Hello,
I applied a questionnaire using the Likert scale of 11 points (from 0 to 10). Now, I thought of forming three groups: (i) with an average below 5 (Bad); (ii) with an average above 7 (Good); and (iii) intermediate group, with an average between 5 and 7. What do you think? I need references to support this classification.
Thanks
Matheus Da Costa Gomes Will be good if you could provide details on sample size and and if you will be using AMOS or PLS for the analysis.
Hello Matheus,
actually one should not split continuous data. If you think this variable is a moderator, better run moderation analysis with a product term.
Dimitruk, P., Schermelleh-Engel, K., Kelava, A., & Moosbrugger, H. (2007). Challenges in nonlinear structural equation modeling. Methodology, 3(3), 100-114.
MacCallum, R. C., Zhang, S., Preacher, K. J., & Rucker, D. D. (2002). On the practice of dichotomization of quantitative variables. Psychological Methods, 7(1), 19-40.
Maxwell, S. E., & Delaney, H. (1993). Bivariate median splits and spurious statistical significance. Psychological Bulletin, 113(1), 181-190.
All the best,
--Holger
Pradeep Paraman and Holger Steinmetz Thank you for your support.
I am estimating a panel event study (total sample = 2.515 firms). The model addresses the impact of credit policies on corporate investments. There are companies that are impacted at different times (n = 1.536 firms) and also companies that are never impacted (pure control group = 979 firms). However, there is a variable obtained using a Likert scale with an important characteristic for the analysis because it indicates companies that should be more impacted (and others that may not). Only 43% of the treatment group has this variable. I don't know how I can consider this variable without practicing dichotomization and sub-samples.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire