I'm analyzing the effect of different categories of expenditures on subjective well-being (SWB). The way I'm doing that is by taking the natural logarithm of the expenditures (expressed in US dollars), to see if higher levels of them are associated with more SWB. One problem I'm facing is that as I just have a cross-sectional dataset, so I can't use the differences-in-differences method (or something similar) to control for unobservable characteristics of the consumers. Therefore, I will probably be capturing not just the association between SWB and higher levels of expenditure in a specific set of goods, but also between SWB and being the kind of person who buys that specific set of goods. Can I address this problem by adding one dummy variables for each category of expenditure that equals 1 if the individual has spent any money on those specific categories and equal 0 if he/she has not spent any money on them?


Thank you in advance!