Two types of people probably lurk on this type of Internet site! The ones who hate statistics and for which numerical logic remains a mystery, or a fear.
And others who gladly seem to be answering other peoples queries! i m looking for those who have knowledge in this area and are ready to share it, i would so appreciate having a little assistance here.
I m conducting a student research , a kind of beginners one. i m using an internet survey. there is a quantitative aspect to my research; i gave out to my sample a questionnaire measuring self efficacy (confidence in doing something) and one measuring self care behaviors for a specific illness. Both questionnaires use Likert type scales.
Here are my questions;
1) if i want to study the relationship between self efficacy and self care, do you reckon that correlations is the one and only measure i could calculate or is there another test that could give me more information?
2) if i want to compare groups of people within my sample (ex, compare the results of those below 20 to those over 40 on the self efficacy dimension), is that a simple t test? when i type t test into google though, it comes up with so many different t tests (simple, multiplem etc)
3) if i want to control for variables like age, type of illness, etc, I can use multiple regression yes? even though they are categorical variables these, and the others are numerical and continuous?
and lastly, if my data for my self efficacy scores and self care scores happen to be skewed, can i still do a t test and multiple regression (which is the same as anova right?)
sorry for the beginners questions. i have some basic statistics going on, and sometimes i think i understand more, and then pfiou the next day its all gone and i start thinking too much again.
Thank you very much!!!!
Advertise on Talk Stats