Many researchers compare rates across populations or analyze trends over time. If groups or years differ with respect to factors associated with the event under study (e.g., age, sex, or race/ethnicity), the comparison of the overall or crude rate may be misleading. One popular way to address these differences in confounding factors is direct standardization, in which a weighted average of stratumspecific rates is calculated. It does not mean, however, that the effect of the considered factor (for instance, age) is eliminated; it is merely kept constant. The purpose of determining such a standardized summary measure is in the comparison with other rates adjusted to the same standard, not in the actual value itself. Several authors have demonstrated that if specific rates in the groups being compared do not bear consistent relations across strata (interaction present), any summarization procedure will yield interpretation problems. Therefore, standardized rates can never be a substitute for the analysis of specific rates but may serve as a convenient summarization measure. While each type of standard has its own profile that may guide one’s choice, the use of a widely available standard (European or World population standard) is strongly advocated in the analysis of routine data sources (mortality, hospital statistics, cancer registry data). This strategy enables other researchers to adjust their rates to the same standard, thereby facilitating the comparison and a more direct interpretation of differences. We manually searched all 1996 and 1997 issues of American Journal of Public Health, European Journal of Public Health, JAMA, and BMJ for articles using direct standardization and determined the type of standard used. In 19% (14 of 75) of these articles, we were unable to identify the standard that had been applied. Only 15% (11 of 75) of the articles used one of the available World Health Organization standards. Surprisingly, this percentage was only slightly higher, 27% (10 of 37), among studies using routine data sources. Because the choice of standard may affect the comparison, readers need to know which standard has been applied. In one fifth of all articles using direct standardization, we could not retrieve that information. Widely available standards to facilitate the comparison between studies using routine data sources were infrequently used.