For each price sample LS test whether over the 1980s–defined as 1980 through 1989–the prices of skilled-labor-intensive products rose relative to the prices of unskilled-labor-intensive products. Each industry’s skill intensity is measured as the industry’s direct employment ratio of nonproduction to production workers (i.e., NPW/PW). LS identify the pattern of price changes two ways. First, (Figures 8 and 9, pp. 196-197) they pool all industries in each price sample and regress industries’ percentage change in product prices over the 1980s against industries’ skill intensity in 1980. That is, they estimate via OLS the following regression.
Second, to complement these regressions for each price sample they construct a weighted average of all decadal price changes using nonproduction-employment weights and then production-employment weights (Tables 3 and 4). For the domestic prices LS also construct these weighted averages for the 1970s and 1960s as well. Link
The basic finding by LS is they estimate b to be zero or negative, not positive, for their various series of traded prices: industries with higher relative direct employment of nonproduction to production workers did not have larger price increases over the 1980s. These estimates are corroborated by the weighted averages. LS interpret these results as evidence against the hypothesis that international trade contributed to rising U.S. wage inequality by raising the relative price of skilled-labor-intensive products.