A. As a student taking Biostatistics, you are confronted with an exam problem wh
ID: 3151883 • Letter: A
Question
A. As a student taking Biostatistics, you are confronted with an exam problem where you are given two sets of data with equal numbers of observations. Also given is a list of values that are the differences between corresponding values between the two sets of data, for every subject that was part of a before stimulus - after stimulus experiment. (I.e., measurements were made on each subject before a stimulus and then after a stimulus; hence the two sets of values.)
A quick look at the differences shows that most values are positive; most differences are small in magnitude; but some differences are quite large (bringing doubt that data are normally distributed). You are asked if the there is evidence that the stimulus caused an increase response, such that even subjects with the highest baseline responses before the stimulus tended to rank lower than the subjects with the lowest responses after stimulus.
The most appropriate test to perform to answer this question (of the choices given) is
two-sample t-test
paired t-test
Wilcoxon-Mann-Whitney test
Wilcoxon Signed Rank test
Sign test
Randomization test
B. As a student taking Biostatistics, you are confronted with a problem as part of a lab exercise where you are given two sets of data with different sample sizes. You are informed that the values are parasite burdens (number of parasites) for organisms collected in two different geographic locations. The data tend to have a lot of small numbers like 0, 1, 2, 5 in both samples, as most subjects are not heavily parasitized. However, in only one sample, there are a few subjects with really large parasite burdens (e.g., 220, 400, 372).
You are asked if mean parasite burden might be greater in either geographic location. Uncertain about this question, you ask the instructor if it is ok to answer this question by determining if one population has a greater propensity to produce large parasite burdens than the others, based on collected samples, rather than compare mean parasite burdens. You are informed by the instructor that you are supposed to compare means, as the question asks. You do not feel comfortable with the violations of the assumptions for a parametric hypothesis test, so you seek a test that does not have such stringent assumptions, but allows you to compare sample means.
The most appropriate test to perform to answer this question (of the choices given) is
two-sample t-test
paired t-test
Wilcoxon-Mann-Whitney test
Wilcoxon Signed Rank test
Sign test
Randomization test
two-sample t-test
paired t-test
Wilcoxon-Mann-Whitney test
Wilcoxon Signed Rank test
Sign test
Randomization test
Explanation / Answer
(A) If the data is not NORMALLY DISTRIBUTED then WILCOXON SIGNED RANK TEST IS SUITABLE OTHERWISE PAIRED T TEST(IF DATA IS NORMAL)
(B) WILCOXON-MANN-WHITNEY 'U' TEST
Related Questions
drjack9650@gmail.com
Navigate
Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.