4. Using an appropriate display compare the success rate between salespersons (hint. A salesperson will have at most as many successes as they have transactions. So if a different employee, say Jennifer, was on the job for 40 transactions, and 32 of those were successes, then her success rate would be calculated as 32/40; dividing instead by the total number of transactions available to all salespersons would be unhelpful.)

(a) (2 points) First plot a graph that ignores (is unconditional upon) shift

(b) (4 points) Now repeat your analysis (from part (a) ) but conditioning on each of the levels for shift, i.e, repeating the analysis for only daytime and then for only evening shifts. Please display these plots in terms of relative frequencies (proportions)

(c) (3 points) Please contrast the conclusions one might draw when conditioning your analysis upon shift versus the original marginal analysis (i.e., when you ignored shift).

5. You are examining a single transaction record corresponding to NoSale for Alphonse. Based on the dataset provided: (a) (2 points) what is the probability that this sale took place in the evening?

(b) (1 point) what is the probability that it took place in the daytime?

(c) (2 points) Would you answer change of the salesperson linked to the record was Constance? If so explain why?

6. (3 points (bonus)) PJamJ Management informs you that the data provided were collected as part of a study investigating the potential gains in success-rate induced by exposing a salesperson to a check-list — the study involved exposing each salesperson to the check-list before his/her night shift. PJamJ Management would like you to estimate the beneficial effect of this checklist intervention — namely by quantifying what the success rate would have been in the absence of the checklist exposure, and noting what improvement the exposure induced. Can this be estimated? Why or why not?

 

sale No sale total alphonse 115 687 802 belinda 124 575 699 constance 110 389 499 total 349 1651 2000