Margin of Error, Statistics Problem Example

The following poll was taken from the Environment website of the polling report.com. The poll can be found at the following URL: http://www.pollingreport.com/enviro.htm. The poll dealt with a person’s view on preventing laws that require the Environmental Protection Agency (EPA) to enforce regulations regarding greenhouse gases and other environmental issues. The website reported the following results from a CNN/Opinion Research Corporation Poll on April 9-10, 2011. The sample size was 824 adults. The adults were polled nationwide. 28% of the polled individuals indicated that they wanted to prevent the EPA from spending money on the enforcing regulations. 71% of the polled individuals indicated that they preferred the EPA to continue funding to enforce laws and 1% of polled individuals were unsure. The Margin of error for the results was +3.5.  Therefore, the results can be interpreted that 28% of individuals plus or minus 3.5% (231 individuals plus or minus 29 individuals) do not want to continue funding to EPA to enforce laws. 71% plus or minus 3.5% (585 individuals plus or minus 29 individuals) want continued funding and 1% plus or minus 3.5% (8 individuals plus or minus 29 individuals) are not sure either way.

There does seem to be a significant difference in response between all three of the results. The poll jumps from 28% of individuals not wanting funding to 71% wanting funding. In addition, there are only 1% of the polled individuals that were not sure. Statistically, there could be a significant difference; however, statistical analyses, such as an ANOVA should be completed in order to verify the significant difference. Furthermore, there were only 824 individuals nationwide polled for the results. This is such as small sample size when looking at the country as a whole. Perhaps are larger sample size would alter the margin or error and give a better understanding on the nations view. A larger sample size could be more valid as well when conducting statistical analyses to indicate whether the results are significantly different.