The result is no less valid with a log scale. It's very typical to determine a linear regression on a log scale. It's a paper aimed at those capable of underst

The result is no less valid with a log scale. It's very typical to determine a linear regression on a log scale. It's a paper aimed at those capable of understanding statistics and analysis. It's assumed that the reader is sufficiently educated to understand the implications of a log scale. If that assumption is out, the rest of the article is going to be meaningless to the reader too. The article does say it has a US sample space too. I didn't read too much into it, for the more detailed conclusions, however from a skim read the high level conclusion "richer people are more likely to happier" is correlated within the given income bracket.