non normal data

[chuie]

Hi There,

I have  Length of Stay (LOS) data and it is  not normal/skewed to right. So my questing is : When we have not normal distribution, how do we interpretate the average of not normal distribution? My understanding for normal distribution is that 68% of the data falls within 1 SD away from the average. But for non normal we cannot say that. when data is not normal it is best to sue median instead of mean.

when I transform the LOS data , how to interpretae the average of log transform ? How do we deal with calculating average of non normal distribution

please help

Thank you

C

[Reeza]

Or you report the Median instead of the Mean.

[Reeza]

You can calculate the mean of the transformed variable and the confidence intervals and then back transform into the normal units, but that's hard to explain.

---

@chuie wrote:

Hi There,

I have  Length of Stay (LOS) data and it is  not normal/skewed to right. So my questing is : When we have not normal distribution, how do we interpretate the average of not normal distribution? My understanding for normal distribution is that 68% of the data falls within 1 SD away from the average. But for non normal we cannot say that. when data is not normal it is best to sue median instead of mean.

when I transform the LOS data , how to interpretae the average of log transform ? How do we deal with calculating average of non normal distribution

please help

Thank you

C

---

[chuie]

do you have link  where I can read about back transformation?

If we back transform the unit then can we interpretae the average as if that is normal distributed? 

[Reeza]

No, that's why it's hard to explain 🙂