Measuring Changes in Body Composition.

By March 21, 2017 April 22nd, 2019 No Comments


Most people interested in fitness are interested in how their body composition stacks up against others.

Essentially, body composition is divided into “fat mass” and “lean mass”. Fat mass is pretty self explanatory, and lean mass is everything else (for example muscles/organs). Body composition is often correlated with the risk of developing certain diseases, and obviously shapes our appearance. Lower body fat levels are favourable for most endurance sports simply because carrying less weight is more efficient. In physique sports, body composition is also important due to the impact body fat and muscle mass have on appearance.

There are multitude ways to measure body composition, each with varying degrees of accuracy. Here’s my take on the two most popular options for at-home measurement of body composition: bioelectrical impedance analysis (BIA) and skinfold measurements.

Bioelectrical Impedance Analysis (BIA)

In a nutshell, these devices (often similar looking to scales) measure the electrical impedance of the body to an electrical current. Fat is less conductive than lean tissue. This will result in a higher level of impedance, corresponding to a higher result for body fat percentage.

Unfortunately, many variables affect the accuracy of a BIA measure…

Differences in hydration status can have a large impact on the variability and accuracy of the measurement and predicted body composition. As can the length of fast leading to the measurement, recent food choices, and recent exercise/activity (2). Also, prediction equations used in BIA devices to predict body composition are based off select populations (eg overweight, young and active, elderly), which can introduce error if you do not fit within the exact framework of the population used in the study validating the piece of equipment. There is also a large difference between required predictive equations for different ethnicities (3).

There is a large amount of variation between how accurate different brands/models of BIA equipment are (2), sometimes errors can be as large as 8-9%. As such (if you care to buy a BIA device) I’d recommend doing your research as to which current models have been scientifically validated with different equations available for different populations (age / ethnicity / activity etc).

A huge difference in appearance exists between 6% and 15% body-fat (relative to comparing 30% and 40%), reducing the worth of a lot of BIA measures when getting extremely lean – they aren’t accurate enough.

If you are using a BIA device to measure your body fat percentage for bodybuilding reasons, the error between measures can really throw things out. Coming into competition, a male will usually want to be between 3-6% body fat. Measuring body fat to the accuracy of 1% is probably not a task for most standard household BIA devices, so picture analysis is likely more useful at detecting when body-fat levels are reaching the “ideals” rather than relying on BIA measures. For an in-season bodybuilder, differences in 1-3% in body-fat look a whole lot more obvious than a difference of 1-3% when at a higher percent; for example comparing 37 % to 40%.

Due to the large error margin, BIA is better to be used over a long period of time to look at trends in predicted bodyfat, rather than one-off bodyfat measurements.

Due to the large variation in quality of BIA products with many not having been scientifically validated, combined with the variation in measurements caused by differences explained above, I do not usually recommend the use of BIA to analyse body composition for at home use, unless you reach all the criteria:

  1. The device you use has been scientifically validated and you are using predictive equations relevant to you (age / ethnicity / activity etc).
  2. You are carefully controlling hydration status at each measure (to be the same each time).
  3. You fast for the same length of time leading into the measure.
  4. Recent exercise / activity levels are same between measures.


Bodyfat conversion tables or formulas based on skinfold measurements turn raw data from skinfold measures (decent information) into confounded data influenced by a predictive error (3). 

Even though I don’t recommend taking skin folds and converting the values into a body-fat % value, I think there is some merit in taking skin folds and using the raw data, as discussed in the following section.

What I recommend to track changes in body composition at low cost:

Does putting a number on body fat percentage really matter? If your goal is to lose fat and you’re losing fat, is putting a % to it really that meaningful? Sure, if you’re a competitive bodybuilder, you’re going to want to get within a certain range of body-fat (eg 3-6%), but measuring to that degree of accuracy is usually out of the picture.

Here are some of my recommendations for measuring changes in body composition. The key to all of these methods of measurements is that it is the trends over time that you should be worried about, not a singular measurement.

Raw data from skinfold measures

One recommendation (if you want some numerical data) is to take skinfold measurements. Just record the raw measures, don’t try and convert it into a percentage. To make comparisons between time points more accurate, ensure:

  1. The skinfolds are taken at the same time of day (preferably morning prior to eating).
  2. Each skinfold is measured 3+ times and either the mean (average of all the numbers) or median (middle number) is chosen.
  3. The same person takes the measures (preferably someone who knows what they are doing!).
  4. Ideally having eaten the same thing in the preceding meal (dinner the night before).

Pictures taken in similar conditions

Those being:

  1. At the same time of day (preferably morning prior to eating) this reduces the influence that the prior meal can have on bloating & hydration status.
  2. In the same room with the same lighting (curtains closed, lights on to reduce light variability).
  3. Ideally having eaten something in the preceding meal (dinner the night before).

Trends in body-weight combined with performance measurements

A downward trend of body-weight over a decent period of time (1 month+) will almost always indicate fat loss. If gym performance is maintaining or improving, you can be pretty sure that muscle loss has been minimised therefore it would have been reductions in fat mass that would have contributed mostly to the bodyweight change.


As interest in the fitness industry grows, so do the number of fitness-related products. A lot of these make products make false claims to entice a sale.

Body fat and body composition analysers, for at-home use, are also often overpriced for how useful they actually are. Many people have the desire to know their own body fat percentage however many popular “at-home” methods are truly inaccurate and unvalidated. Do your research before purchasing any of these products!

Share this article on Facebook & join in the conversation

Please contact me here if you are interested in the nutrition and training coaching services I provide!


(1) Heyward VH, Wagner DR: Body composition and ethnicity. Applied body composition assessment. Human Kinetics. 2004, 135-172.

(2) Pietiläinen, K. H., Kaye, S., Karmi, A., Suojanen, L., Rissanen, A., & Virtanen, K. A. (2013). Agreement of bioelectrical impedance with dual-energy X-ray absorptiometry and MRI to estimate changes in body fat, skeletal muscle and visceral fat during a 12-month weight loss intervention. British Journal of Nutrition, 109(10), 1910-1916.

(3) Wells, J. C. K., & Fewtrell, M. S. (2006). Measuring body composition. Archives of disease in childhood, 91(7), 612-617.

Banner image: By JKSJks111 ( work) [CC BY-SA 3.0 (], via Wikimedia Commons