Product preference Test and Alination

[Kunal]

What is Product preference Test and Alination in SPSS.

[Art Kendall]

Please explain in more detail what you are asking.

In what discipline does the terminology arise?

Is this terminology from a particular textbook?

What does your data look like?

What questions are you using the data to find answers?

[David Marso]

Kunal wrote

What is Product preference Test and Alination in SPSS.

Please provide a freely available reference of what you inquire?
OR a proper formula with well defined entities.
FWIW: There is no such 'canned' menu item of that name.
Sounds like some sort of market 'research' buzzword.
I believe you misspelled Alienation?
Here is one bit of gobblygook I found:
http://www.insightsnow.com/products-0/bid/267512/Research-Tip-Alienation-Analysis
However there is no specific statistical test mentioned.

" For each product, ask their purchase interest and ask preference for the two products directly. The percentage of people who would be alienated can be calculated by identifying the number of people who:
1.Prefer the current product AND
2.Would buy the current product AND
3.Would NOT buy the new product.

Take this count and divide by the total number of people in the study and you have an estimate of the current users who could likely stop buying your product due to the changes. This percentage is often compared to the percentage of people who prefer the new product AND would buy the new product AND would NOT buy the current product.   Assuming you recruited current product users, this group provides a benchmark of the variability in the data and allows for a statistical test to determine if the alienated group is large enough that you may damage your brand by introducing the new product. There are other modifications to this analysis that are appropriate depending on your target market structure, whether you include a no-preference option, and options for considering attributes other than purchase intent. (SNIP sales spiel)'

Sounds like it could be as simple as two computes and a t-test?
I suppose that totally depends upon the nature of the data?
--