Sarita,

you asked three questions:

(1) How to establish item normality in standardizing questionnaires?

If a variable is not normally distributed and you want to change this, you could normalize the data. For instance, have a look at:

http://journals.sagepub.com/doi/abs/10.1177/0013164495055004009

(2) How do we check normality of each item? 

https://www.youtube.com/watch?v=IiedOyglLn0

(3) Is establishing item normality necessary in construction and standardization of tools?

Procedures like Conbach's alpha rely on correlations for which normally distributed data are required.

However, the pearson correlation seems to "be able to deal" with some non-normality, see:

http://www.sciencedirect.com/science/article/pii/S0047259X99918584

and citing research.

.

Cronbach's Alpha and non normal data do not mix well !

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3279724/

so, better check that your data do not strongly  violate the normality assumption

.

Thank you Matthius and Fabrice

 I know that normality of data is to be checked. I also know how important it is. I have clarity about the procedure/steps involved in establishing the normality.

 @ Matthius the link for the video you have shared gives the description of normality of variable.

I want to know if 'normality of item' in any manner is different from 'normality of variable'. As per Cronk's book normality of each item is to be established. As per my comprehension the variable and item normality are two different concepts. 

.

the normality assumption is for test scores, not test items

(items could be ordinal, Likert items for instance, and normality would not make sense in such a context !)

.

@ Fabrice thanks. I also hold the same opinion. Got doubtful when I came across the term  'item normality' in a paper and  book.

normality is a myth. check our paper

Article TO DETERMINE SKEWNESS, MEAN AND DEVIATION WITH A NEW APPROAC...

@Thanks Guven.