2008 Employee Survey Results - Laserfiche WebLink

Home Browse Search

City of Louisville Employee Survey July 2008 Appendix C: Survey Methodology Questionnaire Development The first draft of the 2008 Louisville Employee Survey was based on a previously tested survey of local government employees designed by NRC.The survey consisted of approximately four pages of thematically similar statements grouped into question sets and all asked on a scale with agree- disagree type responses (strongly agree, somewhat agree, somewhat disagree, strongly disagree),as well as a fifth page of sociodemographic questions. Louisville staff provided feedback on the survey and,consequently, some questions were revised and some new questions were added. Survey Administration and Response Rate On May 5,2008, all 143 Louisville City Employees were mailed a postcard inviting them to participate in the 2008 City of Louisville Employee Survey.The survey was to be conducted online using a Web address listed on the postcard invitation. Of the 143 City employees,all had email addresses. Employees received the first email invitation on May 8 and two reminder emails one week apart over the following two weeks.The email messages invited employees to participate in the survey and included a Web link. Centrally located work stations were set up with computers with Internet access for those who do not work at a computer on a daily basis. Shortly after employees received the final email reminder,department heads and supervisors were encouraged to remind employees to complete the survey.Data collection continued through June 12,2008. Of the 143 employees who received a survey invitation, 112 completed the survey,providing a response rate of 78%. Confidence Intervals The 95% confidence interval (or margin of error) quantifies the sampling error or precision of the estimates made from the survey results.The 95%confidence Ievel for the survey is generally no greater than plus or minus two percentage points around any given percent reported for the entire sample (112 completed surveys).The 95%confidence interval around a score on the 0-100 scale based on all respondents typically will be no greater than plus or minus one point on the 100-point scale. Other types of error such as non-response error (those who did not respond to the survey may have felt differently about the issues covered than those who did respond)is a source of potential bias in survey research.Multiple contacts were made with those selected for the survey to help address non- response errors. Survey Processing and Data Analysis Since the surveys were completed online,data were automatically saved electronically.The data were then exported into a text-only format and the electronic dataset was imported and analyzed by NRC staff using the Statistical Package for the Social Sciences (SPSS). For the most part, frequency distributions for each question and mean ratings for each factor are presented in the body of the report.A complete set of frequencies for each survey question,including"don't know"responses is -a presented in Appendix B:Complete Set of Frequencies.App endixA:Results Tables shows all frequencies to D survey questions presented in the body of the report and do not include the "don't know" responses. DRAFT Report of Results 42