Dear RG community,
I would like to use a logistic regression model in SPSS. But there is a problem. My actual dependent variable is continuous. My independent variable is dichotomous. Is it possible to put the actual dependent variable in the independent variable place, and the actual independent variable in the dependent variable place?
I cannot use linear regression, because the assumptions are violated.
Thank you.
The latest version of SPSS (26) has quantile regression, which is an appropriate analysis.
You use independent and dependent variables as if there are requisite analytic distinctions for each. Indeed, I would, and have, argued that these terms really are only appropriate for an experiment in which one manipulates an independent variable. Although I am not sure from your description, my guess is that this is not the case in your study. It is, in this case, an association between the two variables you seek.
I am not sure what you mean by "assumptions are violated" for linear regression. As I see only two variables, a simple PPM correlation, or equivalently a single predictor linear regression would be fine with usual assumptions. If you do calculate a PPM correlation between the dichotomous "IV" and continuous "DV," you will find that this is not only appropriate, but is equivalent to what would be seen as a more typically used analysis -- a t-test between means of the continuous variable for each of the two levels of the dichotomous variable. The question of relationship and difference are the same; the p values will be the same and the test statistics are related as t=SQRT((N-2)*r2/(1-r2)).
I am not sure why you say the assumptions of linear regression are violated. If you fit a General Linear Model (GLM) then you enter the continuous dependent variable as the dependent variable, and the categorical independent variables as factors (continuous dependent variables are called covariates). In SPSS you have to specifically ask for the regression coefficients to be displayed. (By the way, a model with a continuous dependent variable and one categorical independent variable is exactly equivalent to a one-way ANOVA.)
Francois E Steffens John Morris Edward G. Brown Adrian Esterman
Thank you all very much for your answers, they were very useful! I managed to solve the problem, and now everything is fine.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS