It is well known in the survey literature that emotions can color survey participants' responses. Respondents' emotions play a pivotal role in how they perceive and remember behaviors and activities around them. Thus, self-report measures may provide 'biased' representations of reality in areas generally considered to be socially and emotionally sensitive. Using data from a cross-sectional survey and parallel, but independent, observations recorded by study personnel of the Nutrition Education Aimed at Toddlers (NEAT) intervention study, we examine the degree of discrepancy between these two sources of information. The convenience sample for this pilot study consists of 121 Early Head Start families in 28 counties of Michigan. Survey respondents were caregivers of toddlers averaging 28 years of age (range: 17-54), many of whom did not work (47%), were not married (63%), were white (83%), had only high school education or less (57%), tended to be overweight (25%) or obese (34%) and suffered from depressive symptomatology (CESD score of 16+: 38%). Discrepancies between caregiver interview responses and observations were divided into two categories: 'overestimates', if the caregiver reported that a particular activity occurred more frequently than observed by study personnel, and 'underestimates', if the caregiver reported a less frequent occurrence of an activity. For example, a caregiver might underestimate TV use, claiming it is "never" on, while the observations indicate it is "often" on. Twenty-two out of thirty-two activities revealed statistically significant discrepancies the self-report ratings and those of the observers. Depressed caregivers in particular were more likely to systematically underestimate signs of independence among their toddlers. Implications for mealtime behavior studies in particular and self-report measures in general will be discussed.
Session #1191 - Nutrition: Assessment & Intervention
The 29th Annual MNRS Research Conference (April 1-4, 2005)