5. PROGRAM EVALUATION (continued)5.2.2 Survey Response Characteristics
Overall, 60.9 percent of the mailed surveys that were delivered as addressed were completed and returned. A slightly higher percentage of follow-on surveys (62.7%) were returned than of pre-program surveys (59.1%). Approximately the same proportion of returns was received from Phoenix (61.9%) as from Peoria (60.1%). Table 24 shows the number of returns for both waves and both cities as well as the percentage of delivered survey forms that were completed and returned.
Of the first wave of survey returns, 96 (50 from Phoenix and 46 from Peoria ) were received after the publicity and enforcement programs had started in the two cities. They were separated from the pre-program returns and identified in the database as “late returns” in order to be able to examine this small group of returns for possible changes in exposure to program activities.
The resultant study samples by survey wave are shown in Table 25. This table shows a remarkable consistency in returns for each wave for each city, that is, pre-program returns were 41.3 percent of returns in Phoenix and 43.6 percent in Peoria, late returns were 5.7 percent of the returns in Phoenix and 4.9 percent in Peoria, and follow-on returns were 53 percent of returns in Phoenix and 51.5 percent of returns in Peoria.
The survey form used in both cities is shown in Appendix D. It consisted of eight questions that asked the following of neighborhood residents:
Thus, the survey provided information on the following:
A database was created that documented the answers of each survey respondent. Each publicity/police/roadway event or activity that was specified by the respondent was post-coded to indicate whether the activity was perceived to be “definitely program-related,” “probably program-related,” or “not program-related (or program relation was unknown).” Assignment of these codes and the resulting analysis were therefore conservative because they were based on unaided recall of the Heed the Speed program and a high threshold for assuming a program relationship when post coding. As an example, comments that the respondent had seen police ticketing or general mentions of hearing messages were coded as “probably program-related,” although it is likely that most of the police enforcement and media efforts during the program period were program-related. Likewise, a mention of 3-D paint on the road or yellow signs was judged to be “definitely program related.”
Two additional subjective post codes were entered in the database. One provided an indication of whether or not the respondent had referred to the Heed the Speed program by name. For the second, each respondent was assigned one of three codes depending on whether the survey form as a whole indicated that the respondent had definitely been exposed to the program, had probably been exposed to the program or had not been exposed to the program.
An initial examination was made of the returns to determine if the late returns should be maintained as a separate file, combined with the pre-program or follow-on returns, or eliminated from the analysis. For this examination, responses to two survey questions were explored – respondent awareness of speed control activities and exposure to the Heed the Speed program.
The first survey question asked respondents if they were aware of any recent activities to control speeds in their neighborhoods. The responses for both cities are given in Table 26, which shows significant differences between survey waves for both cities.
In both Phoenix and Peoria , the table shows a significant increase in awareness of activities from pre-program data to late returns to follow-on data. It is interesting to note that approximately two-fifths of the respondents in both cities reported being aware of activities to control speeds in their neighborhoods before the program started. In Phoenix , 44.1 percent of the pre-program returns indicated awareness of activities, 74 percent of late returns, and 84.5 percent of follow-on returns. In Peoria , these data were 41.6 percent, 58.7 percent, and 82 percent, respectively. The data for late returns likely indicate a building exposure to program activities in the two cities.
Table 27 shows whether or not the respondent was judged to have been exposed to the Heed the Speed program. The table shows that only one of the pre-program respondents had probably been exposed to the program. However, 22 percent of the respondents who sent in late returns from Phoenix had probably or definitely been exposed to the program as had 19.6 percent of the late returns from Peoria . These data, though small in number, appear to indicate that awareness of the program was already present in the late return data. By the time of the follow-on survey, 69.5 percent of the Phoenix residents had probably or definitely been exposed to the program as had 70.8 percent of the Peoria respondents.
As indicated previously, the number of late returns was quite small. The foregoing tables, however, suggest that the late returns were somewhat different from both the pre-program and program samples. Since there was an insufficient number of late program returns to support analyses of meaningful subsets (e.g., gender, neighborhood) and it was inappropriate to combine them with either the pre-program or program data, they were omitted from further analyses. Subsequent tables therefore include only pre-program and follow-on data.
The information obtained from the surveys is presented in the following paragraphs. Unless otherwise noted, when no data were provided for a given item, the record for that item has been excluded from the summary tables. Significant associations were tested for selected survey results by means of the chi-square test. Differences reaching the 0.05 level of significance or less are noted.