1994 Citizen Attitude Survey__Report of Findings

Home Browse Search

Louisville residents, young or old, tend to rate limited growth strategies higher than annexation strategies designed to encourage residential -commercial development. There is a slight tendency for the oldest age category to give a higher mean rating to residential -commercial growth strategies, although not significant statistically. Open space is a major preference for the young, while being important but not equally as much for the 45+ resident. The other five social -economic indicators (gender, education level, years of residence in Louisville, income level, and size of household) did not have an impact on response patterns, so similar tables for those indicators are not shown here. This is the general format in which the results of the questionnaire will be presented in this part of this report. Additional Data Analysis Part II of the report will provide a more in-depth discussion of the response pattern by social - economic indicator for each of the twelve scaled parts to the questionnaire. Usually, only one of these indicators has a significant imp . t on response patterns for each part. Finally, returned questionnaires frequently contained written comments. All written comments on the 345 returned questionnaires were entered into a data spreadsheet in order that they could be extracted and analyzed by groups or subsets. A summary of written comments associated with each of the twelve parts of the questionnaire are included in a companion report titled Survey Frequency Responses and Citizen Comments. Statistical Testing The statistical test that was used to obtain a better understanding of response patterns was the means test for analysis of sub -sample means. This test is useful in indicating if the mean values of subsets or categories of sample respondents, such as the three categories of age used in this report, are significantly different from each other to warrant the interpretation that they show distinctly different preferences toward the question items. There are many additional tests used with this analysis of mean values to determine if such an inference can indeed be made. These include a statistical test for homogeneity of variance for the subsets or categories being compared, as well as a test to determine what specific subset is the most unique. Statistical analysis has been used very conservatively in this report because such tests generally assume a fully random sample, are sensitive to variance from normal distribution within the sample subsets, and gene- "'y assume homogeneity of variance between subsets. Qualifying significance levels have be set very low (.00) due to the fact that randomization and sampling error can generally not be properly determined for mailed questionnaires. Consequently, a 12