1990 Citizen Survey Results - Laserfiche WebLink

Home Browse Search

rate compared to telephone or in -person interviews. Generally, single shot mail surveys produce a 20-35% response rate. With one or several mail or telephone follow-ups -- for example, sending a reminder letter and another copy of the survey to non - respondents, or calling non -respondents -- the response rate can be increased considerably. The final response rate depends on the number and type .of follow-ups used. Ideally, we would have sent out the survey, then, several weeks later, followed up by sending a reminder letter and a second copy of the survey to non -respondents. However, this would have been costly and, more important, could not be accomplished within the time frame allowed. To preserve options for the city with regard to a follow-up, a postcard was included with each survey sent out, and persons in the sample were asked to return their postcard under separate cover at the same time that they sent back their completed survey. This permitted us to track respondents without in any way violating their anonymity. Thus, although there was no time or provision for A follow-up mailing, the names of non -respondents are available if the city desires to conduct a follow-up in 1990. Since the city of Louisville controls utility billing, the original plan was to use the utility billing list to select a random sample of Louisville households. However, several problems prevented us from using this procedure. First, the city was unable to run mailing labels off the utility billing list, so that any sample selected from this list would have to be sent hand -typed labels. This would have been an extremely cumbersome process and, given the very short time frame for conducting this study, simply was not feasible. Second, we learned that non- resident owners were included on the utility billing list, which meant that the list underrepresented renters. This would have presented a sampling problem and therefore made the decision not to use the utility billing list a positive one. Instead, we used the voter's registration list. This list traditionally has been the best available and most widely used source for sampling adult citizens, and, in the case of Louisville, with its high voter registration rate, provides a very good sample for a citizen survey. In addition, since minorities, the group that is most likely not to be registered to vote, comprise a very small proportion of Louisville's residents, the list should produce only a limited degree of bias. We sent out a total of 1500 survevq 'to a random sample of Louisville's registered voters. Fortunately, the response rate from the single mailin was 421 which is surprisingly high for a 2