Sample Surveys for Monitoring New Hampshire Health Insurance

Sample Surveys

Most reports about insurance coverage in New Hampshire are based on surveys of New Hampshire residents. The annual Current Population Survey of the U. S. Census Bureau collects information each year on the health insurance status of about 1,400 New Hampshire households and uses that sample as the basis for estimates about the whole state. The NH Department of Health and Human Services (DHHS) sponsored surveys of many more New Hampshire households in 1999 and 2001, generating somewhat more precise estimates of the status of the whole population. In 2003, the Endowment for Health and the Healthy New Hampshire Foundation sponsored a survey that covered about 750 households.

Surveys that reach only a small part of a population are subject to what statisticians call "sampling error," the probability that the sample surveyed is not exactly representative of the whole population. That's why statisticians typically report their results as a point within a range of error: for example, "12 percent, plus-or-minus 4 percentage points" means that the statistician is reasonably sure that the actual answer in the total population is somewhere between 8 and 16 percent.

One of the surveys conducted in New Hampshire reported that the percentage of the population that is uninsured is 9.0 percent, with a 95 percent confidence interval of +/- 2.0 percent. That means that the actual "uninsurance rate" in the total population is probably in the 7.0 percent to 11.0 percent range, although there is even a 5 percent chance that it is outside this range. If the population in a subsequent survey shows 9.8 percent uninsured with a +/-2.0 percent confidence interval, the actual rate in the full population is probably somewhere in the 7.8 percent to 11.8 percent range. Accordingly, the apparent increase of 0.8 percent in the rate of uninsurance among those in the two survey samples may not reflect any trend in the population, but only the difference in the samples.

Surveys do not give a sufficiently refined measure to determine if the uninsurance rate in New Hampshire is rising or falling. A review of the surveys conducted in New Hampshire by the U.S. Census Bureau, the NH Department of Health and Human Services, and the Endowment for Health shows that the changes they have reported in the uninsurance rate have been too small to distinguish from the background noise of sampling error.