Body fat scales vs science

New Topic
This topic has been archived, and won't accept reply postings.
 dread-i 15 Nov 2021

I have some scales that measure body fat, muscle mass etc. I stand on sensors, and it supposedly uses a small electric current to measure various metrics.

To test this I measured myself and my body fat was X and muscle mass was Y. I then tried it again holding a heavy weight (15kg). In theory, if it's using a current to measure these things, then my body fat would be the same but my weight would go up. That wasn't the case. Both body fat and muscle mass went up, which make be think its not working as advertised. I think its just using some basic maths to estimate these values.

Can anyone else replicate this ?

Can anyone recommend some scales that accurately measure body fat & muscle mass, that pass such a test. Bonus points if it comes with an app that doesn't require location information and access to your address book.

(Looking at amazon, there appears to be loads, with similar functions, all with the same app. Which makes me think they are all based around the same sensor.)

 chris_r 15 Nov 2021
In reply to dread-i:

I think the the body fat "measurement" is based upon electrical conductivity. With muscle vs fat having different conductivity. The electrical current will take the shortest route between the two points that you're standing on, up one leg and down the other. Holding a weight won't alter the resistance encountered by the electricity.

OP dread-i 15 Nov 2021
In reply to chris_r:

> I think the the body fat "measurement" is based upon electrical conductivity. With muscle vs fat having different conductivity. The electrical current will take the shortest route between the two points that you're standing on, up one leg and down the other. Holding a weight won't alter the resistance encountered by the electricity.

Which is my point. I'm not expecting them to be medically accurate, but to basically work as advertised would be nice.

 nathan79 15 Nov 2021
In reply to dread-i:

To be honest none of these scales can be relied on for those measurements. They just don't provide accurate readings. You'll get a more reliable measure buying a pair of calipers and learning to use them.

The only reading you can rely on is the bodyweight.

Post edited at 13:49
 jezb1 15 Nov 2021
In reply to dread-i:

I have a fairly expensive set, no idea how accurate the fat / muscle measurements are but I find them useful to show a trend.

In reply to dread-i:

I am similarly suspicious of my scales which are a £25 set from Amazon that say "kamtron" on it.

The weight seems very consistent and reliable, but I suspect the body composition details are just projected from the weight, possibly using the height, age and gender details that you put into the app.

I'm not complaining for a cheap set of scales they have lasted a couple of years already and link to the app well so I have a history of my weight fluctuations.

I would be interested to hear if anyone knows if there are some scales that do give scientifically measured accurate body composition details.

 chris_r 15 Nov 2021
In reply to dread-i:

My friend bought some scales recently. She and her partner both tried them and quickly realised they were broken as they were reporting both their weights to be 2 stone too heavy.

They returned them to Tesco, got their money back and bought a different, more expensive model.

Same weight reported.   

In reply to dread-i:

All based around similar sensors and all give a % based on the resistance between your feet.

The scales will assume a leg length based on your inputted height. They will apply an approximation to cross sectional area, assume a bone % then estimate a ratio of muscle:fat from the calculated resistivity of your legs and pelvis.

All this introduces significant errors. 50% of the population carry a pair of fatty lumps away from the path of the signal.

These scales are best used to monitor trends rather than give a value. As such they have questionable advantage over a pair of mechanical scales and simply working on weight. 

 geckoboy 15 Nov 2021
In reply to dread-i:

The scale will be calculating your body composition based on your weigh, impedance and height (depending on the scale). Change any of these and you will get different results.

These sorts of scales are only really any good for tracking a trend rather than looking at absolute readings.

 nniff 15 Nov 2021
In reply to dread-i:

Some years ago my wife joined a gym.  She was particularly lean at the time.  They put her on some of those scales which declared her body fat percentage to be too high.  She enquired which bit they thought was the problem.  They bolloxed away and she cancelled her new membership and joined a different gym.  We concluded that she has little fat toes, a fact that has been scientifically proven.....

 deepsoup 15 Nov 2021
In reply to dread-i:

> I think its just using some basic maths to estimate these values.

And a lot of assumptions - reasonably well educated guesses based on averages that may or may not be a good fit for your body.  Essentially using the electrical measurement, weight, height, age etc.. to look up a 'best guess' in a table of most likely results.

It doesn't matter whether they're cheap scales or expensive ones, it just isn't possible to measure body fat percentages accurately via electrical contacts on the soles of your feet.

In reply to dread-i:

I have some old Withings scales which produce similarly nonsense composition figures, although I like them because they store my weight measurements in the cloud. I’ve also tried weighing myself whilst wearing socks to see if the scales actually measure resistance - same result. And no, I don’t wear Bacofoil socks. Most days. 

 deepsoup 15 Nov 2021
In reply to mountain.martin:

> I would be interested to hear if anyone knows if there are some scales that do give scientifically measured accurate body composition details.

Nope, there's no such thing.  Though the measurements don't necessarily have to be all that accurate to indicate a trend as things change. 

Calipers work pretty well in the hands of someone who knows how to use them, otherwise you may as well just use a tailor's tape measure to compare the circumference of your waist with your neck.  Or, if you can be objective about it (which most of us struggle with tbh) - just make an honest appraisal based on stripping off and looking in a full-length mirror.

If you want a proper 'scientific' measurement you'd be looking at a DEXA scan or maybe even an MRI.  Even the methods that measure the density of your whole body (floatation tanks and those 'airpod' things) involve a bit of educated guesswork about the ratio of fat to lean muscle, bone density etc.

 ianstevens 15 Nov 2021
In reply to deepsoup:

> Nope, there's no such thing.  Though the measurements don't necessarily have to be all that accurate to indicate a trend as things change. 

> Calipers work pretty well in the hands of someone who knows how to use them, otherwise you may as well just use a tailor's tape measure to compare the circumference of your waist with your neck.  Or, if you can be objective about it (which most of us struggle with tbh) - just make an honest appraisal based on stripping off and looking in a full-length mirror.

This. Calipers are often quoted as the tool to use in these threads, but require skill to use properly beyond what you can learn on the internet. Better to have consistent error and worry about trends than inconsistent error and hence meaningless numbers.

OP dread-i 15 Nov 2021
In reply to deepsoup:

>Essentially using the electrical measurement, weight, height, age etc.. to look up a 'best guess' in a table of most likely results.

Thinking about it further, the impedance of a short fat persons legs might be similar to those of a tall skinny person. I guess that the calculation uses weight as an additional factor. One enters age and height as part of the set up process, so in theory, it shouldn't need weight. I guess there is only so much logic you can cram onboard a cheapo microcontroller.

In reply to Presley Whippet:

> 50% of the population carry a pair of fatty lumps away from the path of the signal.

I'll have you know my nadgers aren't fatty lumps...

 deepsoup 15 Nov 2021
In reply to dread-i:

I don't think it's quite as simple as impedance, though I'm not entirely sure - I thiiink there might be something going on to do with a frequency response as well.

> One enters age and height as part of the set up process, so in theory, it shouldn't need weight. I guess there is only so much logic you can cram onboard a cheapo microcontroller.

I'm not certain but don't think the processing power of the machine is really the issue, it just isn't possible to resolve an accurate measurement out of the data.  Otherwise an expensive scale would be significantly better than a cheap one, but I'm pretty sure that isn't the case.  (Not even the really expensive ones that have handles with electrical contacts for your hands as well.)

Also, if it shouldn't need weight then you have to wonder why age would be relevant. 
(The answer is that it's one of the criteria being used to look up a 'best guess' in a set of pre-programmed tables, along with height, weight, sex and presumably the electrical measurement as well.) 

Try lying to it about your age - that shouldn't really have any effect on the result if the number on the screen is determined solely by measurements of your body.

 minimike 15 Nov 2021
In reply to dread-i:

They work by combining your impedance and weight to calculate fat % according to a formula. If you manipulate your weight with some external mass (made of whatever) the equation it’s based on won’t apply so it will produce garbage. It wasn’t designed with the possibility that someone might be holding some heavy objects! 

In reply to dread-i:

Try measuring you body fat percentage before and after filling your underpants with lard. If it’s working correctly you should see an increase in bf%. Do this when there’s nobody else in the house. 

Post edited at 17:37
 flatlandrich 15 Nov 2021
In reply to dread-i:

I have a rather old set for body fat percentage and I couldn't tell you exactly how they work but they are hugely affected by hydration levels.  Apparently a night on the lash and a kebab means 5% less body fat the next morning! Who knew you could drink yourself slim?  

 sbc23 15 Nov 2021
In reply to dread-i:

It’s all just a fudge.

we have some expensive Fitbit aria 2 scales that connect to the Fitbit app. Recently, my wife & I have approached the same weight. She is considerably shorter than me and higher body fat %. 

When I stand on the scale, it doesn’t know which of us it is so it asks. It gives totally different percentages dependant on who you tell it. By 20%.

The system knows our heights and sex, so I assume it just uses that via some look-up table and gives you a %.

Waste of time, although the actual kg logging and graphs are good.

1
 summo 15 Nov 2021
In reply to dread-i:

Just get a mirror, they don't lie (unless at the amusement parks or shop changing rooms).

2
 flatlandrich 15 Nov 2021
In reply to dread-i:

My scales only do body fat but I thought I'd try your experiment.  They are pre-programmed with my height, age and sex. 

81.8kg - 23% bf

Now whilst wearing a rucksack full of climbing gear.

92.7kg - 30% bf 

So with the increased weight and the same electrical resistance they are calculating an increased body fat percentage. Putting all the figures in to some online calculators I get roughly the same results (+/- 2%). So broadly speaking they seem to work, although, as others have said, probably best used to track trends than be totally accurate. 

2
 FactorXXX 15 Nov 2021
In reply to deepsoup:

>   Or, if you can be objective about it (which most of us struggle with tbh) - just make an honest appraisal based on stripping off and looking in a full-length mirror.

I think I must be anorexic because every time I look in the mirror I see a fat bastard...

 CantClimbTom 15 Nov 2021
In reply to dread-i:

There's 2 ways to look at these scales

1) see that the numbers are woefully inaccurate, declare them BS 

2) realise they are woefully inaccurate but the same person on same scales over time can show a trend, and they can be useful to monitor trend even though the % fat etc numbers aren't reliable

 Martin W 15 Nov 2021
In reply to Presley Whippet:

> All this introduces significant errors. 50% of the population carry a pair of fatty lumps away from the path of the signal.

Both the body fat scales I've used have required you to enter the gender of each user as well as their height and age, so the machine should be taking that in to account somehow (otherwise why bother requiring it?)

I was quite worried for a while when my current body fat scales started telling me that my body fat had suddenly gone up two whole percentage points.  After a while it occurred to me to check the settings and I discovered that it had somehow decided that I was female rather than male.  Changed that back and my body fat reading was back more or less where I expected to be.

But, to echo the other comments above, these devices do seem to use a lookup table to give a rough estimate, rather than any kind of clever calculation using the numbers to give an answer with any kind of accuracy or precision.  Within the range within which my body weight typically varies these days the algorithm seems to report almost exactly 0.1% body fat more or less for each 1lb variation in weight.  So yes, useful for monitoring trends (though I could do that one in my head), but I don't expect it to be an accurate figure (then again, it seems it's not all that easy to get a truly accurate figure by any other means either, at least not without the sort of equipment that is only found in hospitals and expensive clinics).

 Dan Arkle 15 Nov 2021
In reply to dread-i:

One of the most accurate ways to measure body fat is via displacement to find the body volume. 

This can be easily and simply performed at home by filling a large bath to the brim and lowering yourself into it until totally below the surface. Fatter people may need a weight belt to sink correctly.

The volume of displaced water can be simply measured by mopping it up in the room below. 

 Toerag 15 Nov 2021
In reply to deepsoup:

> I don't think it's quite as simple as impedance, though I'm not entirely sure - I thiiink there might be something going on to do with a frequency response as well.

...and maybe capacitance?  Impedance won't really take into account anything other than the shortest path, whereas capacitance will.  The chest and heart people come round to our work every other year with some similar scales (Tanita brand)  that give you a printout of things like muscle mass, visceral fat, normal fat and (best of all), 'metabolic age' There's always a lot of anticipation anyone comes back to the office to see how metabolically old they are compared to their real age.

 Tallie 15 Nov 2021
In reply to dread-i:

This seems to be a good summary: 

https://rippedbody.com/how-calculate-body-fat-percentage/

TLDR: US Navy body fat method is free, as good as most and significantly better than scales. 

 yorkshire_lad2 16 Nov 2021
In reply to dread-i:

As others have said, I don't get too fussed about the individual readings, but watch any changes in trend.

I also like the statistic that it comes up with about "Metabolic age" (whatever that is) that has me 16 years younger than I am: flattery gets you everywhere

Scales: Tanita RD-953

The app gives read-out for:

  • Weght
  • BMI
  • Body Fat
  • Muscle Mass
  • Muscle Quality
  • Physique
  • Bone Mass
  • Viscreal Fat
  • BMR
  • Metabolic Age
  • Water content

https://tanita.eu/rd-953/

(I have no connection with Tanita other than as a customer)

 mutt 16 Nov 2021
In reply to dread-i:

There are several methods of measuring body fat. All of them do not perform a full body scan so they won't give you anything approaching the right numbers. 

Why did you buy fat weighing scales in the first place? If it's for a science project then just accept the limitations and follow a procedure. 

If on the other hand if you are trying to loose weight then what you eat and how you excercise are what you need to measure. Wind excercise up slowly and cut down on harmful food. In time you will hopefully feel that you no longer have to measure how much fat you have and harmony will have been achieved..

OP dread-i 16 Nov 2021
In reply to mutt:

>If on the other hand if you are trying to loose weight then what you eat and how you excercise are what you need to measure. Wind excercise up slowly and cut down on harmful food. In time you will hopefully feel that you no longer have to measure how much fat you have and harmony will have been achieved..

I'm trying to burn off pies, before I put them on again at Xmas. I'm also trying to avoid loosing muscle. I'm increasing protein whilst reducing calories. I am keen to understand if the weight I'm loosing is actually lard and not my six pack <cough>.

OP dread-i 16 Nov 2021
In reply to yorkshire_lad2:

>The app gives read-out for: <snip>

>Muscle Quality

>Physique

Its things like these that confuse me. What is muscle quality? Strength or endurance, probably cant be measured by standing on a stationary object. Also physique, unless you are also measuring from arms across the chest and down to the feet, it might be a one sided view.

It seems there is a whole world of curious stats available. I suppose if you're tracking trends its all good.

OP dread-i 16 Nov 2021

In reply to Shani:

> To preserve muscle mass in a caloric deficit you need to increase protein AND train heavy. 

That might be specific to body builders, rather than endurance. I've not increased my training load. I'm running the same sort of distances and at the same sort of speeds. So not sedentary; maintaining. But, I don't want to side track the thread.

 mutt 16 Nov 2021
In reply to dread-i:

Is your heavy weight electrically conductive. If so (i.e.it is metal and not plastic wrapped)  then it will play havock with your conductivity. From memory the resistance of you body is 50ohms. The weight will provide a short circuit of much lower resistance 

Plus you should be pumping that weight if you want to loose fat and improve muscle condition.

Post edited at 14:10
 deepsoup 16 Nov 2021
In reply to dread-i:

> It seems there is a whole world of curious stats available.

Those Tanita scales may as well chuck in a hair lustre coefficient, an eye twinkle quotient, a few biorhythms and a horoscope for what most of those are worth.

It's interesting that the eminently sensible seeming advice linked to by Tallie above suggests not using these things to track trends.  Instead he says you should use some method (most likely a simple tape measure - and it's interesting that he rates that as more accurate than even a DEXA scan) to determine body fat percentage to see what you're dealing with, but routinely track trends only by directly measuring things you can directly measure.  (Weight, waist size, the actual size of your muscles if you're looking to bulk up etc..)

 streapadair 16 Nov 2021
In reply to yorkshire_lad2:

Yah boo, I've got a set of scales that gives 13 metrics. The catchy name NCKNCIZ is on it. I think the kg reading is fairly accurate, not sure about the rest, but I do like the Body Age metric which takes 10 years away.

 yorkshire_lad2 17 Nov 2021
In reply to dread-i:

> >The app gives read-out for: <snip>

> >Muscle Quality

> >Physique

> Its things like these that confuse me. What is muscle quality? Strength or endurance, probably cant be measured by standing on a stationary object. Also physique, unless you are also measuring from arms across the chest and down to the feet, it might be a one sided view.

> It seems there is a whole world of curious stats available. I suppose if you're tracking trends its all good.

Manual at:

https://tanita.eu/media/wysiwyg/manuals/home-use-body-composition-monitors/...

gives some explanation (but the words "snake" and "oil" come to mind!)

 Yanis Nayu 17 Nov 2021
In reply to streapadair:

I borrowed a set of those once that told me I was 18 when I was about 48 - loved them!

OrangeBob 18 Nov 2021
In reply to dread-i:

Can you pinch more than an inch?


New Topic
This topic has been archived, and won't accept reply postings.
Loading Notifications...