While the MEPS-IC survey has a large national sample of establishments, the samples that support the state-level estimates are considerably smaller. When estimates are based on smaller samples, there can be a lot of variability in the resulting data. For this reason, all MEPS-IC estimates are published with standard errors. To determine if the changes in state-level estimates are statistically significant, you need to perform the appropriate statistical test for the difference between the two estimates of interest.
|