Comparison

Asks the same question of several things at once, so the column where the answers differ becomes the reason.

You have three ways to practice a language and a firm opinion about each one. The app feels productive. The podcast fits your commute. The tutor costs money and makes you nervous. You have never once asked the same question of all three, so what you are actually weighing is how each one feels when you think about it, and the winner is whichever one you thought about last.

Comparison puts several things in one frame and asks each of them an identical question. That identical question is the whole mechanism. Once every option has an answer in the same column, differences stop being impressions and become positions you can point at. Most columns turn out to say nearly the same thing about everything, which is itself useful, because it tells you that criterion is not what is deciding. What you are hunting for is the column where the answers split. That column is the reason, and until you draw the grid you are guessing which one it is.

It has six common renderers. A table puts things down the side and criteria across the top. A matrix crosses one list with another and puts something in every intersection. A T-chart splits one thing into two columns, usually for and against. A quadrant turns two criteria into axes and sorts items into four kinds. A deviation chart measures everything against one chosen reference, so you read signed distance instead of raw size. Side by side repeats the same drawing with one thing changed.

Reach for it when you are choosing among a few things, when an argument keeps shifting its criteria mid-sentence, or when you suspect you have been comparing yourself to the wrong reference. It is weak wherever the interesting thing is movement. A comparison holds everything still, so it cannot show that one option is improving while another decays, and it cannot show why. It is also honest only if you fixed the criteria before you looked at the answers.