CONCLUSIONS AND NEXT STEPS
This report describes our analysis of written federal district court opinions, an analysis carried out to provide insight into how judges evaluate and screen expert evidence in the wake of the Supreme Court's 1993 Daubert decision. The findings are based on systematic coding of a substantial number of district court opinions between 1980 and 1999 and statistical analysis of the results. In this concluding section, we summarize our key findings and the conclusions we draw from them. We also identify important gaps that remain in understanding how judges screen expert evidence, as well as research that would help fill those gaps.
Our analysis of district court opinions suggests that following Daubert, judges scrutinized reliability more carefully and applied stricter standards in deciding whether to admit expert evidence. After Daubert, the proportion of challenged evidence in which reliability was discussed and the proportion of expert evidence found unreliable rose. These increases do not appear to be due to a mere shift in terminology: the proportion of challenged evidence excluded rose following Daubert, indicating that the change in standards had real impacts. While the standards appear to be more stringent than what existed earlier, we do not know whether they have led to improvements in the quality of evidence admitted or to exclusion of evidence that should have been admitted. Our findings also suggest that the change in standards was not limited to a relatively small number of judges. However, additional information is needed to determine how evenly the changes were spread across the federal bench.
Considerations That Enter into the Admission Decision More Closely
The Daubert decision did not change the standards for the relevance of expert evidence, the expert's qualifications, and other considerations that enter into a judge's assessment of whether to admit expert evidence, but it did affirm that judges should act as gatekeepers. Our findings suggest that judges scrutinized evidence with increasing care after Daubert with respect to these criteria. We found that the success rate for challenges based on all criteria (i.e., the percentage of evidence rated unfavorably when the criterion was addressed) rose after Daubert, although the increases did not always achieve statistical significance. It appears that once judges