Regular high consumption of alcohol in selected populations have, with high precision, been identified by two new alcohol markers; carbohydrate-deficient transferrin and mitochondrial aspartate aminotransferase. To test these markers in an unselected population, gamma-glutamyltransferase (GGT), carbohydrate-deficient transferrin (CDT), and mitochondrial aspartate aminotransferase (mAST) were measured in the Norwegian population, 310 males and 171 females, aged 18 to 60 years, living at Svalbard. Using self-reported alcohol intake as gold standard, sensitivity, specificity, positive predictive value, and likelihood-ratio were estimated according to different cutoff-points for alcohol intake and for the tests. In contrast to earlier studies, the sensitivity was in general low. With a specificity of 90% or higher, the sensitivity did not exceed 26% for any of the tests. Whereas CDT showed its best discriminatory power at lower intake of alcohol, GGT discriminated best at higher levels. Parallel and serial analysis of CDT and GGT indicated a conditional independence between the tests, as well as at higher and at lower levels of alcohol consumption. mAST was judged as not suitable in population studies.