The Experimental Apparatus

post by Elo · 2018-07-26T22:16:35.782Z · score: 9 (2 votes) · LW · GW · 2 comments

Previous: The problem of other minds (Greaterwrong [LW · GW])

Greaterwrong // bearlamp (my blog)

Science really likes to isolate variables. From first grade, dropping a ball to measure gravity experiment. We talk about the wind resistance on the ball, and how it might affect a measurement. So we repeat the experiment for reliability. Assuming if we have slightly different wind, and we average out the noise we can probably work out the answer on average by getting repeatability.

The “astute” of us will think about who drops the ball, how the ball drops and how we can drop the ball in an impartial way so as to not affect the experiment ourselves when we run it.

This process presumes there is some kind of setup that can isolate us from the equation when we drop the ball. And we can probably drop the ball in such a way that it isn’t too important for the results of the experiment.

The “way too astute for their own good” will start to think about local density effects of having humans around and moving and how humans will still influence the experiment with our teeny gravitational pulls even if it’s irrelevant for the most part.

It’s easier for inert experiments to isolate a variable - It’s just a ball and just gravity. We take this principle from the ball and gravity and we apply it to Psychology. If we can just stand back and watch people. We can do psychology research without disturbing them.

Well actually we know that we have trouble with repeatability of experiments. It’s remarkably hard to do an experiment the same every time. And we struggle with it and we wrestle with it. And we isolate enough variables that eventually we believe we can predict “how often people cheat given certain conditions” (Dan Ariely studies human biases and how certain conditions influence people to cheat, he’s great fun to read but that’s not so important right now).

Somehow there’s this trouble with isolating variables. All of science struggles with it. All of model making struggles with it. Einstein struggled with it. Any time we try to learn we struggle with it. We wonder - Was that the gravity on the ball that I’m trying to isolate and measure, or was that just the wind on the ball...

Next: The Feedback Problem [LW · GW]

2 comments

Comments sorted by top scores.

comment by gjm · 2018-07-26T23:17:13.606Z · score: 8 (5 votes) · LW · GW

"Previous" link is to bearlamp post, not to LW version of it (which seems like it should be the target).

I have to say it doesn't seem like there's a lot of content here. Yes, doing science on people is hard because people are importantly unlike perfectly smooth hard billiard balls, and actually even doing science on perfectly smooth hard billiard balls can be pretty tricky. This ... isn't exactly news. I guess this is part of a sequence of posts that will in due time say some more surprising things; it might be worth either moving faster or (if that's a bad idea for the readership of your blog) making some adaptations (e.g., batching up a few bearlamp posts and condensing them for LW).

Nitpick: The title currently has the non-word "Aparatus" in it; it should be "Apparatus".

comment by Elo · 2018-07-27T02:28:45.797Z · score: 4 (2 votes) · LW · GW

Nitpick fixed. Previous links being finicky but also fixed.

Content: foundation of further developments. Concern that if I move faster I will lose people.