Using either mean, median or mode to describe central tendency of any data



Hello People,
I’m doing the Descriptive statistics course in Udacity. I understand that mean, median and mode, all three are used to describe any data. But if I have to use only one of the three to precisely describe the central tendency of a certain data:
1)On which kind of data should I use the mean?
2)On which kind of data should I use the median?
3)On which kind of data should I use the mode?


Hi there,

According to me you should

1- Use mean when you data is continuous and Uniformally distributed i.e it shows a nice bell shape curve with no outliers.
2- Use median when your data is heavily skewed due to outliers or is ordinal.
3- Use mode when you are dealing with nominal data.

Though the usage will depend on what you trying to do with the data.

Hope this helps.


Thank you NSS. That pretty much solves my question. Still other creative ideas are welcome :slight_smile:


I agree with the solution of first two problems, using mean when data is continuous and uniformally distributed and also using median when data is heavily skewed but you should use mode for categorical variable independent of weather the data is ordinal or nominal.

According to me using median for an ordinal variable will just gives middle value of the distribution that will be totally random and do not have any information about the distribution, thus using mode will give the level which is more prominent.
Please let me know if there is some mistake with this analogy.




Hi there, I would like to draw your attention towards the following points.

1- Median is calculated only when we have sorted the data into ascending or descending order.
2- Ordinal variables have an order i,e sense of ascent or descent and hence can be sorted. Remember this cannot be done with nominal variable.

So, Since ordinal variable can be effectively sorted and this is what median requires. I don’t see a problem in using median as one of the central tendency.

Please let me know if I am wrong.