Most of the research would look at the relationship between the factors or subscales, not individual items. the only thing close to this might be the use of the Cronbach's alpha to eliminate items from scales or subscales based on their lack of correlation. But people usually only list the item(s) that don't correlate and had to be removed. They usually don't discuss the items themselves.

 Are you asking for how to score the EPQ? Do you have the short version or the regular version? do you have the manual?

If you are asking for some raw data so that you can then the analysis of the relationships between items, then you are welcome to use the 3570 cases of answers to the 57 items of the EPi that is available in the psych package in R.

Install R, http://cran.r-project.org)

start R  

install.packages("psych")  #just do this once

library(psych)  #makes the package active every time you want to use it

?epi   #will give you a description of the 57 items as well as examples.

There are a number of different data sets embedded in the psych package, but this is probably the most useful for you.  

See http://personality-project.org/r   for a discussion of how to use R in personality research.  

Is there any specific reason to look for such scale answer?

Yes. FCA works with this kind of values and we want to explore this area

Is this can be analyzed for five- likert scale? eg., strongly agree to strongly disagree

In FCA is possible, but we need the bivalued attributes for a special algorithm.

Understand, thanks for this information, perhaps I should apply this method for further studies.