Morning Pascale! 

Go to Analyse-  general linear models - univariate. Add your dependent and independent variables. Those covariates which are categorical go to "factors", the scales go to covariates. 

Then go to the model tab. By default the full factorial model is set. It might be useful if you have a short number of control variables and if you are interested in moderation effects between them. I assume it is not the case, so define the model with only main effects.

In options choose homogeneity tests, parameter estimates, observed power and effects. (May be you need more statistics, but these indicators are a minimum got a good ANOVA report.

Run!

I hope it helped.

Sincerely,

Eugene 

Everything depends on the hypothesis you are considering. But,

What is your dependent variable? If it is categorical then you should go with logistic regression including other variables also into the model.

If it is continuous (since you talk about t-test) then you should go with linear regression. Also include other variables as covariates.

Morning Pascale! 

Go to Analyse-  general linear models - univariate. Add your dependent and independent variables. Those covariates which are categorical go to "factors", the scales go to covariates. 

Then go to the model tab. By default the full factorial model is set. It might be useful if you have a short number of control variables and if you are interested in moderation effects between them. I assume it is not the case, so define the model with only main effects.

In options choose homogeneity tests, parameter estimates, observed power and effects. (May be you need more statistics, but these indicators are a minimum got a good ANOVA report.

Run!

I hope it helped.

Sincerely,

Eugene 

Hello Pascale,

From the way you describe your question, I gather that your dependent variable is continuous in nature.  This would suggest that logistic regression isn't appropriate because logistic regression is suitable only when the dependent variable is categorical (usually dichotomous, but it can be more than two categories).  So I think you should aim for what is generally called multiple regression - even though logistic regression can be multiple in a sense, too, because "multiple" really refers to there being more than one dependent variable brought forward for analysis.

It is also important, as Mehmet wrote, that you consider your hypothesis (or research question) - and do so carefully.  If you are primarily interested in social versus commercial enterprise as your independent variable but want to CONTROL for the effect of the other variables that you mention, I think you should consider conducting a hierarchical (sequential) multiple regression rather than a standard (simultaneous) multiple regression.

If, however, you are wanting to consider social versus commercial enterprise along with the other variables you mention, perhaps to see which of them are the strongest predictors of EO, I think it might be best to run a standard "simultaneous" multiple regression.

Your choice of the kind of regression to run is based on your research questions(s).

Of course, do look carefully at your standardised and unstandardised coefficients to see what they tell you about your results.

There are some good YouTube sites outlining how to conduct both simultaneous and hierarchical multiple regression, most using SPSS.

Also, ensure that you have met the assumptions concerning your data if you want your analyses to be valid.  Multiple regression has about 8 assumptions that should be satisfied.  One of these is that there is at least SOME relationship between the variables that you are relating to your dependent variable before you start. If, from correlations or whatever else, you can winnow out some of your prospective dependent variables before throwing them into the regression pot, that might be useful.

Every good wish with your work!

Robert.