A paper I made in the summer vacation was made by using Logistic model and SPSS 17.0, all of which were self-taught (I was a senior, and it was very hard to teach econometrics and SPSS software at the undergraduate stage in our school), so I won't talk nonsense and cut to the chase directly.

According to my understanding in the past two months, the Logistic model can be divided into two types, one is binary and the other is multivariate. The difference between them is the value of the dependent variable. The requirements of independent variables are the same, and it doesn't matter how much, continuous and discrete, as long as there is no correlation between independent variables. The binary Logistic model requires that the dependent variable obey the two-point distribution of 0- 1. The dependent variable values of multivariate Logistic model often need to be converted into variables in the range of [0, 1] by quartile method and percentile method, which are also discrete, but more than two values can be obtained.

I'm not sure about your sentence "The total scores of their three scores are not equal". Do you mean that the ranges of these three variables are different or that you give them different weights? It doesn't matter if the range of values is different. If the weights are different, we can't use Logistic model, only AHP.

Also, if the econometric software is used for Logistic regression, the software will provide Hosmer &；; Lemeshow test (Likelihood Chi-square test), if the significance level is higher than 5%, the model has a good fitting degree, which shows that Logistic model can well describe the relationship between independent variables and dependent variables provided by you.

I hope this helps.