Knowledge transfer in public school food service: Post-implementation of the process approach to HACCP food safety system

Lydia Rebecca Frass


Based on Havelock's (Havelock & University of Michigan. Center for Research on Utilization of Scientific Knowledge, 1969) Knowledge Transfer Model and using Kirkpatrick's Training Evaluation Model (1996a, 1996b) levels of Reaction, Behavior, and Learning, this study analyzed knowledge transfer between public school district managers, cafeteria managers, and line workers. These employees were trained through cascade training methods during the federally mandated implementation of the Process Approach to HACCP food safety system beginning in the 2005-2006 school year. Measuring Behavior, an Observation Checklist (based on HACCP's 7 steps and 10 FDA foodborne illness risk factors and interventions) was used to determine if knowledge transfer occurred producing observable behaviors in line workers. The researcher's observations in a selected school district indicated that the district had implemented the system, line workers appeared to be properly following their district's plan and SOPs, and knowledge transfer seemed to have occurred. The Training Evaluation Assessment questionnaire consisted of 15 demographic, thirty-six 4-point Likert scale, and 11 matching items (measuring Reaction and Learning). Child Nutrition Program employees were mailed the instrument in North Carolina, Kentucky, Mississippi, and Wyoming. States were selected based on health regulations adopted by health agencies during 2005-2006 (1976 Model Foodservice Code Guidelines and 2005 FDA Food Code). A 2x3 Factorial Analysis of Variance (ANOVA), used to analyze for statistically significant differences in (learning) knowledge by Job Positions and Health Regulation Version, found significant differences for Job Positions. Multiple Analysis of Variance (MANOVA) tested for statistically significant differences in Reaction scores pertaining to factors of training (environment, materials, and relevance), training outcome, reaction to Process Approach system, supervisory support, and trainer effectiveness between Job Positions and Health Regulation Version. Job Positions had a significant main effect. Using the Bonferrroni method, each ANOVA was tested at the .007143 level. The two factors of Reaction to Process Approach and Supervisory Support were significant by Job Positions. Cascade training is still the quickest way to disseminate knowledge between multiple levels of workers. However, it may not be the most effective for long-term knowledge retention in an environment where hands-on, on-the-job training is most dominant and training resources are limited.