Diet Screener in the 2005 CHIS: Definition of Acceptable Dietary Data Values
Data collected on the 2005 California Health Interview Survey (CHIS) Screener are coded by frequency and time unit -- times per day, week, or month. We used USDA's 1994-96 Continuing Survey of Food Intakes of Individuals (CSFII) data on reported intakes over two days of 24-hour recall to make judgments about reasonable frequencies of consumption that were reported on a per day basis. This helped us assess values from the CHIS Screener, some of which are highly unlikely.
Maximum daily average frequencies (averaged for each individual across the two days of report) in the CSFII ranged from 1.5 times to 12 times per day for the relevant food groups. We accepted frequency values that were reported in the CHIS on a per day basis up to the maximum average values (rounded to the next whole number).
|Food Group||Maximum Daily Acceptable Value|
|Other white potatoes||3|
|100% fruit juice||4|
|Cookies, cake, pie||5|
In addition, we applied judgment to determine the acceptability of frequency reports for the weekly and monthly time periods (see below). For example, a report of 25 times may be most logically associated with a month or year time period, but not so logically associated with a week time period. We applied this judgment to all foods.
|Time Period||Acceptable Frequency|
|Week||≤ 14 / week: acceptable
> 14: assign a missing value
|Month||≤ 60 / month: acceptable
> 60: assign a missing value
Depending on the intent of the analysis, a researcher can exclude a person with a missing value for any of the 11 foods, or with missing values only on foods needed to estimate a particular dietary intake variable. In our analyses of CHIS data, we excluded individuals only for the dietary variable for which they had missing food-level data.