The manager of a retailer has been experiencing quite a few online "order errors" (i.e. retrieving

the wrong item). A training program then introduced as the manger believes the source of these

errors are due to being as an inexperienced workers who have never attended an on-job

training session. The 37 collected randomly selected sample data below are provided

with number of errors, whether the worker attended a training sessions (train=1, otherwise=0),

years of experiences.

obs. Errors Experience Train

1 13 9 0

2 3 27 0

3 12 6 0

4 0 20 1

5 5 9 1

6 49 26 0

7 39 4 0

8 8 5 1

9 6 28 0

10 12 2 1

11 5 17 0

12 1 5 1

13 21 8 0

14 73 2 0

15 28 5 0

16 0 29 1

17 6 25 1

18 3 23 1

19 4 20 0

20 4 17 1

21 2 30 1

22 1 27 1

23 3 23 0

24 1 24 1

25 6 5 1

26 3 25 1

27 6 14 0

28 4 21 1

29 7 2 1

30 4 24 1

31 21 8 0

32 73 2 0

33 28 5 0

34 0 39 1

35 9 17 0

36 1 23 1

37 6 20 0

A) Estimate two linear regression models - one with experience and train as independent

variables, and another one with an extra feature for the interaction between experience

and training. Both models come with an intercept. You can use basic excel (Data

Analysis adds-in) or Minitab 19 for this part.

B) Comments on coefficient of determination, and the significance of the estimated coefficients of both models.

C) Clearly explain- in a nontechnical language- the followings:

D) Significance or insignificance of the interaction term estimated in part (a).

E) Interpretation of the estimated coefficients of both models.

F) Report which model you will go with for predicting the errors, and why?

G) Use the selected model in part (d), and predict the number of errors for an employee with

20 years of experience who attended the training session as well as for an employee with

20 years of experiences who did not attend the session. Any comments? [No need to use

Minitab 19 for this part].

H) Discuss any potential influential observations for both of the estimated models.

I) List some other potential features may have impact on the errors made by workers of

this retailer, and explain how you would collect data for these potential extra features.