Nutrient data analysis techniques and strategies

Analyses of nutrient data pose special challenges to investigators. In such analyses, investigators need to consider: 1. Possible over-or under-reporting of intakes, leading to “impossible” or extreme values in the data set 2. How to adjust for total energy intake 3. How to model nutrients, e.g., as continuous or categorical variable 4. How to avoid multicollinearity, particularly when nutrients are expressed in absolute amounts, e.g., grams/day 5. How to analyze dietary supplement data 6. How to account for large day-to-day variability in intakes, which can lead to misclassification of individuals with respect to usual intake The objectives of this section are to examine various approaches to addressing the above issues; to briefly describe the common types of observational and experimental studies that collect nutritional data; and to describe the most common methods of analysis used in the types of studies described.