I am doing a survey regarding people's awareness on fire safety.
In Part I, I gather information like gender (2 categories), education levels (5 categories) etc.
In Part II, I set 9 questions with likert scale from 1-5 to examine people's attitude towards fire safety.
I want to find the correlation between the attitude and the categorical variables in part I. What should I do?
QUANTITATIVE. Treat the Likert scale (1,2,3,4,5) as quantitative data. Treat the demographic data, such as gender and education as nominal data. OCrrelation may be calculated by using biserial correlation coefficient. See link below for paper with example:
Article Correlation Coefficient According to Data Classification
Paul Louangrath
Thank you. I am not statistic savvy but this article helps. What if the demographic data is not dichotomous such as education level divided to five levels. Is it a must to dichotomise the categorical variables first?
I agree with Paul, it can help the option to segment in 2 groups the VCs and make the bi-directional correlation
https://www.youtube.com/watch?v=rmDaE_9Prh4&t=262s
https://eddata.com/wp-content/uploads/2015/11/EDS_Point_Biserial.pdf
https://www.youtube.com/watch?v=4R4mZ2SSsw0
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire