BACKGROUND: The valid estimation of the usual dietary intake remains a challenge till date. We applied the method suggested by the National Cancer Institute (NCI) to data from the 2nd Bavarian Food Consumption Survey (BVS II) and compared it to an individual means approach. METHODS: Within the cross-sectional BVS II, 1,050 Bavarian residents aged 13-80 years participated in a personal interview and completed three 24-h dietary recalls by telephone interview. For the 13 main food groups and 23 subgroups the usual intake was calculated by (1) an individual means approach and (2) by the NCI method. RESULTS: The distributions derived by the individual means approach are wider than those derived from the NCI approach. For a majority of food groups and subgroups, the proportion of participants who meet the dietary recommendations published by the German Nutrition Society is higher when the NCI approach is applied. The proportions of participants above or below recommended amounts differ greatly for "meat and meat products" and "cheese." CONCLUSION: The mean intake at the groups level can easily be derived from the individual means approach. Since only the NCI method accounts for intra-personal variation, this method provides more valid intake estimates at the individual level and should be applied when, for example, individual intakes are compared with dietary recommendations.