Orange County NC Website
Browse       Search
7 <br />Paper surveys were used. When the data collector was unable to obtain a completed survey <br />from the sampled customer, he /she attempted to survey the next customer instead. Situations <br />requiring this type of substitution included customers who did not speak English, refused to <br />participate in the survey, or did not live in Orange County. Substitution was discouraged and <br />data collectors were trained to be persistent and try to get at least the key piece of information <br />from the sampled customer. A survey was deemed "complete enough" if the data collector <br />was able to ascertain whether or not the customer lives inside town limits; substitutes were not <br />sought in situations where this information was determined, even if the respondent was unable <br />or unwilling to answer the remainder of the survey questions. The number of substitutions at <br />all sites was low, ranging from 0 to a maximum of 3 per site per day. <br />In total, 1,335 surveys were completed during the 2 week data collection period. The number <br />of surveys completed at each site by day of week is shown in Table 1. <br />Table 1. Number of Surveys Completed by Site and Day <br />Monday Tuesday Thursday Friday Saturday Sunday Total <br />[dhoa Qu. 31 27 37 31 126 <br />70 86 88 73 35 24 376 <br />55 44 65 61 225 <br />59 49 75 20 203 <br />ve 76 92 86 83 38 30 405 <br />146 323 174 276 250 166 1,335 <br />OCSWM staff reviewed the survey forms and keyed the data into an online data entry program <br />provided by the Odum Institute. Only six forms had to be omitted due to data collector error <br />making it unclear whether the customer lived inside or outside town limits; all other forms <br />were of good quality. <br />WEIGHTING <br />Data collectors recorded the total number of customers who entered each site each day of data <br />collection. These totals allowed us to generate precise weights for each survey date and site. <br />The weight was generated simply by dividing the total number of uses for the day /site by the <br />number of surveys completed for that day /site. Due to good fidelity to prescribed data <br />collection procedures, the weights varied little from the expected values based on the sampling <br />design —that is, a weight of "5" for surveys at most days /sites and "15" for Eubanks Rd and <br />Walnut Grove Church Rd on Saturday and Sunday. One exception was that the Bradshaw <br />5 <br />