higher if three instead of two buffets are set up while all other variables are kept at the same level. The coefficient for “Austria” indicates that if the percentage of Austrians staying at the hotel is 1 percentage point higher, the average food waste observed is 0.1 g lower. The estimate for “Younger guests” indicates that on days when the guest mix includes more young guests, the average food waste is 3.4 g higher compared with days when the share of older guests is higher. The standardized coefficients in Table 2 indicate that the importance of the coefficients for variables measured in percentage is in fact higher than their unstandardized coefficients would suggest. Furthermore, the dependence between the number of buffets set up and the number of guests staying at the hotel as well as the age distributionof guests changing with the occupancy level in the hotel leads to higher VIF values for these variables. Figure 1 illustrates the relative association of each of the predictor variables on the dependent variable (food waste). For each predictor variable selected by the stepwise procedure, a bar is added where the length is proportional to its size and they are ordered by absolute value. The direction of the bar depends on its sign. For each bar the standard errors of the estimated coefficients are indicated by the lines. The gray bars indicate that the t-tests have a p value smaller than 0.05, indicating a significant effect, while white bars are used otherwise.