Objective: Investigate differences in dietary recall accu- racy by interview content, retention interval, sex, and grade. Design, Setting and Participants: For a validation study during the 2010-2011 school year at 4 public schools, each of 48 third-grade children (half girls) and 48 fifth-grade children (half girls) was observed eating school-provided breakfast and lunch and interviewed about his or her observation day. For each grade, interview content was ‘‘diet only’’ for 24 children, and ‘‘diet and physical activity’’ for 24 children. Retention interval was same-day recalls in the afternoon (SDA) for 12 children per interview-content group per grade and previous-day re- calls in the morning (PDM) for 12 children per interview- content group per grade. Assignments were random. Outcome Measures and Analysis: Rates for omissions (foods observed but unreported) and intrusions (foods re- ported but unobserved) were calculated. A multivariate ANOVA was conducted for omission and intrusion rates, with interview content, retention interval, grade, sex, and their interactions as explanatory variables. Results: The interview-content-by-grade interaction was significant for intrusion rates (P ¼ .007), which were lower (better) for third graders for diet-and-physical-activity in- terview content but better for fifth graders for diet-only content. The retention-interval effect was significant (P < .001), with omission and intrusion rates lower (better) for SDA than PDM recalls. The sex effect was significant for omission rates, which were lower (better) for girls (P ¼ .02). Conclusions and Implications: Results concerning the interview-content-by-grade interaction and sex were unex- pected. The retention-interval results confirm previous findings. Shorter retention intervals should be used to en- hance dietary recall accuracy. Results provide empirical ev- idence emphasizing the importance of validation studies to investigate methodological aspects for improving child- ren's dietary recall accuracy. Funding: National Heart, Lung, and Blood Institute, NIH R21HL093406. O7 An Assessment of Tray Waste Measurement Techniques A. Hanks, PhD, ah748@cornell.edu, Cornell University, 17 Warren Hall, Ithaca, NY 14853; D. Just, PhD; B. Wansink, PhD Objective: To determine which of 2 tray waste methods provides the most accurate approximations of actual waste compared with weighing tray waste. Design, Setting and Participants: Researchers set up a measurement station in an elementary school cafeteria in western New York, where they carried out 3 waste mea- surement methods: photographing trays; reporting whether none, a quarter, half, three-quarters, or all of each selected item was wasted (quarter); and weighing the amount of each wasted item. At a later date, researchers examined the photographs and reported whether 0% to 100% of each item was wasted. Outcome Measures and Analysis: Researchers weighed a serving of each item offered in the cafeteria so that the vi- sual estimates for food waste could be converted to grams and compared with the weighing method. t Statistics of the difference between the weighing method and the other 2 methods provided evidence for each method's accuracy. Results: Except for green beans (P < .01), there were no statistical differences between the weighing method and the quarter method. Statistical differences between the weighing method and the photograph method were signif- icant for chicken nuggets and chicken strips (P < .01 in each case), applesauce (P < .05), 1% milk (P < .01), and fla- vored milk (P < .01). Conclusions and Implications: Visual methods—pho- tograph and on-site methods—for measuring tray waste are less costly than weighing waste of each item but vary in accuracy. Photograph methods are useful but limit food visibility, whereas on-site visual estimates are much more accurate. Funding: USDA, Economic Research Service 59-5000-0- 0090. O8 Enhancing Quality of EFNEP Data: Designing a Diet Recall Form for the Low-literate Participant in a Group Setting M. Townsend, PhD, RD, mstownsend@ucdavis.edu, University of California, Department of Nutrition, UC Cooperative Extension, Davis, CA 95616; C. Schneider, PhD, RD; C. Ganthavorn, PhD; M. Johns, MPA, RD; M. Neelon, MS, RD; C. Lamp, MS, MPH, RD; B. Roche, MS, RD Objective: Design a 5-step multipass dietary recall form for a group setting. Use of Theory or Research: Visual Information Process- ing and Cognitive Load Theories. Target Audience: Sixty Expanded Food and Nutrition Education Program (EFNEP) educators and 9,000 EFNEP participants in California. Description: Educators consistently reported that the re- call data collection process was tedious and difficult with a group and with low-literate learners. Improving the func- tionality of the recall process was accomplished by chang- ing the name, reducing the respondent burden, and easing administration difficulties for the educator via use of color coding and design elements. A portion-size game initiated the recall process. To promote reliability, a 20-page educa- tor's instruction guide provided detailed instructions that include a script, background information for educators, potential participant questions, and recommended re- sponses. The recall process was supported by 2 additional tools: the EFNEP Food Descriptions Guide and Portion Size Guide. Evaluation: Educators (n ¼ 54), county advisors (n ¼ 8), and, indirectly, participants guided the development of the new form, using qualitative methods. Educators voted for a less academic name, Food Tracker. Color coding helped educators guide participants to the same step. Three-, 4-, and 5-syllable words were replaced. The O6 (continued) Journal of Nutrition Education and Behavior  Volume 44, Number 4S, 2012 Oral Abstracts S15 Continued on page S16