Sritha Zith Dey Babu P. MADHUSUDANA Patra Wegayehu Enbeyle Lalgoulen Khongsai M. Rashid Nizamani Jonathan James Olmedilla Canete any one who know please give feedback
Cell size is not the important factor. a good rule is probably 10 or more observations per predictor variable. You can check for more options by a google query. Best, D. Booth
I would agree with David, there is no minimum cell size. If you are doing a frequentist hypothesis test then you want to run a power analysis before the data is collected. Some people use a cell size of 10 as a rule of thumb, but is depends on how big an effect you expected. If you are doing a Bayesian regression you could argue it does not matter at all, it just changes the uncertainty in the estimates.