03 - Laserfiche WebLink

City of Pleasanton Satisfaction Survey Proposal Page 2 voters, who were surveyed in 2007. Likely general election voters comprise about two-thirds of the total pool of voters in Pleasanton. Changing the sample from likely voters to all registered voters means the sample universe will include 100 percent of Pleasanton voters, broadening the universe of respondents from which we can interview, which will more closely approximate opinions among the general population of Pleasanton residents. By including in the sample voters who are less electorally engaged, the research will reveal a more general view of what residents think about the quality of life and issues in Pleasanton. Depending on the length of the questions, a 20-minute survey will include about 50-65 discreet question items. A voter list sample contains both listed and unlisted telephone numbers, as well k s cell phone and landline telephone numbers. A random sample of 600 respondents yields a m trgin of error of plus or minus 4.0 percent at the 95 percent confidence level. Having a statically randomized sample ensures that the results can be generalized to the entire survey population. Moreover, having a listed sample enables more than one person in the household over the age of 18 to offer their opinions if they are randomly selected. FM3 recommends interviewing at least 600 Pleasanton voters, which is the !,ame number of voters interviewed in the 2007 follow-up survey. In addition, we also recomme id oversampling Latino and Asian voters in Pleasanton. Census data show that while Latinos mace up 10 percent of Pleasanton residents, they comprise just six percent of voters; the disparity is even starker for Asians, who comprise 23 percent of the population but just eight percent of voters. This means that with a total sample size of 600, there would be as few as 36 Latino respondents and 48 Asian respondents—too few for drawing statistically robust conclusions about these groups' opinions toward City issues. So while this sampling methodology would accurately represent the City's voter demographics, we would not be able to draw many conclusions about the opinions of Latino and Asian residents, given the smaller sample size that would yield. Given the rising proportions of the population these two groups comprise, FPi13 recommends oversampling these two groups in order to interview 100 respondents identifying with each; this would allow FM3 to analyze opinions of these demographic groups independent ly, and offer the City more definitive research about each population's views toward the quality of life in Pleasanton. In Table 1 below, FM3 has provided two separate sampling schemes for the City to consider: Fairbank, Muslin, Maul/in, Metz& Associates FM3