Question

Questions for Exercise 34

Directions: For each of the following excerpts from journal articles, indicate whether the sample is biased or unbiased and explain your choice. In some cases, you may answer "not sure" because there is insufficient information to make a judgment. If you answer "not sure" to an item, describe the additional information you would like to have before making a choice.

1. Researchers state, "Smokers (defined as individuals smoking 10 or more cigarettes per day) were approached at various locations on or near the University of Pittsburgh campus. The sample (N = 102) included 42 men and 60 women, with an average age of   27.78."

2. Researchers state that "Participants were 307 students in a convenience sample of volunteers in 18 elementary and middle school classes. After we accounted for missing data, including student absences from school during 1 or more days of the intervention, the actual sample size for the statistical analysis was 268."

3. Researchers state that "Briefly, a stratified random sample of 4,907 coronary artery disease patients undergoing an angiogram at one... New York State hospital… was selected for the medical records abstraction portion of the study."

4. Researchers state that "We commissioned Survey Sampling Incorporated to provide a random sample of 1,000 Florida residents. Of the initial sample members, 198 had moved, had incorrect addresses, were deceased, or were otherwise unreachable. These individuals were replaced with additional randomly selected residents. Thirty-three of the replacements also could not be contacted but were not replaced. Thus, the total number... was reduced to 967.

5. Researchers state that "a standardized questionnaire developed for use in this study was mailed to a stratified random sample of 750 hospitals drawn from the American Hospital Association membership list. Hospitals were stratified according to size, geographic location, and stage of managed care development. ... "

Answer 1

1. Not Sure, as objective is not open.

2.Unbiased, because the researchers already assumed 10% drop-out rates.

3.Biased, as sample should from different hospitals where angiogram is going on based on their similar characteristics.

4.Unbiased as Non-Smpling error ie. Non-respondants happen here, which can be estimated.

5.Unbiased , basic theory matches