You get better answers when you have generate questions.

You generate better question when you know your outcome of interest, your $y$ value.

Here’s a silly example.

Question: What happens when you eat more slices of pizza?

That’s hard to answer. We don’t know what we’re outcome we’re trying to measure. What happens to your enjoyment? What happens to your weight? What happens to your inclination to make bad decisions — such as listening to cheesy K-Pop boy bands? Weight, enjoyment, inclinations — these are outcomes, $y$ values.

The $y$ value comes from an x-y graph. The $x$ axis is horizontal, left to right, and the $y$ axis is vertical, up and down. As $x$ increases or decreases, there’s an effect on $y$ — the outcome of interest.

There’s a difficulty here — how do you know if you’re focused on the right outcome?

That’s tough to answer. If we’re talking pizza, does one of the outcomes above matter more than the other? That seems like personal preference. If you’re in business, likely profit matters more than other outcomes, but maybe not. Context matters.

I’m not here to tell you what to do in each context. Instead I recommend that knowing the question isn’t enough if you don’t know the $y$ value behind the question.

Last modified on 2026-02-07