Authors' reply

摘要

We agree with Winkler that the interpretation of dietary data requires appropriate perspective, context, and recognition of strengths and limitations, as is true for any scientific research. But Winkler's concerns reflect only a partial understanding of the strengths and limitations of various nutritional assessment methods. The interpretation and validity of individual level dietary data depend on the dietary assessmenttool used (such as diet records, diet recalls, semi-quantitative food frequency questionnaires, tissue biomarkers) as well as the level of assessment required (such as individual intake vpopulation orgroup mean intake) and the nutrient, food, or dietary pattern of interest. In relation to these considerations, various dietary instruments have different strengths and weaknesses, and theirvalidity and reliability have been extensively documented.12 Overall, the validity and longterm reliability of food, nutrient, and diet patterns assessed by such instruments are reasonable, similarto other widely used measures such as blood pressure and blood cholesterol. Furthermore, when used to assess population means, as in our study (ratherthan to assess individual intakes), these dietary instruments have even strongervalidity, especially when using diet records or recalls, which were the main instruments used in ouranalyses.