Orange County NC Website
u <br />more. A data collection period longer than one week would result in a large number of <br />customers being asked to do the survey a second, third, or even fourth time. Cooperation rates <br />would be expected to be poor in such situations, which would lead to missing data that could <br />potentially bias the survey results. Further, given that most customers use the SWCC at least <br />once a week, no new information would likely have been gleaned by a longer data collection <br />period: in expectation, since most users come to the SWCC once a week or more, one week is <br />the same as any other week, provided no holidays or severe weather situations cause unusual <br />aberrations. <br />Having identified one week as the appropriate period of data collection, we turned our <br />attention to how customers would be selected for the survey. In order to make proper usage <br />estimates, we would either need to survey every single customer (infeasible given the large <br />numbers of customers entering the site in a very short period of time at some SWCCs, or we <br />needed a scientifically rigorous way to select customers —a random probability sample. In a <br />random probability sample, every element of the target population (instance of use of an <br />SWCC, in this case) must have a known probability of being selected into the sample. For the <br />SWCC survey, the way to implement this would be to "count cars" as customers entered the <br />site, and to systematically survey every "nth" customer. The "n" needed to be a number that <br />generated sufficient surveys for a reliable survey estimate, but also one that the data collection <br />staff could feasibly implement. We estimated that it would take 2 -3 minutes for the data <br />collector to follow a selected customer to their stopping point, wait for them to get out of the <br />vehicle, and then administer the survey. At some sites at some hours, several vehicles may <br />enter the SWCC during that 2 -3 minutes when one data collector was unavailable. Through <br />discussion with OCSWM staff, it was determined that data collection staff could reliably survey <br />every 51h customer at most SWCCs on most days. The exceptions were Eubanks Road and <br />Walnut Grove Church Road sites on the weekend days when these sites receive heavy usage. <br />For those sites on those days, a sampling rate of every 15th customer was planned. These <br />sampling ratios were expected to yield sufficient surveys for reliable estimation of the <br />parameters of interest. <br />DATA COLLECTION <br />OCSWM staff hired, trained and supervised the data collectors. Surveys were conducted at the <br />Eubanks Road and Walnut Grove Church Road SWCCs for the one -week period spanning <br />Thursday, January 22 to Tuesday, January 27, 2015. (All SWCCs are closed on Wednesdays). <br />Surveys were conducted at Ferguson Road, Bradshaw Quarry Road, and High Rock Road SWCCs <br />on Friday, January 30 through Sunday, February 1 and on Tuesday, February 3 —a period <br />encompassing all operating days in a one week period for these centers, since they are closed <br />on Mondays, Wednesdays and Thursdays. <br />C! <br />