Study of Prejudice Against American Indians
In October 2007, The Tulsa World published a story about a study regarding racism toward Native Americans. According to the researchers, who were from the University of Tulsa, “The findings support the idea that although overtly racist ideas toward African-Americans appear to be less prevalent in contemporary America, overt racism towards Native Americans is present."
One thing I found absolutely shocking about the study was its methodology. According to the published reports, the results "were from a written survey of 55 white, middle-class college students in their 20s at TU who had been in college for more than a year."
Excuse me? Fifty-five college students? I will admit that my statistics class is a distant memory, but I seem to recall rules and formulas for calculating sample size. Indeed, five minutes of searching on google not only revealed the basic rules, but turned up a "sample size" calculator. You input the confidence interval (the old standard "plus or minus 4 percentage points") and the confidence level (95% is standard), and it tells you what your sample size needs to be. I selected a confidence interval of 5 and a confidence level of 95%, and the calculator told me I needed a sample size of 384.
If I reduced the confidence interval to 13, I could make do with a sample size of 57. Can you imagine the political pollster reporting that candidate X has a 4 point lead in the polls, with a margin of error of plus or minus 13? What does that really tell us?
And why survey only students at the University of Tulsa? To get a truly representative sample, shouldn't students from other universities, at the very least the University of Oklahoma and Oklahoma State University have been included?
Fifty-five college students sounds like a class project, not a major research study. Fifty-five college students sounds like the initial springboard giving rise to a bigger, more comprehensive study, rather than the actual end of the line.
Given the problems with sample size, I would be curious to know the rest of the survey's methodology. How was the written survey administered? What were the study conditions? What questions, and how many, were asked? How were the answers caculated and analyzed?
The study itself is an intriguing one, and I would be absolutely interested in seeing the results of a properly conducted survey, or in seeing an explanation of why this survey was properly conducted. Perhaps I am wrong, but I am reluctant to jump to conclusions and accept the results as accurate, when for all we know, fifty-five hungry (and maybe hung over) college students were given a "check the box" quiz about product labels while waiting in line for breakfast Sunday morning - which is more offensive, the Land o'Lakes label or the Aunt Jemima label?