Could you tell us how YES and NO are coded? There is a table from the output you could show. Generally, a negative constant is the predicted value when all the independent variables are equal to 0, which doesn't make sense unless 0 has a meaning (like the variables are "centered" at 0). You could center the variables by removing the mean from each (think "z-scores, for example) one, then re-run if you want a meaningful constant.
------------------------------
Rick Marcantonio
Quality Assurance
IBM
------------------------------
Original Message:
Sent: Fri May 07, 2021 04:14 PM
From: Vittoria Bsn
Subject: Problem with logistic Regression on SPSS
Hello!
I'm writing because I seem to have a problem understanding the output of a logistic regression I ran on SPSS. The logistic was based on:
- Dependent variable: Would you try this product (YES/NO)
- 4 Factors drawn from a factor analysis ran on the following variables:
- 8 questions like: What would push you to try this product on a scale from 1 to 10
- 6 questions like: What would prevent you from trying this product on a scale from 1 to 10
I got 5 different factors. I ran the logistic regression using them as independent variables, and I got the following result using SPSS
You can see only 4 factors here as the 5th was not significant. Anyways, I'm really struggling to understand first of all how to interpret the negative constant (-0,858). Moreover, it seems to me that, counterintuitively, the factors that should push a person to try the product under investigation (specified benefits with certified information, and positive feedbacks) have a negative effect on the dependent variables, while the ones that intuintevely should prevent a person from trying it (health-related fears and fear of disgusting flavour), have an overall positive effect. I've tried the analysis many times but I can't really understand my mistake. Can you help me?
------------------------------
Vittoria Busani
------------------------------
#SPSSStatistics