"Just be Honest With Yourself" isn't the advice you think it is!

This week's research is the reason I no longer fully believe that.

Armitage and Conner, and later McEachan and colleagues, ran large meta-analytic reviews looking at how well intentions predict behaviour but they split the data by how the behaviour was measured. Some studies asked people to self-report what they did. Others measured it objectively with attendance records, purchase logs, physical tests, actual counts, not someone's memory of their own week.

When behaviour was self-reported, intentions explained 26–31% of it, roughly in line with what I've already told you.

When behavior was measured objectively, that number dropped to 12–20%.

Read that again. The gap between what people say they did and what they actually did is big enough to roughly halve the predictive power of their own stated intentions. Not because people are lying, mostly. Because self-report is a memory and a story, and both are generous to the version of us that meant well.

This is the part that should worry anyone who's built their entire self-improvement system on the honor system. "Just track yourself honestly" assumes the tracking is accurate. The data says it systematically isn't and it isn't in one direction. People overestimate their own follow-through, consistently, across domains.

I don't say this to make you distrust yourself. I say it because it changes what actually needs to be built. A system that only checks in with your memory of the week is checking the wrong data source. You need something closer to the objective record, "did the thing actually happen?" verifiably, not "did it feel like it happened?"

That's not a minor design detail for Brick. It's close to the whole point.

-Abdurrahman Ishaq Muhammad