The evaluation of a Web-based food frequency questionnaire
Lane, Kimberly Gibson
MetadataShow full item record
My Food Choices.com is a web-based food frequency questionnaire developed for use with African-American fourth graders to assess dietary intake of the previous week. The questionnaire includes digital food photographs to assist in recall of recent intake. This research evaluated the My Food Choices.com questionnaire in a series of three studies. First, the questionnaire was validated against multiple 24-hour dietary recall interviews. The second study was a reproducibility study in which the reported food frequency questionnaire intakes on two occasions were compared. The validation and reproducibility studies were conducted with 132 African-American fourth graders from urban public four schools. These participants completed a questionnaire and participated in a dietary recall interview on two consecutive Saturdays as well as participated in a dietary recall interview during the school week. In a final study, the web-based questionnaire was compared to a paper-based version of the questionnaire. This study was conducted with 1252 fourth graders. All of these students were participating in a pre- dietary intervention data collection. Pearson correlations were used to for the between instrument comparison in the validation study as well as between the two occasions in the reproducibility study. In the comparison between two versions of the questionnaire, analysis of variance, analysis of covariance and multiple regression analysis were applied. The questionnaire had poor validity (r = 0.06 to 0.42) as compared to the validation standard of three 24-hour dietary recalls. This correlation range is similar to published traditional food frequency questionnaire validation studies with pre-adolescents and 24-hour dietary recalls as the validation standard. However, the questionnaire had moderate reproducibility correlations (r = 0.45 to 0.72) between the two administrations. For the paper-based and web-based comparison, no significant differences were found in reported intake between the two instruments when differences in school membership were controlled. The average number of reported servings was lower for all food categories with the web-based questionnaire; however the difference was not statistically significant. In summary, the questionnaire performed poorly as compared to the validation standard, performed moderately well when administered on two occasions, and yielded lower mean intakes than the paper-based version.