"Adults Who Stutter: Responses to Cognitive Stress," published in the Journal of Speech and Hearing Research in 1994, reported speech rate, in words per minute, for nine stutterers under conditions of stress.
a. If a comparison is to be made between these and speech rates under conditions of stress for nine non-stutterers, would this be a paired, two-sample, or severalsample design?
b. Would the data best be summarized by reporting means and standard deviations of speech rates under stress of stutterers and non-stutterers, respectively, or by reporting the mean and standard deviation of the differences between speech rates, pairing up each stutterer with a non-stutterer?
c. Would the data best be displayed with a histogram or side-by-side boxplots?
d. In order to draw conclusions about speech rates for all stutterers under stress conditions compared to non-stutterers under stress conditions, should we perform inference about md based on or about µ1 µ2 based on ?
e. Which form of inference should be performed first: a confidence interval for population difference in mean speech rates (stutterers versus non-stutterers), or a hypothesis test to determine if stutterers in general are different from nonstutterers in terms of their speech rates?