Standardising scales

[Indy]

Hi

I have had to combine 3 dimensions of a Big Five factor personality scale (openness, conscientiousness and agreeableness) due to multicollinearity.

I had to reverse score some of the items on the above dimensions previously and I labled them as variables BFIo, BFIc & BFIa.

I then took the items incuding the reversed scored items from the three variables and computed the Zscores via SPSS. I combined the Zscores into the new varible BFIoac.

Having done this I performed cronbach's alpha and had good reliability on all items in that variable or scale.

I ran the hierarchical multiple regression and I was surprised to find that the BFIoac had a mean of 0.00.

Is this normal and if not what did I do or didnt do?

kind regards

[JohnM]

By definition, z scores, representing the standard normal distribution, will have a mean of 0 and std dev of 1 (of course in samples they will differ from those values....)