Data analysis
We usedmixed factor ANOVA to analyse the effects ofMeal (2 levels:
meal and no meal) and Before/After (2 levels: before and after meal/
wait) on hunger and food liking, on the different measures of food
reward (desire to eat, etc.) and on responding on the work for money
task.We used standardmultiple linear regression (Tabachnick & Fidell,
2007) to test for the independent contributions of liking and hunger
(predictor variables) to food reward (independent variable). Data for
bothmeal and nomeal participantswere included in this analysis,which
ensured a large range of scores for each of the various measures. All
data were normally or near normally distributed. The bivariate corre-
lation between liking and hunger was r = 0.36, ruling out collinearity
as a problem in the regression analyses.