Design and validation of a children's food safety survey

Abstract

The purpose of this project is to design a survey instrument to study children's knowledge, attitudes, beliefs and behaviors related to 3 food safety issues: food additives, pesticide residues in food, and food biotechnology. Two drafts of questions for 11 and 12 year olds were developed to test 28 hypotheses and administered to two sixth grade classes--one in a rural low-income community and the other in an urban middle income community. Children were asked for two types of feedback: 1) their responses to each question and 2) a description of what each question meant to them. This method provided an opportunity to evaluate the construct validity of each question. Overhead transparencies of each survey page were used to help students track the discussion. As suggestions for improving the wording or response options of questions were elicited from students, changes were made on the transparencies to confirm our understanding of their suggestions. This method was followed for each question on the survey. Two survey versions of the third draft were developed and administered to two classes in another urban middle-income community. This tested the effect of question sequence for those questions that were consistently difficult for children to understand. The fourth and fifth survey versions were tested in schools with a high population of low-income, ESL students (primarily Latino). We continued to simplify questions and format. However for some questions, wording could not be altered without changing the meaning. The sixth version was critiqued by an elementary school reading specialist. The process we undertook to reach the final version of the instrument required 6 months time and repeated interactions with children from varied socioeconomic levels, geographic locations, ethnicity and gender. We learned that children are eager to assist in survey development when their participation is clearly valued by adults. The children are aware of fine distinctions in their feelings and experiences and want to have their responses to questions reflect those distinctions. To date, the survey and protocol have been tested with 200 children. The final product bears little resemblance to the initial version. The final survey and protocol is available for wide spread use.