HarvestPlus Reaching End Users Orange-Fleshed Sweet Potato Project 2007

A 24-hour recall questionnaire was developed specifically for the survey and was administered by trained fieldworkers. A training manual and a food intake booklet were developed and employed, as appropriate, for the administration of the questionnaire. The multiple pass 24-hour recall interviews were structured into four steps (passes) to maximize respondent recall of foods eaten. The first step, the ‘quick list’, involved respondents supplying a broad description of all food and beverage items, including snacks, consumed in the previous day, commencing with the food or drink taken immediately after they woke up and ending with the last food or drink taken before going to sleep at night. In the second step, a detailed description of each food or beverage item on the quick list was ascertained through a series of questions and prompts (generated by interviewers) which were specific to each item. Respondents were asked to give information on the ingredients and preparation methods of mixed dishes. The third step was to estimate the amount of each food and beverage and their ingredients consumed. The final step was to review and check the recall responses in conjunction with food picture charts to clarify and confirm responses given about the previous day’s intake.
All days of the week (including the weekend) were proportionately represented in interviews across the sampling units to negate any day-of-the week effects on dietary intakes. The multiple-pass procedure allows for enhanced recollection of the various foods and beverages consumed, as well as improved estimation of portion sizes, by asking respondents to:
a) record all foods eaten by themselves on food picture charts;
b) instead of eating from a “common plate” as is the norm, serve food in individual bowls;
c) estimate the quantities of main staple food items consumed using replicas of these staples, such as various porridges (e.g. matooke, posho, and atap or kalo) to improve visual impression of the foods consumed.

Repeat dietary recalls were conducted on ten percent of the sampled households in each region on a nonconsecutive day, to allow for an estimation of the distribution of usual nutrient intakes by the population, thereby reducing variation due to day-to-day (intra-individual) differences in intakes. All interviews (including the 24-hour dietary recalls) were conducted in the first language of the person being interviewed whenever possible. As such, the instruments were translated into local languages, Luganda (for Kampala), Runyakitara (South-West), and Luo (North). Enumerators were recruited to match this need.