1)- is it adequate to run a multinomial logistic regression test using spss on two categorical variables to test the influence relationship, knowing that data is collected via likert scale items and then computed into 5 classes : example from 5 to 10: weak, from 15 to 20: moderate, from 20 to 25: strong etc ..

2)- is it possible to test the correlation and/or regression between two variables measured with different likert scales (one is measured with a 5 likert scale item, the other is measured with a 7 likert scale item) ( both will be computed into 5 categories)

thank you for your kind assistance