to be 445 mg/day (139 percent of the RDA) to have a 97.5 percent confidence that intake was adequate with only one day of observed intake. However, a mean observed intake of only 349 mg/day (109 percent of the RDA) would be needed with 7 days of observed intake.

• For a given confidence level, the larger the SD of daily intake, the greater the intake level needed for intake to be assessed as adequate. If the SD of magnesium intake were 25 percent larger, then intake would need to be 486 mg/day (152 percent of the RDA) to have a 97.5 percent confidence of adequacy with one day of observed intake, and 362 mg/day (113 percent of the RDA) with 7 days. If the SD were 50 percent larger, then the intakes would need to be still higher to have 97.5 percent confidence of adequacy.

To simplify this approach for nutrition professionals, institutions, and agencies may wish to establish levels of intake that they consider adequate for a given nutrient. For the example shown here, a level of 377 mg/day might be chosen as the level of adequacy of magnesium intake for women 31 to 50 years of age, by an institution that typically collects three days of dietary data for its patients, and wanted a high level of confidence (97.5 percent) that intake was adequate.

To summarize, despite the fact that neither individual requirement nor usual individual intake is available for dietary assessments of individuals, some inferences about individual adequacy can be made by looking at the difference between observed intake and the median requirement. Shortcomings of this approach are described in Appendix B. For example, the approach cannot be used when observed daily intakes are not normally (or symmetrically) distributed around the individual 's usual intake. An indication that the within-person intake distribution is not normal (or symmetrical) is the size of the within-person standard deviation in intake relative to the mean intake. When the SD of daily intake is high enough so that the CV of daily intake is larger than approximately 60 to 70 percent, then the approach proposed here is not appropriate. Appendix Table B-2 and Table B-3 indicate that for vitamin A, carotenoids, vitamin C, and vitamin E, among others, the CV of daily intake is very large, above 70 percent. For those nutrients, it would be incorrect to apply the method described in this section to assess adequacy of an individual's diet. At this time, no alternative can be offered, as much research is needed in this area.

It is also possible to calculate observed nutrient intake levels with an 85 or 97.5 percent confidence of inadequacy. Intakes with a high probability of inadequacy are below the EAR. For confidence (at