Bloodeyes
Arcane
- Joined
- Jan 30, 2007
- Messages
- 2,908
I like doing surveys on the codex. It's fun because it's effortless and quick, and you usually get quite a few responses to whatever question is on your mind. Obviously there are going to be significant cohort effects when only codexers can reply to polls, but it's still fun and discovering those cohort effects is actually half the fun. You never know what the hivemind is going to come out with.
The problem is you can't give full surveys because you're limited to a single question. If I wanted to, for example, give the 21 question depression anxiety stress scale to the codex I couldn't do it. I'd have to link to an external site and ask people to post their results in the thread, which wouldn't really work.
Like my current poll on dairy is just limited to milk and treats milk drinking as a binary variable (which it clearly isn't). How much cooler would it be if I could have had a section on milk consumption, a section on health beliefs relating to milk and a DASS 21. Then I could have done regression (using software lol not manually) to see if there were significant correlations between codexers milk consumption and their scores on the DASS scale and if this effect was mediated by health beliefs. Much more interesting than a single question of if y'all drink milk or not.
We need a survey option that allows for multiple responses and reports results not just in a bar graph but also in a spreadsheet of the raw data (anonymized of course). Like you wouldn't want to display the raw data on the thread page because most people wouldn't care, but if you had a button that generated a downloadable .txt or something that would do. Then I could plug that into jamovi and actually report informative results. It would be fun.
The problem is you can't give full surveys because you're limited to a single question. If I wanted to, for example, give the 21 question depression anxiety stress scale to the codex I couldn't do it. I'd have to link to an external site and ask people to post their results in the thread, which wouldn't really work.
Like my current poll on dairy is just limited to milk and treats milk drinking as a binary variable (which it clearly isn't). How much cooler would it be if I could have had a section on milk consumption, a section on health beliefs relating to milk and a DASS 21. Then I could have done regression (using software lol not manually) to see if there were significant correlations between codexers milk consumption and their scores on the DASS scale and if this effect was mediated by health beliefs. Much more interesting than a single question of if y'all drink milk or not.
We need a survey option that allows for multiple responses and reports results not just in a bar graph but also in a spreadsheet of the raw data (anonymized of course). Like you wouldn't want to display the raw data on the thread page because most people wouldn't care, but if you had a button that generated a downloadable .txt or something that would do. Then I could plug that into jamovi and actually report informative results. It would be fun.