Part I: Learning by Doing We replicate parts of an academic journal article "Toward an Understanding of Learning by Doing: Evidence from an Automobile Assembly Plant, " abbreviated Levitt et al. [2013). Datasets: For Levitt et al. (2013): learn do weeklyxlsx, where "lea rn do\" abbreviates "Learning by Doing" from the title and "weekly" refers to the fact that these are the weekly [not daily) data. 1. Replicate Figure 1 below. Create the yaxis variable and name it ave_def_pc. Like the authors, display only weeks with the production of at least 100 cars in the gure. [For the xaxis tic values in Figure 1, just use the week number.) 90 Average Defects p-el' Car 10 20 3U 4U 5U 60 70 EU D 'LIIllIIl_ h. h. N \"b *1. '1; "Ir s' e" i" it'll will a\"? 4' at\" if ll Production Weakear Figure 1 plots the average number of defects per car by week. When production begins in midAugust, average defect rates were around 75 per car. Eight weeks later, they had fallen by twothirds, to roughly 25 defects per car. These strong initial learning effects are consistent with findings in the broader literature on learning by doing." [pp. 6534) 2. Replicate the simple regression in Table 1, Panel A, Column (1). First, create a variable named cum_prod for the cumulative production in previous weeks. Next, create the x variable named In_cum_prod. Create the y variable, In_ave_def_pc. Run the regression: use only weeks when at least 100 cars are produced (your number of observations should match the table). (Note: Cumulative production includes all weeks, including those with production below 100 cars.) To visualize, see Figure 2 below. Table 1: Estimates of Learning By Doing (1) (2) Panel A. Weekly Data 0.289* 0.335* Estimated learning rate, (0.007) (0.017) Time trend 0.007* (0.002) Observations 47 47 R2 0.961 0.969 Panel B. Daily Data -0.306* -0.369* Estimated learning rate, B (0.006) (0.014) 0.001* Time trend (0.0002) Observations 224 224 R 0.931 0.943 Notes: Column (1) in both panels shows estimation results for In (D,) = a + B In(E, ) + Et , where D, is the average defects per car in time period t and E, is production experience up to that point: cumulative number of cars produced before the current period. Column (2) in both panels shows estimation results for In(D, ) = a + B In(E. ) + y * t + Et. Heteroskedasticity-robust standard errors are in parentheses. * Significant at the 5 percent level.Table 1 shows the results of estimating these specifications with our sample. Panel A shows results from specifications using weekly data (average defect rates over the week and production experience at the week's outset); Panel B shows results using daily observations." (p. 655) 4.5 + 3.5 Natural log of average defects per car this week 4 6 8 10 12 Natural log of cumulative production of cars in previous weeks Figure 2: Like Figure 2 on p. 656 of Levitt et al. (2013), except that it shows the weekly data (instead of the daily data). This figure corresponds exactly to Table 1, Panel A, Column (1).3. Replicate the multiple regression in Table 1, Panel A, Column (2). With regular (not robust) standard errors you get (0.016), not (0.017). But, rounded, it's (0.02) either way. 4. Construct a correlation matrix for the variables: In(average defects per car), In(cumulative production), and the time trend. Use the data from part 3. Verify it matches: time_trend In_cum_prod In_ave_def_pc time_trend In_cum_prod 0.870334115 In_ave_def_pc -0.809082222 -0.980156745 5. Table 1, Panel A, Column (1) shows a simple regression, which means an uncomplicated relationship between the correlation and slope coefficient: they will have the same sign. Verify that the slope coefficient in the standardized regression equals the correlation in part 4 (-.9802). 6. Table 1, Panel A, Column (2) shows a multiple regression, which means the correlation and the slope coefficient can differ wildly and can even have opposite signs. 7. Why are many of the values in the multiple regression output for part 6 (standardized) identical to part 3 (not standardized)? Why does standardization change the regression coefficients on the x variables? 8. Fully interpret the number -0.289, which appears in Table 1, Column (1), Panel A. Write an interpretation that would be clear to someone who has not read the Supplement. Answer with 1 precise sentence that clearly explains what that number means in a practical sense. 9. How many regressions are reported in Table 1? Which of these corresponds to Figure 2? Answer with a number of regressions and the column number and panel that goes with Figure 2. 10. By the last time period of the available data, approximately how many cars had the manufacturer produced in total during the study period? Answer with a number and your work. 11. Compute the 99% confidence interval estimate of the learning rate using the results in Table 1, Column (2), Panel A. Answer with a quantitative analysis. 12. Table 1 reports the usual test of statistical significance for each coefficient at a 5% significance level (in other words, the statistical test that standard software packages, like Excel, automatically conduct). All are marked with a "*" as being statistically significant. For the "*" next to 0.007 in Column (2), Panel A, what are: the formal hypotheses being tested, the value of the test statistic, and the relevant critical value(s)? Answer with formal hypotheses in standard notation and the other requested values