My row data has 2000 observations.
Sleeping hours fluctuate from 1h (min) and 15h(max) so I want to categorize them into categories, but I don’t know what is important to consider, taking into account min. and max. values are really extreme. I was thinking in two possibilities first (8h) and second (9h), but do I need to think about frequency? Does categories have to be homogeneous? I am using SAS.
N=2000
Sleeping hours mean=7h
Skewness=0.0638
Thanks
Assume you enjoy the statistical approach: why not putting the data into quartiles, pentiles or more detailed into10 classes.
On the other hand your physiological/pathophysiological suggestion of 8h youngsters, depressives and hypersomniacs. I would prefer your second proposal with 5 classes - somehow a compromise between statistics and biology. SAS will do everything.
Dear Friedhart Raschke I really appreciate your answer. Actually, my population are people aged between 30 and 65 years old. But, as you say the deliberated way how I was thinking to classify the sleeping variable could have implied other problems. So, I decided to distribute it into quartiles. However, the new classification bring me to new questions. Is it appropriate include decimals in categories classification?
For example: for sleeping categories: 8h.
Thanks for your reply
Hi Diana! It's up to you, which kind of category you prefer. The use of decimals is OK.
- Badges
- Science topic
- Similar topics
- Psychology