Reliability of Estimates

The standard error of an estimate is primarily a measure of the sampling variability that occurs by chance because only a sample is surveyed, rather than the entire universe. Because the NHAMCS is a sample survey, users should be aware of the reliability or unreliability of survey estimates, particularly the smaller estimates. NCHS considers an estimate to be reliable if it has a relative standard error of 30 percent or less (that is, the standard error is no more than 30 percent of the estimate). It should be noted, too, that estimates based on fewer than 30 records are also considered unreliable, regardless of the magnitude of the relative standard error.

Standard errors can be calculated for NAMCS and NHAMCS estimates using various types of statistical software packages including SAS, Stata, SUDAAN, SPSS, and others. They can also be approximated less accurately using coefficients derived from a generalized variance curve. The annual public use documentation provides such curve coefficients and approximate lowest reliable estimates for each data year based on the generalized variance curve. The documentation also provides more information about using masked sample design variables to estimate variances for NAMCS and NHAMCS estimates using a variety of statistical software packages.