The ranks in some tables are based on estimates derived from a sample(s). Because of sampling and nonsampling errors associated with the estimates, the ranking of the estimates does not necessarily reflect the correct ranking of the unknown true values. Thus, caution should be used when making inferences or statements about the states' true values based on a ranking of the estimates. As an example, the estimated total (average, percent, ratio, etc.) for State A may be larger than the estimates for all other states. This does not necessarily mean that the true total (average, percent, ratio, etc.) for State A is larger than those for all other states. Such an inference typically depends on --among other factors-- the size of the difference(s) between the estimates in question, and the size of their associated standard errors.
In other tables, the ranks are based on a complete enumeration of the target population, or on complete administrative reporting from the population. In such cases, sampling is not used, and there is no sampling error component in the estimates. Still, care should still be taken when making inferences or statements based on the rankings. The table values may still exhibit nonsampling error originating from such sources as coverage problems (missing units or duplicates), nonresponse, misreporting, and others.Last Revised: September 27, 2011 at 09:43:17 AM