Logit analysis defines the functional relationship between stated purchase intentions and preferences, and the actual probability of purchase. A preference regression is performed on the survey data. This is then modified with actual historical observations of purchase behavior. The resultant functional relationship defines purchase probability.
This is the most useful of the purchase intention/rating translations because explicit measures of confidence level and statistical significance can be calculated. Other purchase intention/rating translations include the preference-rank translation and the intent scale translation. The main disadvantage is that the software is not easy to find.