# How can I rank values that are not normally distributed?

I have one metric value changing between 0 and 1 for 650 observations. When I analyze frequency distribution I see that the values are right skewed. I need to classify values as high/medium/low. I used 25th and 75th percentiles of the values. Next, I regarded values as "low" below 25th percentile and "high" above 75th percentile.I used average (between 25th and 75th percentiles) and classified it as "medium". Is it logical or should I do an another test?

• Thanks but I do not think that my question is out of statistics. May 6, 2020 at 2:58
• researchgate.net/post/… May 6, 2020 at 2:58
• All of the answers in that post are different because there is no one correct way to do it. None of those answers are more valid or more useful than another without knowing why the values need to be binned, what the binds mean, and what the variable means. I can give you a complex algorithm to create bins but there is no way, based on the information you provided, to determine whether that method is any better than any other method. YOU have to decide what method works for your purposes.
– Noah
May 6, 2020 at 3:03
• Thanks for clarifying, Noah. This is really helpful. May 6, 2020 at 3:06

Yes, it is logical to classify values as high, medium and low using the quartiles, like you did. In fact, many people do it this way, especially for non symmetric distributions. This is not the only logical way of approaching the classification, of course.

• Less logical than intuitive, no? May 22, 2020 at 1:16