Conversation

I asked ppl "how poly/monog are you", on a scale from 0 (very monog) to 5 (very poly) should i present the data granularly, or would it be easier to just have "monog people" (0-1-2) and "poly people (3-4-5) grouped as two groups? Or "monog (0-1), "medium" (2-3), and poly (4-5)?

Aella

@Aella_Girl

granular is usually better, but the v poly bin sizes start to get a bit low, and i'm not sure if the difference between ppl marking "mostly poly" and "fully poly" is worth distinguishing?

10:30 PM · Jan 10, 2023

25.7K

Views

Replying to

It might not be, but if it wasn't worth distinguishing, then why ask it in a way that distinguishes?

Replying to

granular

Replying to

you shouldn't adjust your bin sizes to make your point more salient, that's bad science

Replying to

unless "semi-poly" actually means regular guy who would fuck around if given the option? a certain c'est threshold must be crossed to be included

Replying to

of course granular. also, reminder - can you please give us descriptions for each column of your recently released dataset (from the relationship survey)?

Replying to

Why the 0-5 scale?

Replying to

I like monog, medium, poly

Replying to

A cumulative histogram might work well. Plot number of people who said at most x (instead of exactly x) for each x. It's just as granular, but it sidesteps some visual issues arising from small bins. It contains the 0:2 vs 3:5 case as the bar for "at most 2".

Replying to

You can use a stacked bar chart to show both, ie you would bin 0-1, 2-3, 4-5 where each bin is a bar and the bar is split into segments representing the individual responses

Replying to

You should run a binary survey as well - curious if your vins compare well to self identified bins