Dear Buddhika,

In an asymmetric data, the distribution is highly skewed. When this occurs, mean and standard deviation may not be good measures of central tendency and dispersion respectively. By this, inferential  analyses become problematic. You can use alpha 3 measure of skeweness as:

α3 = Σ (x – x)3 / n X sd3 (pls note that the second x of the formula in the bracket is x-bar: mean value and figure 3 is superscript)

Preferably, you can normalize your data by computing a new variable based on the square roots of the original (Skewed)  data. You can do this by  Box Cox technique and  Statistical Analysis Software or  by data transform in SPSS.

If you wish to continue with the non-normal distribution, you can use Minitab software package.

Thank you

Hello Adebukola,

Thank you very much for your valuable input

I wonder how they calculate these values using in spss or strata software

Please guide 

http://www.biomedcentral.com/1471-244X/3/3

Kind regards

Buddhika