Poll on Name Change

[Guest guest]

Pam: To use statistical techniques you must randomly select people for the poll, otherwise you can't make inferences about the population based on the sample. To be 95 percent confident that a least 75 percent of the members want a name change at least 80 percent of the people polled out of a sample of 200 would have to vote in favor of the name change. If you took a smaller sample of 52 then at least 85 percent of those polled would have to be for the name change. The above numbers are based on testing the null hypothesis that less that 75 percent of the people would prefer no name change against the alternative hypothesis that 75 percent or more of the people would prefer the name change. This is a one-tailed test using the normal approximation (z statistic) to the binomial distribution. Since you already have some responses to your poll I would just use what you have and make a decision based on the results. I don't think its worth the effort to conduct another poll by randomly selecting people and if you did I think you would arrive at the same conclusion as with the existing poll. The sample sizes and the percent in favor requirements stated in the previous paragraph can guide your decision. If you haven't had much of a response (at least 100) to your poll so far I would conclude that it is not a big issue to most people and leave the name the way it is. If more than 100 people have responded than I would require that at least 80 percent be in favor of the change. Although 100 or more respondents out of 690 is a lot for most polls I don't think it is too much for a simple e-mail poll. Good Luck Dick Swanson