3. Specialized Bicycles wants to increase the efficiency of their manufacturing workers. They believe that a new training program would increase the number of bikes that could be made per day. The company currently has 224 employees working in manufacturing and they invite all of their employees to take the training. Around half of them do. The company then runs some regressions to try and see if the training is working. The following table shows the results of three different regressions, each with the dependent variable (or y-variable) being the number of bikes made per day: Effect of Training on Production Efficiency (2) 2.29** (0.972) took training tenure HS diploma N 2.31*** (0.113) 224 0.740* (0.221) 224 (3) 1.99** (0.883) 0.521 (0.334) 1.14** (0.213) 224 a. How do you interpret the coefficient (the main number) in column (1) for "took training." (Note, took training is a binary variable that =0 if the worker didn't take training and=1 if they did). b. The number changes in column (2) and column (3), but not by a whole lot. What do these small changes say about the training program? I. Regardless of how long someone has been with the company or whether they have a HS diploma, training still improves their efficiency. ii. Tenure doesn't seem to have any impact on production efficiency ii. Tenure doesn't seem to affect the effectiveness of the training very much iv. Having a HS diploma doesn't seem to have any impact on production efficiency v. Having a HS diploma doesn't seem to affect the effectiveness of the training very much

3. Specialized Bicycles wants to increase the efficiency of their manufacturing workers. They belleve that a new training program would increase the number of bikes that could be made per day. The company currently has 224 employees working in manufacturing and they invite all of their employees to take the training. Around half of them do. The company then runs some regressions to try and see if the training is working. The following table shows the results of three different regressions, each with the dependent variable (or y-variable) being the number of bikes made per day: a. How do you interpret the coefficient (the main number) in column (1) for "took training:" (Note, took training is a binary variable that =0 if the worker didn't take training and =1 if they did). b. The number changes in column (2) and column (3), but not by a whole lot. What do these small changes say about the training program? 1. Regardless of how long someone has been with the company or whether they have a HS diploma, training still improves their efficiency. ii. Tenure doesn't seem to have any impact on production efficiency iii. Tenure doesn't seem to affect the effectiveness of the training very much Iv. Having a HS diploma doesn't seem to have any impact on production efficlency v. Having a HS diploma doesn't seem to affect the effectiveness of the training very much c. There may be a concern of omitted variable bias and that the results shown above are just spurious correlation. Which of the following would be the main concern for Specialized Bicycles? i. That people who took the training program didn't pay very good attention ii. That people who took the training program might already be better workers than those who didn't take the training program iii. That the timing of the training program might be correlated with something that affects worker productivity, like weather or holidays iv. That people who took the training program are more educated than those who didn't take the training program d. What would be the best way to make sure the results are not coming from spurious correlation? Another way to say this is, how can Specialized Bicycles make sure the effects are actually causal? 1. Force all workers to take the training program ii. Randomly choose a group of workers to take the training program iii. Randomly change the type of training each person who takes the program gets iv. Only allow educated workers to take the training program v. Only do the training on sunny/warm days