Rather than comparing multi-establishment and single-establishment firms by estimating Regression 2 separately for the two subsamples of firms, suppose instead that you estimate a version of...

Rather than comparing multi-establishment and single-establishment firms by estimating Regression 2 separately for the two subsamples of firms, suppose instead that you estimate a version of Regression 2 (that includes an interaction term) only once, using the full sample. Write down the regression equation you would estimate, and explain how you would use the parameter estimates to measure the relationship between wages and health insurance.