Measure Twice, Intervene Once

Gary Wolf, writing about how he and Kevin Kelly founded the Quantified Self Movement:

"In 2007 we began looking at some new practices that seemed, loosely, to belong together: life logging, personal genomics, location tracking, biometrics … some of [this] was coming from “inside” as our friends and acquaintances tried to learn new things about themselves”

While I feel that Quantified Self (QS) has evolved beyond the practices mentioned above, the key idea of experimenting or measuring something in order to learn (or alter) things about yourself hasn’t changed. It doesn’t even need to involve technology in the modern sense; using pen and paper instead of a device to learn something new is perfectly valid.

A process for personal 'experiments'

Goal setting, or setting a target to change something about yourself, generally involves a ‘before’ measurement and an ‘after’ measurement - the change may be impossible to detect otherwise. For example:

I weighed 200 lbs - I changed how I eat/exercise - now I weigh 175 lbs
I smoked 10 cigarettes a day - I used patches - now I don’t smoke at all

If I didn’t weigh myself or measure my waist circumference or take some other ‘before’ measurement, how would I know if it worked? Even if I don’t write or record anything on paper but can see visually that I’m slimmer, I’m still doing a comparison: my memory of how I looked before and after - but hard numbers are always better.

There is an element of scientific experimentation to this - taking a single person (you!) and ‘experimenting’ by making a change to achieve a goal. This process of experimenting could be described as follows:

Set your Goal and what you’re going to change to achieve it.
Take ‘Before’ Measurement(s).
Make the change (’intervention’) - exercise 3 times a week, for example.
Take ‘After’ Measurement(s).
Compare your results to the Goal you set.
Repeat/Stop/Adjust the ‘intervention’ depending on success.

From making a single change like this, you can then see if the change brought you closer to or achieved your goal. From there, you can decide if you want to do more of it, less of it, something else, or forget about it all together!

This doesn’t mean that you don’t take measurements during the change, and only before and after. If something isn’t working for you, that doesn’t mean you keep going until the end. If a change is having the wrong or opposite effect and you need to stop, then do!

The Science-y bit

This concept of a ‘before/after’ study is part of the area of experimental design. While scientists are generally studying changes at a population level and would use other types of study design, it’s suitable (and often the only choice) for use on a personal level.

Other types of studies include randomised trials with control groups, where the individuals taking part are randomly made part of the control group (they don’t do the thing being tested) and the intervention group (they do the thing being tested). However, you can’t run a trial where you’re in both the control and intervention group at the same time. Your best choice is a n of 1 trial.

How do I know what worked?

One final thing - it can be hard to determine the exact cause of the change. For example, take the weight loss example above -

I weighed 200 lbs - I changed how I eat/exercise - now I weigh 175 lbs

In this example, it looks like 2 changes were made, eating less and exercising more. Which one caused the weight loss? Both? Did one have more of an effect than the other? Why? Does it even matter?

There can be an urge sometimes to figure out exactly why something happened, or why it didn’t happen. It isn't always possible to get an exact answer like that.

Using the example above, I don’t think it’s too important to say ‘it was 70% diet and 30% exercise’. Sure, if you’re interested, go ahead (it’s hard, if not impossible, though!). But just identifying a process that worked, in this case, some combination of exercise and diet change, is enough information for me, and probably for you too.