The Dietary Screener in the 2005 California Health Interview Survey: Background
The screener used in the 2005 California Health Interview Survey (CHIS) was derived from the Five-Factor Screener in the 2005 NHIS Cancer Control Supplement (CCS). The 2005 CHIS screener asks respondents for information about how frequently they consume foods in 11 categories. No portion size questions are asked.
This screener does not attempt to assess total diet. The questions allow researchers to gather information about intakes of fruits and vegetables and teaspoons of added sugar. Fruit and vegetable intake is quantified using two different metrics. The Pyramid servings metric is based on the 1992 definitions of servings from the Food Guide Pyramid. The cup equivalents metric is based on the 2005 definitions, derived from Dietary Guidelines for Americans.
The 2005 CHIS Diet Screener is composed of QA05_C14 to QA05_C24 of the 2005 CHIS Adult Questionnaire The following variables in the 2005 CHIS Adult data were derived by the procedures outlined here: FV, FV_ADJ, FVNB, FVNB_ADJ, FVNF, FVNF_ADJ, FVNFB, FVNFB_AD, FVCE, FVCAD, FVCNB, FVCNBAD, FVCNF, FVCNFAD, FVCNFB, FVCNFBAD, SUG, and SUG_ADJ. Note that the variables SUG, SUG_ADJ, FVNB, and FVNB_ADJ were corrected/modified 2/27/2008.
In CHIS 2005, we applied rules for excluding extreme data responses, described in Definition of Acceptable Dietary Data. The process of scoring the individual response data is described in Scoring Procedures. A description and guidelines for the appropriate uses of the screener-estimated dietary intakes is found in Uses of Screener Estimates. Validation data for the CHIS 2005 screener are presented in Validity Results.
NOTE: The dietary variables on the CHIS dataset are in their natural units. For analyses, however, they must be transformed, first, to approximate normal distributions. For servings of fruits and vegetables and cup equivalents of fruits and vegetables, use the square root transformation; for teaspoons of added sugar, use the cube root transformation. After analyses, the result variables can be back-transformed for easier interpretation.