My study is on electric vehicles and I want to test whether factors such as cost and range anxiety have significant effects and their degree of influence when analysing the intent to adopt.
I am first asking questions on age/gender/current engine and then using a likert scale for the factors. My final question is on how likely they are to adopt in % based on their final feeling of 'intent'.
Which regression model should I be using? Thanks.
Is anxiety your dependent variable? Multipel regression (linear) is ok.Then you get from the squared part correlation the % each factor contributes with to the total variance in your type of anxiety.
I think it is important to specify how you define "intent", because this will say if you need a linear regression (if this is a continuous variable) or logistic regression (if this is a yes/no variable). As to the multiple or hierarchical, I would go for hierarchical, including age, gender and current engine at Step 1 and then the factors at Step 2. Step 1 will answer your first question, and Step 2 will answer the second, controlling for age, gender and current engine.
The intent to adopt is a continuous variable, so I will use linear hierarchical regression. Thank you Georgios Vleioras and Béatrice Marianne Ewalds-Kvist for your help.
I agree - Multiple linear hierarchical regression sounds like a good step forward - ensure all assumptions are met :)
Good luck with your work!
Hi there, I am up to the analysis stage of this research but I was wondering whether I treat this as parametric or non-parametric for the tests as I have seen different answers. Samantha Curle Georgios Vleioras
Hi Thomas. When you write "I treat this as parametric or non-parametric" what do you mean with "this"? Stated differently, I am not sure I understand your question...
Sorry Georgios Vleioras I did not make myself clear. Do you know what statistical test I should use for Likert items for predictors and a continuous dependent variable? I am confused as to whether I am allowed to perform a multiple regression with 1-7 scored likert items? Some have argued that I can only treat Likert as ordinal and so cannot run a multiple linear regression
Béatrice Marianne Ewalds-Kvist Could you please offer your expertise on the above issue have? Thank you both for your time in advance.
If your sample is greater than 30 you call the linear regression parametric, even if the sample is less than 30 in this case you use the same method. You use the multiple regression analysis model 1 = linear
then you get the result and see if the model is significant.
After that you look at your output and you see which of your independent variables are significant. Then you look at your part correlation and square it, Then you wrire
age group contributed uniquely and significantly with 11% (example) to the total variance in R square (R2) although gender did not reach significance (p
- Badges
- Science topic
- Similar topics
- Psychology
- Emotion