What is this 'Body Analysis Scale' actually doing?

Factual Questions
1

I recently purchased a Conair Model WW79 scale (very similar tothis model), which purportedly uses a small electrical current to measure ‘body fat, body water, bone mass and BMI’. However, I suspect that instead of measuring total body fat (and body fat %), the scale is actually calculating these values from a simple BMI based formula.

To marshall my arguments-

  1. The scale wants to know my age, which is commonly used in these BMI to body fat % calculations.

  2. Surveying a variety of body fat percentage pictures, such as the one on this page, I think I come in between the 15-20% pictures. The scale is commonly reporting a BF % of 25%.

  3. Using a variety of web based algorithms based on body measurements similar to this one I come up with estimates ranging from 14-19% body fat. Again, the scale is coming in at 25%.

  4. A common formula for estimating BF% based on BMI tracks the scale’s ‘measurements’ pretty well, but comes in about 1% low. Modifying the formula to-

       {[(1.5 x BMI)+(.075 x AGE)] -18}

-tracks the measurements more closely.

I know this all smacks of someone who’s carrying more fat than they want to admit, but I’m not really invested in the number beyond its accuracy.

So, did I get hoodooed into buying a scale that doesn’t do what it says it does? And further, if that BF % is incorrect, are the other numbers the scale purportedly measures (water %, bone %) any more accurate?

2

This electrical resistance method was never very accurate. The method you want to use is a water displacement or air displacement test. That’s a lot more accurate. Apparently there’s a way to x-ray and see the fat and muscle mass directly which is more accurate still.

3

I have a Tanita scale which is supposed to measure body fat %.
I can’t tell you how accurate it is (I’ve never had a water-immersion measurement made), but I will say that it is very precise.
I measure my weight and BF % every day, and record it. I can tell when I am putting on more fat than muscle just by looking in the mirror, and the scale accurately reflects that. When the scale said I was at 11%, I had significant visible resting vascularity in my forearms. Now the scale says I’m a few percent higher than that, and I agree with it.

So, they may not be clinically accurate, but these scales can give a very good trend indication, which is really what most people want.

4

Bioelectrical Impedance Analysis

As Habeed says, it’s not a particularly accurate method of determining body composition - however, as the article says…

“BIA is considered reasonably accurate for measuring groups, or for tracking body composition in an individual over a period of time, but is not considered sufficiently accurate for recording of single measurements of individuals.”

5

Agreed. That’s why I got the scale, but there’s a difference between ‘inaccurate’, ‘so inaccurate as to be useless’, and ‘not even doing what it claims to do.’

If it was off by a % point or three, I’d say the first; if it’s off by 11 % points, as some of the other estimations claim is possible, it would be the second. But the fact that it tracks so closely with a generic BMI conversion formula makes me suspect the third.

I would have sworn to you that the scale wasn’t even taking the impedance measurements at all, and was just doing BMI calculations and its derivatives, but I just broke out my digital multimeter and it is applying ~1 VAC and 300 uA (micro amps) during the measurement interval.

So at the very least, it is applying the electricity.

6

Try standing on the scale without touching the electrodes, or short them together. Find out if this changes the reading at all. Maybe the engineers in China found this circuit wasn’t working, so they just quietly commented out the part of the code that reads it and switched it to a BMI calculation.

7

Done. Unfortunately, the scale did not like nor appreciate this. It returned an error message for all of the impedance based measurements with the electrodes open or shorted together. Multiple attempts in different positions didn’t fool it.

I did, however, use my thumbs instead of my feet as the conductors, and all the measurements returned pretty much the same as if I had stood on the scale like a sane person.

Now I wonder if it isn’t just a simple filter, along the lines of-

Is it human?

If so, return BF/BF%/Water %/Bone % calculations.

Or perhaps I’m just off my nut.

8

They actually do work, but people when they post their reviews and tests on the internet often don’t get how they work or understand the limitations. I use the four electrode model (they might call it eight, but it is a scale you grab the handles and pull up when you stand up).

Tanita BC-558

I’ve used it almost every day since 2008 - so I’ve taken well over 2,000 readings with it.

  1. unless they are incompetent - they would have to be at least as accurate as BMI as they already are putting your BMI into the equation they are using

  2. it is measuring your hydration levels (more or less) - since body fluid moves around during the day - this especially impacts the leg to leg (or hand to hand models) - as your fluid will shift due to gravity. I believe Tanita recommends 3 hours and Omron recommends 2 - maybe it is the other way around after waking up before taking a measurement (personally I’d never buy a machine if it wasn’t from these two). They make respectable medical equipment used by hospitals all over the world and I wouldn’t trust a company that is known for hair dryers (I think I read they make good clippers though - so maybe they are experts in some field I don’t know).

So all of the equations that have gone into building the statistical model are based on the factors they list in their instructions - regarding water and food consumption, exercise, alcohol, waking up.

How many people really are still around their house three hours after waking up. Yes you can take it at other times during the day, but it is less likely to be precise at those times unless your activities are the same every day.

Anything you do that changes your hydration status will effect the reading. Want to see if the scale really does something - simple:

A) Wait three hours after having been in bed, done any type of what would be considered exercise, eating or drinking - Take a reading
B) Chug a 32 oz glass of water (you don’t have to chug it) - take another reading.

It should go down (bodyfat%) - and take another one 90 minutes later and 90 minutes later. By that time it will probably have returned to what it was before.

I’ve only done this test a few times, but it very obvious to me that it works. It absolutely detects hydration status. I’m doing some dietary experiments. The other day I took the most salt it have ever taken in my life - 8 grams - at once. I (probably you can guess) - got very thirsty - the next day when I stepped on the scale I had what was probably the biggest one day drop in body fat ever 4% (sometimes - maybe one out of 50 it messes up and gets one of the limbs wrong - it’s obvious when this occurs as the measurements are way off and don’t match the other limb - this wasn’t one of those times - and I’m not including those times.)

Some people would say this means it’s inaccurate - as obviously I didn’t lose 4% points of bodyfat in one day, but they are looking at it wrong. It can’t tell how much bodyfat you have - it’s not psychic. It is predicting the bodyfat based on an equation that takes into account your physical parameters and the impedance to get a pretty good idea of the water content/hydration level of your body. Since fat and muscle are different - they use fancy math - I’ve seen it - it involves trigonometry (wait are sin and cos trig?) and a bunch of stuff to come up with an estimate of your bodyfat.

I’ve had my bodyfat measure twice with what are considered more accurate measures - both the air plesmowhatever - and DEXA. In both cases I asked the operator/or my doctor to guess what they thought I’d be and both guessed like 3 points higher (geeze thanks - I don’t think my doctor was trying to be kind, but I think the bod pod guy actually rounded his real guess down) than my Tanita.

Both were within 1.5% of my Tanita - which generally varies about that much day to day, but I took extra care to make sure I was up the right amount of time on the testing days. I think I got a little lucky, but it still is pretty good - when I get the flu/cold/feel like shit - most times I’ll notice a bigger change in the scale - same with look at my face in the mirror and look bloated - almost ALWAYS a change.

I’ve done lots of analysis of my data and found that there aren’t any real obvious trends from day to day. I don’t make an effort to do it the same time every day - so I wouldn’t necessarily expect that. - I almost always take it before I shower - which might be 10 minutes or six hours.

Even on a weekly basis it is hard to see trends, but I absolutely see trends on a monthly basis.

Many people - especially when they are dieting - go through pretty drastic shifts in hydration. They are going to overwhelm the algorithm in looking consistent day to day, but over a long period of time I’ve been extremely happy with it. I can’t think of anything else that is voluntary ritual and takes a tiny bit of effort that I’ve done like 95% of days for 7 years.

I’m sure most people are less anal than I am and don’t want to do it everyday, but if kinda like it - and like that I have a pretty little graph that has been going down pretty steep in the last few months :slight_smile:

9

The bone calculations aren’t real - it is basing that off equations based off how much muscle it thinks you have which is correlated to bone density - which is related to TBW. I would take that with a large grain of salt. At NO TiME should you consider it a substitute for having your bone dentistry actually measured (if you are supposed to). It’s not actually measuring bone density and can only give broad population like estimates for it. Since bone is such a small part of your body it is going to be much less accurate on that then the other measures.

TBW basically determines most of the outputs - unless your model also has visceral fat which I think might actually be a separate measurement. I’ve been meaning to run a machine learning also on it to see if I can predict it 100% of the time just using the other data, but have been lazy.

10

If you are indeed doing a thumbstand on this device naked in your bathroom as I imagined when I read this, then you are certainly in fine enough shape to not need to worry about any of this.

11

DataX, according to the scale, my Water % almost never moves. I think it has been between 54.0% and 54.5% almost every time I’ve stepped on the thing.

I just tried the 32 oz trick. Results-

------------------Pre------------------Post
Weight-------202.6 lbs-------------205.3 lbs
Total Fat---------52 lbs--------------53.5 lbs
BF %-------------25.7 %-------------26.1 %
Water %---------54.4 %-------------54.1 %
Bone %-----------8.6 %---------------8.5 %
BMI---------------27.4-----------------27.8

So, I ‘gained’ 2.7 lbs and the scale slid 1.5 lbs of it directly into the fat column, despite the fact that -if I’m understanding you correctly, my BF % should have gone down. Also, my Water % went down, which is hilarious.

Also, before I did this, I messed around with the scale’s multiple user features, and told it that I was 18, 43, and 63 years old, while keeping my height and gender constant. For the record, I’m actually 33 years old.

It looks like, to adjust for age, it moved my Bone % up for younger and down for older, then added about half of that missing weight directly to the Total Fat column, except for the time I ‘aged’ 48 to 63, in which case every single category is exactly the same except for Bone %, which decreased from 8.0% to 7.4%. Where that extra weight went is a mystery, since my BF % remained the same.

Perhaps the scale just told me I’m full of shit?

For the Record, here are the rest of my stats (8 consecutive day average), in case anyone sees anything out of the ordinary that might influence the readings-

Gender-----Male
Age-----------33
Height--------72"
Weight-----200.8 lbs
Fat-----------50.4 lbs
BF%---------25.2%
BMI----------27.1
Water %-----54.3%
Bone %--------8.5%

Sorry about the coding.

Funny image, snfaulkner, but what I actually did was…you know what, print the legend on this one. Naked thumbstand it is.

12

I hope that isn’t bone % - it’s lbs - unless you are really dense :slight_smile:

And I think that con air might be cheating - here is my water for the last few days:

56.9
56.5
58.4
57.4
59.7
60.7
58.7
57.3
56.9
56.7
55.6

They don’t have to be “cheating” - they can just be putting more weight on the BMI part - do you think your figures have less variance than mine?

I’m using a $350 scale that does feet and legs at the same time. Not “bragging” (maybe stealth - I don’t know), but I would think mine should be less than yours.

13

DataX, I hope I haven’t misled you, but the scale in question is about 1 step removed from a cheap digital scale. Nothing, up-…errr…high-class about it.

And to answer your question, yes, my Water % has significantly less variance than yours. I haven’t been recording it, partially because it never moves out of the 54-55% range.

Also, the ‘Bone %’ measurement is displayed as a percentage, which leads me to believe it means that 8.5% of my weight is due to my skeleton.

I’m new to all these measurements, and, outside of BMI and BF%, I have no idea what a healthy/normal range or daily/weekly variance is for any of them.

14

It looks like they are trying to copy Tanitas display, but Tanita specifically marks it as lbs both on the scale and the data sheets.

The math only adds up for me if I do it as lbs (why they do it that way - I don’t know - they also give muscle mass in lbs)

But I don’t think yours displays muscle mass - so you can’t do it to check.

15

And if you check the sheet at the bottom you’ll see you are pretty close to what they would predict for you for bone mass
Eta - oh and two posts up I meant “arms and legs” not “feet and legs”!