In the original publication, there are three typos which were unfortunately introduced by the publishers during the production process. The following are the corrected sentences: 1. In the Online First version of our article, section 1, 8th para says: “First, our analysis builds on actual consumption patterns of a very high number of households that are representative for Germany and, moreover, surveyed at several points in time (whereas Haisley et al. use lottery ticket purchases of > 200 subjects after an experimental task as data).” The correct statement reads: “First, our analysis builds on actual consumption patterns of a very high number of households that are representative for Germany and, moreover, surveyed at several points in time (whereas Haisley et al. use lottery ticket purchases of less than 200 subjects after an experimental task as data).” 2. In the Online First version of our article, section 4, 1st para says: “First, we perform logit estimations with a binary dependent variable that indicates whether a household participates in gambling (i.e., has gambling expenditures < 0).” (page 10) The correct statement reads: “First, we perform logit estimations with a binary dependent variable that indicates whether a household participates in gambling (i.e., has gambling expenditures greater than 0).” 3. In the Online First version of our article, section 4.1.2, 5th para says: “Our dependent variable In(Gambling expendituresi) captures the natural logarithm of household i's spending on gambling.” The correct statement reads: “Our dependent variable ln(Gambling expendituresi) captures the natural logarithm of household i's spending on gambling.”.