This dataset contains the United States unemployment rate, Federal Reserve Board index of industrial production, and year of the decade for 1950-1959. Unemployment is a natural dependent variable.
A regression of Unemployment versus FRB index shows that unemployment rises as industrial production rises, but the R-squared for this analysis is small.
Adding the year variable to the regressionboosts the R-squared to 86.6%, but the coefficient of FRB index is now negative. The problems of this analysis stem from the high correlation between FRB index and year, resulting in collinearity.