyed this series on body fat testing, and I hope it opened some eyes regarding the techniques that are out there.  Again, body fat testing isn't useless, but you do need to be careful in how you interpret the results.

25 Responses to “The Pitfalls of Body Fat “Measurement”, The Final Chapter

  • I disagree with #3. I have developed a measurement method that I perform on myself every month or so and it is very sensitive to changes in body density. I am talking about changes only here, not the absolute fat% value. I agree with you about the percentage fat being inaccurate and that number is of less value to me anyway.
    I measure my own weight hydrostatically. I practiced exhaling to my minimum residual lung volume, until it became very consistent (many sessions over a month long period). Then, I used empty bottles of various sizes as buoys and determine how big of a bottle I could be in “equilibrium” with. The bottle(s) and myself are both submerged and we are neither sinking or rising. I calibrated the buoyant forces of all my bottles at home with a good scale and I have one bottle that is graduated (marks) for fine tuning.
    I am determining my “unadjusted” body density and don’t bother to account for the residual air in my lungs, because I am interested only in the change, not the absolute. I assume that the residual lung volume, my bones, and organs stay the same, so the changes in underwater weight are due only to muscle and fat changes. The underwater weight is not affected by my state of hydration (because that water is neutrally buoyant), and I can guzzle a couple of pounds of water and repeat the test getting the same value for my underwater weight.
    As I have lost significant fat and gained muscle, my underwater weight has increased from 500 grams to 3000 grams. Yes, I can now sit on the bottom of the pool with three empty one liter bottles in my lap. My error for an individual test is less than +- 100 grams. I avoid days where I have any intestinal gas, because that throws it off significantly. I switched to rigid bottles (calibrated) because the light drinking bottles dent in some when they are underwater.
    I’ve been using this method for 8 months. During periods of very regular, heavy, weight training I can gain about a pound of muscle a month while my dry land weight stays constant. I am now in a period of lighter/inconsistent strength training and doing much endurance work (for races) and my underwater weight and dry land weight have both been constant for three months.
    To use this method, a person has to be very comfortable in the water, and have access to a (non-saltwater) pool. After expelling all your breath, staying underwater for more than 20 seconds is difficult and takes practice.

  • I want to thank you for saving me $49 !

    I had a DEXA scan done in May 2009 and was pretty disappointed to be told that I was 29.7% fat. The likely hood of my building 10 lbs of muscle are pretty slim! And it is highly unlikely that I will lose 10 pounds of PURE FAT and then maintain a BMI of 21. It was depressing to realize that a body fat percentage of nearly 30% (thus practically “obese”) was my lot in life.

    I’ve been toying with the idea of having another (since I think, MAYBE I’ve lost some fat in the past year?)… but am clueless about WHAT would I do with the information or data. Will I (realistically) change my eating or exercise bases on the results of another test? No. I lost 35ish pounds 6 years ago… and my eating and exercise are pretty much as “perfect” as they are going to get.

    Really, the only function another test would serve would be to tell me how “good” or “bad” I am. Or it would tell me how close I am to the American cultural “ideal”. Oh yeah… such an emotionally healthy reason to have a scan. 🙂

    Your blog on inaccurate retests was pretty much the final nail in the coffin and put the question to rest, “Should I get another test?”. Soooo… I’m putting that $49 towards this year’s season ski pass. Thanks!

    Great blog! OK if I put a link on my page? (a lot of Weight Watcher members read my stuff).

  • Tjun Kiat Teo
    9 years ago

    I am curious. Has there been any study that established sort of a working conversion between waist hip ratio and body fat percentage. For example a waist hip ratio of 0.8 for male is equivalent to ?certain level of bodyfat percentage

  • When you say 4-5% error rate, does that mean 4-5% BF or 4-5% of the total?

    For example I had DXA put me at 8.1% BF. Would it then be 7.6-8.5% (4-5% either way) OR

    3.1-13.1% BF?


    • I mean 4-5% BF.

      So If DXA has you at 8.1%, the true range would be 3.1 to 13.1 (although in that case, the lower end is highly unlikely).

  • I’m somewhat curious that you suggest that the only way to accurately tell your body fat percentage is to strip your body down and weigh each of the component parts. How then can you tell how much error there is in the methods you examined, considering there is no true baseline to compare it to?
    I’ve had a few DEXAs done over the past couple of years, and the total body weight the DEXA machine gave for me has always been less than 1% variation of the scales. For a tiny beam of x-ray to be that close, surely that means it must be reasonably accurate?
    The technician reminded me each time that I should try to maintain the same routine regarding diet, hydration, exercise and time of day for each scan to get the most accurate comparison. In particular he said the glycogen content of the muscles will vary depending on what exercise I do, and this can affect the results because of the sensitivity of the equipment.
    I know our local football team trainer says they measure skinfold before and after each game – I suspect if there can be a difference in skinfold measurement in a couple of hours of extreme physical activity, then the changes are unlikely to reflect overall body fat percentage?

  • What are your thoughts on ultrasound testing? They just use a wand similar to what they use for pregnant women on up to seven skin fold sites.

  • Thanks james

  • I have always sensed that weight loss measured on the scale tends to lag fat loss as inferred from changes in my body circumferences and skin fold thicknesses. The calipers are a useful tool. I would like to invent electronic underwear that gives me the answer more quickly and accurately.

    • Judith,

      Body weight is heavily affected by factors such as water retention, while body circumferences and skinfold thicknesses aren’t as heavily impacted by these. This is probably why you sometimes notice a lag between the two. Of course body circumferences and skinfolds can have their limitations as well, and it is critical that the same person is doing the measurement each time with an identical technique.

  • Thank you for the thorough, well-researched summary. This is the most valuable collection of information on the subject available.

    Regarding your advice to take measurements less frequently when the error rate is high, my intuition would tell me to do just the opposite. Whether a measurement has errors due to inconsistency or a consistent bias, I would expect that more frequent measurements, plotted on a chart, would be the best way to show the true trend, and allow you to make adjustments to your diet and exercise.

    What I do is use my body fat scale every day and plot the results, and then use calipers every week as another data point. The day to day ups and downs of the chart get smoothed out over time. The higher the sampling rate, the better!

    • Thanks for your comment, Rob. Yes, I can see what you are getting at with more frequent measurements, and taking the average of those measurements. This can be beneficial to people who see the “big picture” like you do. However, for people who get hung up on specific day-to-day numbers, this method could be detrimental psychologically. So it really depends upon the person.

  • Thank you James for your excellent article on bodyfat measurement methods.

    I used 2 DEXA scans on me in one week. Both on Mondays using the same routine each time(exersice, food, time of the exam and I believe hydration) and with the same machine operated by the same technician. My bodyweight in my first exam was 117kg with 180cm height and my BF% according to the DEXA scan was 42,1%. On the following Monday my measurements where: 114kg with no lean mass loss but with increased BF% 42,6% ! In this week I followed two vey popular weight loss systems. 1. The slow carb diet as described in the “4 hour body” book by Tim Ferris (protein from lean meat with legumes and vegetables) and 2.the p90x3 fitness video series from Tony Horton. My frustration and dissapointment was overwhelming after the results of my second bodyscan despite my first enthusiasm of the body weight loss.

    Should I change my Bf measurement to something simpler and cheaper like the calipers?

  • How do you reply to the researchers that have conducted extensive body composition research using a DXA and declaring it a “Gold Standard” means of assessing body composotion? There are dozens of journal articles that conclude DXA’s being a “Gold Standard” method (the vast majority are not industry sponsored).

  • I found this article relieving to read. In the last 6 weeks, I have lost about 4kg. However, I measure 7cm less on my waist and 5cm less in my hips. I do HIIT training as I am fit despite a 10kg weight gain in the last year. Weights are sacred to me because I value being strong to the point where I would rather be strong and a bit chubby than skinny and weak. I got a DEXA scan today and it claimed I lost up to 3kg of lean body mass and gained a whopping 3% body fat, this is despite the fact that I look slimmer, feel stronger and fitter, and measure less! The man asked me if I changed the way I eat. I said ‘yes, I no longer eat junk food.’ Then he told me it could be that due to having less carbs, fluids in my body were less, and hence the drop in lean muscle mass. Oh dear, what an insufferable disappointment. Your blog has cheered me up, and more importantly, when I become a PT in the near future, I will not be recommending DEXA scans to my clients. I will measure and look at their posture myself. Thank you for your insights, once again.

  • I forgot to add today was my second DEXA scan, the last one was 6 weeks ago, when I definitely had much more body fat and was not feeling as strong (the proof being I now lift heavier weights, squat deeper, run faster, have a lower RHR, etc. )

  • If you have an excellent server support plan, it will let you send
    a support ticket instantaneously and get fast response. A dedicated rented machine can be rented for operating complex web applications and for the multiple sites interconnected to your business.
    The laws have been changing and many portals that offer services are finding
    it harder and harder to get their word out there.

  • In point 5 you say that you don’t need to have your body fat tested. You say that body weight and circumference measurements are enough. I disagree with this. I think that knowing your absolute body fat percentage is very important. How else are you to know if you should lose (or gain) weight in the first place? And how else are you to know when you should stop losing (or gaining) weight?

    Consider the following scenario. A man has a BMI of 26. He thinks he’s muscular and lean–thus weight loss is not on his radar. But because of curiosity he decides to have his body fat tested anyways. The test says that he is 32% fat! This is highly useful information for him (assuming that it’s accurate). It tells him that he’s a lot fatter than he thought and that he needs to go on a diet.

    I can’t tell how many people I have run into who are like this. Most people are substantially delusional about their body fat levels.

    Going back to the scenario, let’s suppose that this man has been dieting for a while. How does he know when to stop? Again, we can’t rely on his intuition since he’s delusional about his physique. This is where body fat testing comes in again. An accurate body fat test would be invaluable in telling him when he has reached a healthy and aesthetic amount of body fat.

  • I have read all your articles. Fantastic!

  • I know this is an old post, but I just wanted to thank you for putting together such a well researched, informative article. I was trying to decide between getting a bod pod or spending the big bucks on DEXA to track my weight loss progress. Having read your article, I’ll be perfectly happy to have the same trainer at my gym check me each week. Now I know enough to capture the measurements as well as the final number. Knowing what I know now, I won’t get hung up on the “predicted” body-fat % number, I’ll be focusing on the trend line of the skinfold measurements. In any case, you helped me a bunch and saved me a pile of money. Thanks!

  • “And how else are you to know when you should stop losing (or gaining) weight?”

    How about a mirror? You mention an “aesthetic” amount of body fat, but judging by what you wrote, you don’t seem to know what that word means.

  • Omron body composition monitor shows 6% body fat on me which is impossible. My abs don’t show at the bottom nor do I look ripped. I looked at pictures similar to me and they were more near 15% body fat at least. I look great but I have seen better abs on prisoners war people than me ( I don’t recommend that way because their caloric restriction included malnourishment, looking good but unhealthy ). I am 5 foot 9 inches (69 inches) weighing in at 140 lbs. According to the government I can be as low as 126 lbs and still be right at a BMI of 20 which they consider healthy.
    Omron body fat monitor does not measure below 5% so it calls my body fat an error sometimes <5%. So based on what you said about compounding errors, I changed one input on the scale to get a more realistic measurement of 15% body fat. I reduced my true input height from 69 inches to 65 inches as an input for programing the Omron scale. Of course it was trial and error but this setting where I am 4 inches shorter virtually brought my numbers more in tune with reality.
    You may be asking why would I do this? The scale gives up if it thinks your body fat is below 5% and I want to use it!

Trackbacks & Pings

Leave a Reply to James Krieger Cancel reply

Your email address will not be published. Required fields are marked *