What is "Truth" in reporting survey results? Doctors were surveyed to find out what brand of butter substitute they recommend for their patients concerned about cholesterol. The results:
Recommend no particular brand: 80%
Recommend brand A: 5%
Recommend brand B: 4%
No other brand is recommended by more than 2 percent of the doctors. The firm owning brand A runs an ad that states: "More doctors recommend brand A than any other brand." Is this ethical? Why or why not? What kind of ethical guideline, if any, should be used to address this issue?
If a layperson would examine these results they would in fact come to the conclusion that yes, the data is correct, and ethical - numbers speak. However, when trying to understand statistics used in the media you have to be very cautious since the numbers are not what they seem.
Problem 1: Understanding statistical testing:
When you conduct research, your goal is to see if there is a significant difference between groups. The scope of this research would be to see if doctors significantly prefer one brand of margarine over another. To get statistical significance you need to perform a t-test. You get you test statistics, and you can then assess if there is in fact a difference between groups.
Usually, to get a significant difference between groups, you need to have a wide spread range of data, or have a small standard deviation, or have a very very large population.
Now, you can't make assumptions based on mean percentage. This raw data means NOTING unless you use it ...
This posting dissects the results of a survey to assess the ethical implications of the results. It examines the following topics: Understanding statistical testing; sample size, and selection; Validity of the results; Margin of error.