For each row in this box select whether the predictor variable is expected to have a positive or a negative direction of effect. The user has to make an a priori assumption for each predictor variable: a positive direction of effect is a predictor variable that will increase the value of the dependent variable, i.e. it is considered to be a source of the dependent variable and the beta coefficient is expected to be positive. A negative direction of effect is a predictor variable that will decrease the value of the dependent variable, i.e. it is considered to be a sink of the dependent variable and the beta coefficient is expected to be negative. These specifications will be used as model selection criteria in the statistical analysis; therefore, the user must consider carefully whether each predictor variable has a positive or a negative direction of effect. Incorrect specifications will lead to incorrect LUR models!
For example, the distance to a polygon that will increase the dependent variable (i.e. a source polygon) is assumed to have a negative direction of effect (i.e. it is expected to have a negative coefficient), because as distance increases the value of the predictor variable increases, while the actual effect of the polygon decreases. Conversely, the inverse distance and inverse distance squared to a polygon that will increase the dependent variable is assumed to have a positive direction of effect, because as distance increases the calculated value (i.e. 1/distance) of the predictor variable becomes smaller, as does the effect of the polygon.
After all predictor variables in the list have been defined as either positive or negative, click the Done button. A green tick mark will appear and the Next > button will be activated. This completes the Distance to and/or value of nearest Polygon step. The newly created predictor variables will be listed in the Predictors Added box on the next page.