FIGURE H-8 Boxplots of accuracy index (A) values for datasets from laboratory studies, by funding source.

NOTE: ACR, academic research organization (n = 19); EO, externally funded by agency without polygraph program (n = 14); EP, externally funded by agency with polygraph program (n = 5); IO, internally funded by agency without polygraph program (n = 1); IP, internally funded by agency with polygraph program (n = 13).

Accuracy in Field Studies Figure H-10 displays A values for the 7 field datasets (i.e., specific-incident polygraphs) discussed in Chapter 5, in the same manner as Figure H-3 for the 52 laboratory datasets. As noted above, the median value of 0.89 is roughly the same as the median of 0.86 for the laboratory datasets, with the difference about what might be expected from the more frequent use of trapezoidal estimates of area for the laboratory studies. The standard errors are a bit smaller for the field studies (which have larger sample sizes) than for the laboratory studies, with that for the first study shown in Figure H-10 artificially small due to the proximity of the estimate to the maximum. The two studies with lowest values of A were done 20 years apart, one in the late 1970s and the other in the late 1990s. The five other studies were done in 1988-1991.

The types of funding used to support of these studies illustrate why the categories we used, which are reasonable and quite distinct at face



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement