Empirical Evidence Against Decision Augmentation Theory

Ð A reference on the Decision Augmentation Theory (May et al., 1995) includes a claim that certain data from the Princeton Engineering Anomalies Research program support the DAT model while refuting bitwise influence models at the 8.6s level. We present here an analysis of the entire PEAR database published in Jahn et al. (1996), which shows that the database as a whole is consistent with bitwise influence, and rejects DAT at well over the 5 s level in a linear regression test. The subset of the data used by May et al. is examined in detail, and it is shown that the 8.6 s figure results from an erroneous analysis procedure. Both the data set available to May, and the overall database, are strongly inconsistent with DAT predictions while remaining consistent with a bitwise influence model.