What Are the Advantages and Limitations of CPS?

Advantages and Limitations of CPS

Advantages of the Current Population Survey (CPS)

It is conducted every year and the methodology is consistent from year to year (and from State to State), which facilitates its use in making comparisons over time.

48,000 households are interviewed each month and the response rate is unusually high (90 percent). This broad national sample makes national-level estimates reasonably reliable.

The large percentage of in-person interviews improves coverage and reliability (25 percent of the data is collected in person and 75 percent by phone).

A rotating sample is utilized in which half of the respondents of each year’s survey are interviewed again the following year. This is useful in measuring changes over time.

Limitations of the CPS

State-level estimates of the uninsured in the annual March CPS are unreliable due to small State sample sizes:
- Approximately 2,000-3,000 households are needed in a State sample to generate reliable estimates of subpopulations such as uninsured children below 200 percent of the Federal poverty level. In the CPS, there are only 11 States in which the sample size is more than 1,000 households. This results in large margins of error, especially for sub-populations.

The CPS may overestimate the number of full-year uninsured persons. Although the survey asks whether the respondent has been uninsured for the entire past 12-month period, participants may respond yes if they have been uninsured at any point during that period.

The CPS may underestimate the number of people on Medicaid. Based on studies conducted by the Urban Institute, HCFA administrative data show 20 to 30 percent more Medicaid enrollees than the CPS responses indicate. It has been speculated that respondents do not think of Medicaid as health insurance and that the data reflect this perception.

The CPS does not directly ask people whether they are uninsured. The survey asks about coverage of specific types of insurance and respondents who answer no to all of the categories are considered uninsured. There is concern that some respondents do have insurance, but were listed as uninsured.