Client Survey (3.4)
Preparation for the client survey began with the selection of the A2H providers where interviewing was to take place. As previous A2H surveys had done, the client survey in the 2005 study focused on obtaining data on emergency food providers in the A2H system and on the people those providers serve. The three types of providers whose clients were included in the 2005 survey (and previous A2H surveys) were food pantries, emergency kitchens, and shelters. Many food banks also provide food to other types of agencies, such as those serving congregate meals to seniors and agencies operating day care centers or after-school programs. These other types of agencies perform important roles, but they were defined to be outside the purview of the study because they do not focus on supplying emergency food to low-income clients.
At the outset of the 2005 study, we asked the A2H food banks that chose to participate to provide MPR with lists of all the agencies they served, indicating whether each agency was involved in emergency food provision and, if so, what type of agency it was (pantry, kitchen, shelter, or multitype). MPR sampling statisticians then drew initial samples of the agencies where interviews were to take place. These selections were made with probabilities proportional to a measure of size based on reported poundage distributions as the measure of size; that is, large agencies had greater probabilities of selection.
After the initial sampling, MPR asked the food banks to provide detailed information for the providers or programs in the sample of agencies. The information sought included when they were open and the average number of clients they served per day. For small food banks (as classified by A2H), the sample of agencies for this detailed information was about 47. For larger food banks, the sample of agencies could be as high as 79. MPR then used the detailed information from the sample of agencies to form three pools of providers, and we drew samples of providers for the client interviewing. At this time, we also selected a reserve sample to account for possible refusal or ineligibility of a provider selected in the primary sample.
For each sampled provider or program, MPR selected a specific day and time when the interviewing was to occur, based on the detailed information the food bank had sent to MPR. We also provided a range of acceptable dates and times if our selection was not workable for the data collectors. The food banks were responsible for sending staff or volunteers to each selected program at the specified date and time to conduct the interviews. The data collectors were to use (1) the client selection forms developed by MPR and approved by A2H, and (2) an interview questionnaire that MPR and A2H had designed jointly. Clients at the facilities were selected for the interviews through locally implemented randomization procedures designed by MPR. In total, more than 50,000 clients were interviewed. MPR had another firm (a subcontractor) optically scan the completed questionnaires into an electronic database, and the resulting data files provided the basis for the client analysis.
During the fielding, we used randomly selected site replacements only when an agency, provider, or program refused to participate in the client interview effort or if, after conferring with the food bank and agency, we determined the provider to be ineligible for the study. In cases where food banks did not have reserve sample, we drew a supplemental first-stage sample and requested additional information or assigned an additional visit to a program among the programs already sampled. In some instances, we discovered while obtaining additional information that an agency (or provider) was no longer operating or did not run a pantry, kitchen, or shelter. In such instances, we dropped the agency (or provider) from the sample.
MPR prepared bar-coded labels with identification numbers for the client questionnaires. We also developed and printed, for use by interviewers, client selection forms designed to allow the interviewer to randomly select program participants and to enumerate the number of completed interviews, refusals, and ineligible sample members during on-site data collection. We shipped these materials and client questionnaires to food bank for distribution to the individual data collectors.
After data collection at a provider was completed, the food bank study coordinators shipped questionnaires and client selection forms back to MPR. MPR staff then logged each questionnaire into a database by scanning the bar-coded label on the cover page. As with the agency survey, each Monday morning MPR sent an e-mail to the members listing the agencies where client questionnaires were completed the previous week. The e-mails allowed the member study coordinators to monitor their progress in completing the client survey portion of the study.
After MPR received the questionnaires and MPR staff logged them into the database, the questionnaires were shipped to the subcontractor for data capture and imaging. The subcontractor optically scanned the questionnaires and produced data files for MPR. As with the agency survey, MPR received data files and electronic images of all completed client questionnaires on CD-ROMs. Chart 3.4.1 summarizes the sequence of activities in the client survey.
