Exam 2 coupon users

Question 1

hypothesis test

conclusion (at alpha = .01)

Answer 1

go where it has independence to find p

alpha = .01 > p-value = .0001
reject Ho

Sufficient evidence to indicate that Type of customer and gender are related (dependent)

Question 2

why conducting a series of t test is dangerous strategy?

Answer 2

leads to an inflated probability of making at least 1 type 1 error

Question 3

explain theoretically the meaning of the phrase 95% confident

Answer 3

in repeating sampling, 95% of similarly constructed interval will contain the true loyalty score of the coupon user when age = 55 years and income = 70 thousand dollars.

Question 4

which regression statistic can be artifcially forced to 1 by simply adding independent variables (good or bad predictors) to the model?

Answer 4

R^2

Question 5

why aren’t all the actual values on my regression line?

Answer 5

we do not have a “deterministic relationship”: it is “probabilistic” (involves error)