new sheet, since the filter tool seems to just hide the other observations). That's probably the easiest way to set up the data before creating the correlations and scatterplots. I used Excel. A data problem. Use the famed Penn World Tables. Report the (Pearson) correlation between 2015 GDP per capita (rgdpe/pop) and: (a) Capital per capita (cn/pop) (b) National investment rate (csh i) (c) TFP (ctfp) (d) Labor share of income (labsh) Also, report scatterplots of each of these relationships, with GDP per capita on the Y-axis. You'll see that the correlation between capital per capita and GDP per capita is really high-indeed, it's pretty close to a linear relationship, not a diminishing returns one! This is what the Solow model predicts in steady state, if TFP differences (or to be more precise, labor-augmenting productivity dif- ferences) are the big difference across countries: Double the labor-augmenting productivity creates creates double the capital in the long run. Suggestion: When you download Penn, you'll see you can just filter the 2015 data and drop the rest (you might want to cut-and-paste the 2015 data into a