OP, Why are you weighing yourself so often? I thought if you were tracking your weight for health reasons you were suppose to weigh yourself about once a week or so.
OK, you pushed a button! I should maybe add this to the 'pet peeves' thread!
I've seen this recommendation by dietary 'experts' a million times - that you should weigh yourself only once a week. But those people are not
measurement experts (metrologists), and weighing is in the field of metrology. So they are out of their field of expertise.
A little background - most of my career was involved in measuring things. All kinds of things. To be efficient and effective, you need to know something about the thing you are measuring ('characterize' it), it might be stable, variable, drift with temperature or time (short and/or long term drifts), or be affected by the measuring instrument itself. And then you need to know the goal, do you need a quick & dirty estimate, or a very accurate, traceable measurement, etc.
So how does a metrologist tackle this? OK, so we know that a person's weight exhibits short term variations for many different reasons (all discussed in this thread). But someone on a diet is really interested in their longer term weight change. If I'm dealing with something that has short term drift, I certainly don't take fewer measurements,
I take more measurements, more often, and then apply some filtering tools to those results.
So let's look at this for weight loss. You might have a goal of losing one pound a week. And lets say you were meeting that goal. But we have seen that a single reading might easily vary by plus/minus two pounds due to all these short term variations:
Case A) Imagine our short term variations were just one pound on the light side on the first Wednesday, and the next Wednesday we were just one pound on the heavy side.
We would think we gained one pound, and likely would look into changing our diet further. But a daily measurement averaged would have washed out those variations, and shown that we were on goal to lose a pound a week, keep up this diet plan.
Case B) Imagine our short term variations were just one pound on the heavy side on the first Wednesday, and the next Wednesday we were just one pound on the light side.
We would think we lost three pounds, we might celebrate and pig-out, or ease up on our diet. But a daily measurement averaged would have washed out those variations, and told us to continue what we were doing.
We might make big changes to our diet plan, based on two very different 'results' from totally false information from 'noise' and too few measurements to filter out that noise.
If the reading fluctuates, take more readings, and apply some averaging functions. When I was tracking my weight, I did this. I weighed myself on the same scale every day at work, at the same time, under as closely repeatable conditions as I could. I'd throw out the high and the low for the week, and average the rest, and track each week. That smoothed the variations pretty well. Even w/o any math, just take the median each week (throw out highs and lows until just the middle number is remaining) would be pretty good.
Taking one reading a week is the worst advice one could give/get!
Just some technical bits about electronic scales - In my consultant days, I designed an electronic office scale for a well-known scale company. ....
Oh, about the care and feeding of digital load cell scales - Don't leave loads on them, don't store stuff on top of them, and keep them upright, don't stand them on their side for storage. And don't overload or shock-overload them. They will "appear" to be all right, but a permanent tiny deformation created in the load cell will always be there, and it will introduce a non-linearity that was never intended by the developers. A resulting fixed-offset can be "zeroed" away, but not the created non-linearity.
Thanks for that - I do tend to store our kitchen scale with some things on top. So I just changed that and stack it vertically.
-ERD50
