Linear regression (AskiaAnalyse)

This analysis type attempts to create a model which estimates a target question with explanatory variables. The target question is a numeric variable (numeric or scaled response), and the explanatory variables are numeric. The results consist of a single table containing:

• the coefficient (a_i) of the model (Y = a_i + cste);
• the constant of the model (Y = a_i + cste);
• the probability given by the Student test (when this value is close to 100%, the item explains the variable fully);
• the (R^2) (the coefficient of determination indicates how well the data fits the statistical model. The coefficient lies between 0 and 1);
• the ({ over R^2}) (adjusted (R^2)) : (measures the variability of (Y) when we increase the number of explanatory variables. ({ over R^2}) lies between 0 and 1));
• the Sigma for Fischer test.

To perform a linear regression analysis:

1. In the general tab, select linear regression.
2. Add a numeric or scaled response question to the to explain tab (for example, drag and drop them from the questionnaire list).
3. Add several numeric responses questions to the descriptive tab.
4. Click results to view the analysis.

You can choose how to handle inactive interviews in the regression analysis (for example, you may want to display these interviews, but not include their values). To do so, click analysis options.... For details, see multivariate analysis options.

It is possible to save the linear regression model by saving it as a numeric variable. To do so, click create variable.... For details, see creating a variable.