Similar to our biases sample experiments before, we create a biased sample of 800 claims with a fraud rate of 15 percent (versus an 8.
Further, we demonstrate that PRIDIT-FRE produces very reliable estimates of the population fraud rate even when only a very small and largely biased sample is available for use.
Again, the results are consistently robust; that is, PRIDIT-FRE predicts accurately the population event rate rather than the biased sample event rate.
2500 Note: This data set ("population") has 1,995 claims and the biased sample has 1,000 claims.
After the two biased samples are created, subsamples of different sizes are randomly taken from the biased samples and fraud classifications in these subsamples are then used in Equation (3) to estimate the population fraud rate.
In addition, the difference in mean ratings for the full sample compared to the biased sample was almost twice as large for those physicians in the lowest satisfaction quartile (.
With respect to selecting respondents to simulate a biased sample, level of satisfaction was the only variable considered in determining likelihood of responding.
Our analysis of actual patient satisfaction data suggests that the most satisfied patients may be the most likely to respond, and our simulation demonstrated how such a response bias might jeopardize the validity of interpretations based on a biased sample.
