N Equals Many All articles
Quantified Self

Stop Testing One Tuesday: How to Actually Experiment on Yourself

N Equals Many
Stop Testing One Tuesday: How to Actually Experiment on Yourself

Here's my opinion, stated upfront: personal experimentation is a skill, and nobody teaches it. We teach people to track their steps, log their sleep, journal their meals. Then they stare at the dashboard and draw conclusions using methods that would get a first-year stats student laughed out of office hours. "I slept great the three nights I took magnesium" is not an experiment. It's a coincidence with a spreadsheet.

The good news is that fixing this takes about twenty minutes of setup. Let me walk you through a real one.

The Question: Does the Morning Walk Fix My Afternoon Slump?

A friend of mine — let's call her Dana — read that a 15-minute morning walk in sunlight resets your circadian rhythm and kills the 2 p.m. crash. She tried it for a week, felt amazing, told everyone. Then week two happened: three rough afternoons, and the walking routine quietly died.

Dana ran a sample size of five, with no baseline, no plan, and a conclusion she reversed within fourteen days. Let's redesign her study.

Step One: Define the Outcome Before You Touch the Routine

The most common self-experiment failure isn't bad data — it's a moving target. Dana's real question was about afternoon energy, but "energy" is a fog. So we pinned it down: rate your slump from 1 to 5 at 2:30 p.m., every workday, in a note on your phone. One number, same time, no negotiation. If you change how you measure mid-experiment, you've started a new experiment and burned the old one.

Pick something you can actually record. "Productivity" is unmeasurable. "Minutes of focused work before checking my phone" is measurable. The quantified self movement's biggest trap is tracking everything and concluding nothing.

Step Two: Build a Baseline (This Is the Control Group)

Before changing anything, Dana logged her 2:30 p.m. slump scores for two ordinary weeks. This is her control group, and it cost her nothing but attention. Without it, she has no idea what a normal Tuesday afternoon even looks like. Most people skip this step, which is how a good week gets mistaken for a breakthrough.

Your baseline should cover at least one full cycle of your life — a full work week, ideally two, because Mondays and Fridays are different animals.

Step Three: Randomize, Don't Alternate

Here's where self-experimenters fool themselves most reliably. The instinct is to alternate: walk Monday, skip Tuesday, walk Wednesday. But your weeks have structure. If you always walk on weekdays and skip on weekends, and weekends also involve sleeping in and brunch mimosas, you've confounded your walk with your entire lifestyle.

The fix is randomization. Before each two-week block, Dana flips a coin for each workday: walk day or regular day. Random assignment breaks the link between your intervention and every other pattern in your life. It's the same trick that makes clinical trials credible, and it scales down beautifully to a sample of one.

Step Four: How Many Tuesdays Do You Actually Need?

The honest answer: more than you want. For an effect as noisy as daily energy, you want roughly 15 to 20 days in each condition before you trust the comparison. That's four to five weeks per arm, or about two months for the full trial. Yes, that's annoying. It's also the price of not lying to yourself.

Why so many? Because daily outcomes are dominated by noise — a bad night's sleep, a stressful meeting, weather, whatever you ate. A single day's slump score tells you almost nothing. Averages over many days are where signal lives. If an intervention's effect is real but modest, three days of data will bury it under noise, and you'll wrongly conclude it doesn't work — the mirror image of Dana's original mistake.

If the effect is huge — you start a new routine and your afternoons transform immediately and permanently — you can conclude early. Big effects announce themselves. Small effects, which are most effects, need patience.

Step Five: Analyze Like a Grown-Up

When the trial ends, don't eyeball your best walking days against your worst regular days. That's cherry-picking, and your brain is a world-class cherry-picking machine — it will find the streak that confirms whatever you already believed. Instead, compute the average slump score for walk days and for regular days, and look at the gap. A difference of half a point on a 5-point scale, sustained across 40 days, is meaningful. A quarter-point gap might be noise.

And here's the step nobody does: write down your prediction before you start. If you're 100% sure the walk will work, why are you experimenting? Write the prediction down because in eight weeks, your memory will quietly rewrite itself to match whatever the data shows. The note on your phone doesn't lie; your recollection will.

When the Answer Is Boring

Dana ran the full eight weeks. The result: walking days averaged 2.4, regular days averaged 2.6. A real but tiny effect — nowhere near the transformation she remembered from her first glorious week. She kept the walks anyway, because a small, free, reliable improvement is still an improvement. But she stopped evangelizing.

That's the mature outcome most self-experiments should produce: a modest, honest answer instead of a dramatic false one. Most things you try will do almost nothing. Knowing which ones is the entire game.

Your commute, your coffee timing, your screen curfew — each one is a hypothesis awaiting a trial. Run it honestly. Flip the coin, collect the Tuesdays, and let the streaks you remember lose to the data you wrote down.

All Articles

Related Articles

Tonight's Dinner Is a Coin Flip: How Your Chain Restaurant Quietly Enrolls You in an Experiment

Tonight's Dinner Is a Coin Flip: How Your Chain Restaurant Quietly Enrolls You in an Experiment

You'd Never Ship a Feature to a Sample of One. So Why Is Your Sleep Schedule Any Different?

You'd Never Ship a Feature to a Sample of One. So Why Is Your Sleep Schedule Any Different?

You Were the Guinea Pig Before Breakfast: Inside the Experiments You Never Signed Up For

You Were the Guinea Pig Before Breakfast: Inside the Experiments You Never Signed Up For