The weakness of my study severely impacted the validity of my results. Of coarse, with a survey I had non-response bias. However, my initial goal was to get a sample of 50 people. Due to non-response bias, my sample size was only 36. This significantly decreased my cell counts, making some of them less than five. I cannot make a valid conclusion form my significance tests. In addition, my sample size was also reduced due to the fact that I had such a small amount of blue collar workers. This could be due to the fact that half of my subjects were white collar workers, high school teachers. This was due to convenience. Moreover, my sample was not truly representative of the Cleveland area and its surrounding counties (Lake County, Medina, Erie County, Cuyahoga County, Ashland County). The other half of my sample was all of North Olmsted adults. This, again, would not be representative of the Cleveland area and its surrounding counties.
Extrapolation
If my results had allowed for a valid conclusion, I would have been comfortable extrapolating my results to North Olmsted, Cleveland, Cuyahoga County, lake County, Erie, county, Medina, and Ashland county. Ohio has always been a swing state; geographically there are multiple very conservative and vey liberal pockets. Therefore, I could not have been able to extrapolate my results beyond this area, for I only surveyed within this area. however, since I cannot make a valid conclusion, I cannot extrapolate any of my results.
Further Work and Research
Before I sent out surveys, I researched recent (2008 and 2004) presidential elections' exit polls in Ohio. This gave me an idea of where the biased pockets were, and was the basis for where I had initially hoped to extrapolate. Current and historical county polling results would allow me to, if my results had been valid or significant, compare my results to real ones. Also, these in consideration with he current political climate would have given greater insight into the results, and would have led me to ask more questions, or even question my results.