SPSS 8: Logistic regression

SPSS 8: Logistic regression

1. Analyze> Regression> Binary logistic

   • logistic regression. independent - covariates, dependent - dependent. OPTIONS- classifications plots, iteration history, include constant in model

2. output: case processing summary: information about how many data points have been included in the analysis

   • dependent variable encoding: how the DV has been coded 1-rude, 2-polite
     • Block 0 - beginning block. 'predict the most frequent outcome'
       • classification table -more classed as 'rude' means predicted outcome set to 'rude'. Gives % correct e.g. 55% of sentences correct to predict 'rude'
         • variables in the equation + variables not in the equation = in Block 0 no variables have been entered into the model
       • Block 1: Enter method - should be better than block 1
         • iteration history = process of approximation in the construction of the model.
           • -2 log likelihood = success of model. needs to be low, gets lower each iteration
             • omnibus tests of model coefficients - tests general fit model. chi squared, df, significance
               • model summary = approximates R2
                 • classification table = summaries prediction, compare to Block 0.
                   • variables in the equation = B - coefficient for each predictor. Wald & Sig = usefulness as a predictor. EXP(B) - change in predicted odds for each unit change in predictor variable
                     • chart maps the observed against predicted. if sig = more observed will match the cases

Imagens e arquivos do mapa