STAT 701-Exam 2

The primary difference between a one mean t and one mean Z test is that the
(A) Z test is appropriate only when sample size is large (e.g. > 30) while the t test is only appropriate when sample size is small (e.g. < 30).
(B) one mean t test tests a more complex null hypothesis than does the one mean Z test.
(C) Z test requires knowledge of  while the t test does not.
(D) t test tends to be more powerful than the Z test.

A test that, under violation of some underlying assumption, produces a Type I error rate that is similar to the rate that would have been realized had no violation occurred is said to be
(A) liberal.
(B) anti-conservative.
(C) conservative
(D) robust.

A Type I error occurs when a
(A) false null hypothesis in not rejected.
(B) true null hypothesis is not rejected.
(C) true null hypothesis is rejected.
(D) false null hypothesis is rejected.

The probability of committing a Type II error is symbolized by
(A) α (alpha)
(B) 1-α
(C) β
(d) 1-β

Suppose that researcher A uses a one mean t test to test the hypothesis : μ = 50 against the alternative  : μ > 50 at  α= .05. The result is that the null hypothesis is rejected. Suppose further that researcher B using the same data tests the same hypotheses but chooses α = .01.Which of the following is true concerning researcher B's test?
(A) Researcher B also rejects the null hypothesis.
(B) Researcher B fails to reject the null hypothesis. 
(C) It is not possible to know whether researcher B rejects or fails to reject.
(D) Researcher B will inevitable make a Type II error.

Suppose researcher A forms a two-sided 95 percent confidence interval to estimate μ (the population mean). The result is L = .30 and U = .60.Researcher B, using the same data, performs a two-tailed test of the null hypothesis μ = .45. Which of the following is true concerning the result of the hypothesis test?
(A) Researcher B will make a Type I error.
(B) Not enough information is given to know if researcher B rejects or fails to reject.
(C) Researcher B fails to reject the null hypothesis.
(D) Researcher B rejects the null hypothesis.

Confidence intervals are used to estimate
(A) statistics.
(B) parameters.
(C) sampling distributions.
(D) standard errors.

14. A paired samples t test may be thought of as which of the following?
(A) McNemar’s test when it is referenced to a chi-square distribution.
(B) A two tailed Z test where  is known.
(C) A one mean Z test conducted on a small sample. 
(D) A one mean t test conducted on a set of difference scores.

In a study designed to compare a non-invasive method of monitoring dermal interstitial glucose with an invasive method of monitoring capillary blood glucose, one glucose assessment of each type was made on each of 50 Type I diabetic subjects. After performing paired samples t tests on a number of glucose measures and finding non-significance for each, the authors report, “There was no difference in a variety of glucose measurements including the mean glucose excursion, peak concentration and time course to maximal levels.”

Do you believe this was a proper interpretation of the results of their statistical tests? Why so?
(A) Yes, assuming the tests were done correctly and non-significance was found for each, the researchers have a good basis for concluding that there was no difference between the two monitoring strategies insofaras mean difference of the tested variables is concerned.
(B) No, because the researchers had no way of knowing what the Type I error rate for their tests was.
(C) Yes, because the researchers wold be able to control the probability of a Type I error for each of their tests.
(D) No, because the researchers had no way of knowing what the Type II error rate for their tests was.