1.Before evaluating the effect of an experiment, it is important to make sure that the experiment was executed correctly. Check whether the test and control groups are probabilistically equivalent on their observables? a. More specific, compare the averages of the income, gender and gamer variables in the test and control groups. You should also report the 93 difference in the averages. Compute its statistical significance. [2 pts] b. Briefly comment on what these metrics tell you about probabilistic equivalence for this experiment. [2 pts] 1:. If you had run this type of analysis BEFORE executing an experiment and found a large difference between test and control groups, what you should do? [5 pts] d. {Open/ Ended Question] If you had millions of consumers, your "classic" statistical signicance tests would not work [this is because the number of samples is used to compute those classic statistical tests}. Do some researches online and propose what significance test would you do in case you had "big data"? [5 pts] 2.Evaluate the average purchase rates in the test and control for the following groups. For each comparison, report the average purchase rate for the test, average purchase rate for the control and the absolute difference (not the % difference] between the test and control. 3. Comparison 1: All customers [2 pts] b. Comparison 2: Male vs Female customers [2 pts] 1:. Comparison 3: Gamers vs NonGamers Customers [2 pts] d. Comparison 4: Female Gamers vs Male Gamers [2 pts] 3.Assess the expected revenue in the test vs. control for the following comparisons: 3. Comparison 1: All customers [4 pts] b. Comparison 4: Female Gamers vs Male Gamers [4 pts] 4.Based on your previous answers, provide a brief recommendation to your management team su mmarizing the expected financial outcome for GameFun. a.Should GameFun run this promotion again in the future? If no, explain why. If yes, should GameFun offer it to all customers or a targeted segment. [20 pts] Appendix KFactor Explained Assume you have a market (e.g., a US state or a country} where you have almost zero daily organic downloads. You agree to buy users though targeted advertisement. Let's say that you buy 10,000 users. If after a specific period of time, you suddenly have 12,000 users, you know that anon nf them came as an indirect FPCI Ilt nf vnllr nairl user arnllicitinn ramnaipn n 'vnllr