Evaluating the Dietary Guidelines: The Development of an Evaluation Tool

Purpose

Due to the length of most valid nutritional questionnaires, e.g., Healthy Eating Index, dietary analysis and prescription is often time consuming and unwieldy. The construction of a succinct, less cumbersome nutritional screening instrument that might also serve as a referral tool for further nutritional analysis might alleviate the time and tedium associated with its more burdensome alternatives. The purpose of this study was to develop and field-test a concise, yet valid and reliable nutritional behavior survey to serve as a basic dietary screening tool.

Methods

A 26-item, close-ended, questionnaire configured on a five point Likert scale was developed. A panel of experts reviewed and revised the questionnaire to ensure validity; a post-hoc Kaiser-Meyer-Olkin measure of sampling adequacy (.82) established reliability. Data were collected from 421 university students enrolled in a fitness for living course at a mid-western university. Instructions were delivered by e-mail, and subjects were forwarded to an InQsit website where they completed the survey.

Analysis/Results

Mean scores (M) and standard deviations (SD) for each questionnaire item were determined and rank ordered. Highly ranked-consumed items were: breads, pasta, grains, poultry, and fish. Intermediate ranked-consumed items included: vegetables, fruit, prepackaged-precooked meals, and dairy. Low ranked-consumed items included: tofu, peanut butter, fast foods, eggs, frozen or dried fruit and meat with visible fat. M and SD ranged from highest consumption, 4.13 (0.83), for bread or pasta to lowest consumption, to 1.77 (0.94), for meat with visible fat. Exploratory factor analysis using principal factor solutions with iterations and varimax rotations returned six manageable factors with eigenvalues greater than 1. Factors 1, 2, 3, 4, 5 and 6 had group M of 2.57, 2.65, 2.57, 2.57, 2.91 and 3.89; each accounted for 18.96%, 13.86%, 5.93%, 5.64%, 4.79% and 3.94% of the explained variance respectively. Factor labels with group mean ranks, e.g., Factor X (2), were: Factor 1 – fruits/vegetables (tie - 4), Factor 2 – meats (3) Factor 3 – healthy choices (tie - 4), Factor 4 – unhealthy choices (tie - 4), Factor 5 – dairy (2) and Factor 6 – grains (1). These factors accounted for 53.13% of the total explained variance.

Conclusions

This questionnaire yielded a concise yet valid and reliable measurement of the nutritional practices of this population. Suggestions for refining the questionnaire and increasing its ability to account for additional explained variance were: include more survey items reflective of grain intake, and eliminate/re-word items that represent combined nutritional groups.