Research that depends on the accuracy of people's responses to questionnaires -- called self-reporting -- is dubious if there is no independent or objective verification of those responses, say experts.

This was shown in a study that examined how accurately patients recalled the recentness and results of screening tests for cholesterol and Pap tests. The researchers found self-reporting failed to uncover one-third to more than 50 percent of those who needed such tests -- and failed to uncover more than half of those who had high or abnormal test results.

In a rural area of Australia, researchers asked adults in randomly selected households to recall whether or not they had a cholesterol screening test within the previous five years or Pap test in the previous two. They were asked whether or not the results had been normal.

The researchers were able to test the accuracy of the responses (with the patients' consent) by comparing the answers to the records of the two labs in the area that performed such tests.

• About one-third of those who reported they had received a cholesterol test had not -- amounting to about one fifth of the total who were not screened.

• More than one quarter of women self-reporting they had received a Pap test had not -- and more than half of those who did not receive the test did not accurately report that fact.

• Furthermore, almost half the individuals self-reporting normal cholesterol levels actually had elevated levels, and more than half of those with elevated cholesterol levels did not self-report that information accurately.

• One-tenth of the women self-reporting normal Pap test results actually had not had the test, and more than half the women who had abnormal Pap test results did not accurately recall that fact.

Source: Sallie Newell, et al., "Accuracy of Patients' Recall of Pap and Cholesterol Screening," American Journal of Public Health, September 2000.