Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

thebassoonist

macrumors 6502a
Original poster
Sep 25, 2007
500
1
Davis, CA
Hello!

I'm currently brainstorming ideas for a statistics paper. The only requirements is that it must have three, and only three, variables. I'll be doing cross-tabulations (possibly standardized cross-tabs as well) and chi2 tests. One idea I have is how educated a person is and how often they pray and how that effects confidence in the scientific community. However, I can't show causality, because perhaps people who have little confidence in the scientific community are both less likely to get graduate degrees and more likely to pray.
I'm using data from the GSS, so there are a lot of choices for variables. If you have any ideas (don't worry about them being in the GSS, I'll take care of that, obviously) that would be interesting, I'd love to hear them! Any ideas! Any ideas at all! Political, Social, Family of Origin, Gender, Education, Gun Laws, Abortion Views, etc.
 
However, I can't show causality, because perhaps people who have little confidence in the scientific community are both less likely to get graduate degrees and more likely to pray.

No, you can't show causality because you're doing an observational and not an experimental study. This is a very important concept to understand. No Experiment = No Causality.

Also, the bulk of your post is very difficult to comprehend, can you be more clear?
 
No, you can't show causality because you're doing an observational and not an experimental study. This is a very important concept to understand. No Experiment = No Causality.

Also, the bulk of your post is very difficult to comprehend, can you be more clear?

Never mind.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.