Discuss the types of hypothesis tests we have learned to date. When are each applicable?

1) one sample z-test

2) one sample t-test

3) matched pairs t-test (correlated groups)

4)two sample t-test (independent groups)

5) tests of proportion

On slide 16 of lecture 8, the test statistic is calculated to be 0.39 which is NO WHERE near the size needed to reject Ho.

That is mainly because of two things:

1) the difference in the mean life length being 'only' 13 hours.

2) the size of the standard deviations, and thus variance.

Assume that the variance holds to be about where it is for each machine at each additional test. How big of a difference would we need in the average life length for the two machines in order to reject Ho? ie, when would the difference become 'too much'?

slide 16

Ho: 41 = #2 Ha: M1 # #2 Test Statistic: ($1- X2) - (41 -H2) (375 - 362) - (0) (13) - (0) t = X = 375 1102 1252 x2 = 362 25 + 25 12100 15625 n2 S = 110 25 25 $ = 125 13 13 13 n= 25 = 33.3 0.39