4. Would your answer to C13 differ if the benchmark for our pvalue was 1% instead of the default 5%? NO. Using EmailSubscribe, as the measure of 'best', figure out which variant is the best variant to use. 6. Would your answer differ if the benchmark for our pvalue was 1% instead of the default 5%? 2'. Calculate the average amount a user donates for each variant. If our goal is the receive the most amount of money from users, which variant would we want to use? Why? B. Is there a specific age group within the Variant you selected in Question 2', which should be targeted if we are wanting to receive the most amount of money? Note, we are not interested in the number of people within the segment who donate, but rather how much on average is donated. *Hint, you might need to use Excel modeling here, as opposed tojust using AB Testing. One of the campaign coordinators suggests that it doesn't matter how much someone donates, as long as they donate. The thinking is that down the line, someone who donates once will likely donate again. Based on this logic, if our goal is to get the most number of donations {not necessarily the highest amount], which variant should we choose? 10. Based on your answer in question 9, is party affiliation a confounding variable? 11. Based on your answers above, overall which variant do you believe is the best? Support your answer quantitatively [*Hint - try to figure out which variant meets most of the criteria for the questions posed). 12. Your calculations show that if the wrong link is shared to a potential Democratic voter, there is a strong chance it could result in them changing their vote to Independent or Republican. If you are currently using a pValue threshold of 5%, should you consider changing the threshold? If so, would you increase or decrease your p-Value threshold? Why? 13' Let's assume you wanted to re-run your calculations to find which age group 81 gender combination would result in the highest donations for each variant. Does this calculation make sense, and would this be possible with MB testing? If so, how? [You don't need to actually calculate this, just comment on how you would do it if, and if it's possible to do}