Method of assessing parent-child grocery store purchasing interactions using a micro-camcorder

The purpose of this study was to assess the validity of using participant worn micro-camcorders (PWMC) to collect data on parent-child food and beverage purchasing interactions in the grocery store. Parent-child dyads (n = 32) were met at their usual grocery store and shopping time. Parents were mostly Caucasian (n = 27, 84.4%), mothers (n = 30, 93.8%). Children were 2-6 years old with 15 girls and 17 boys. A micro-camcorder was affixed to a baseball style hat worn by the child. The dyad proceeded to shop while being shadowed by an in-person observer. Video/audio data were coded for behavioral and environmental variables. The PWMC method was compared to in-person observation to assess sensitivity and relative validity for measuring parent-child interactions, and compared to receipt data to assess criterion validity for evaluating purchasing decisions. Inter-rater reliability for coding video/audio data collected using the PWMC method was also assessed. The PWMC method proved to be more sensitive than in-person observation revealing on average 1.4 (p < 0.01) more parent-child food and beverage purchasing interactions per shopping trip. Inter-rater reliability for coding PWMC data showed moderate to almost perfect agreement (Cohen's kappa = 0.461-0.937). The PWMC method was significantly correlated with in-person observation for measuring occurrences of parent-child food purchasing interactions (rho = 0.911, p < 0.01) and characteristics of those interactions (rho = 0.345-0.850, p < 0.01). Additionally, there was substantial agreement between the PWMC method and receipt data for measuring purchasing decisions (Cohen's kappa = 0.787). The PWMC method proved to be well suited to assess parent-child food and beverage purchasing interactions in the grocery store.