" ...in the U.S. General Social Survey one can regress happiness on income and income squared. The latter term is always significant, which is why the single term log Y is used in annex 4.1."

I suppose the author uses "regression analysis" and "regression equation" to analyze the relation between happiness and income. But I have no background in this matter.