Episodic Memory
Regarding the 100 iterations bootstrap test’s performance on less familiar, Episodic information, we found that participants with an estimated BSITER score of 90% may present anywhere between 85.12% – 94.88%. In rare cases, the score may vary as much as +/-15% (75% – 100%). Once again, although participants with very large BSITER scores may be unlikely to encounter issues with diagnostic precision, participants who scored near the 85% threshold may be at risk for inconsistent classification.
Therefore, we tallied how many participants in the Episodic group had mean BSITER scores with 95% CIs that overlapped with the 85% cutpoint. Again, we found that 100 iterations produced reliable diagnoses for the majority of participants (21/29 or 72.41%); however, we had <95% confidence in over a quarter of the sample’s diagnoses (8/29 or 27.59%). All 8 of these participants had BSITER scores <= 90%. Importantly, 7/8 of these participants had BSITER scores equal to or exceeding the 85% threshold, meaning they were correctly classified as “guilty/knowledgeable”, but not with 95% confidence.
In conclusion, we found that the 100 iterations bootstrap test provided consistent and reliable results for the majority of our Episodic and Semantic groups (n = 62/81 or 77%); however, participants whose BSITER scores occurred near-threshold (approximately a quarter of our sample), were not classified with at least 95% statistical certainty. We recommend increasing the number of iterations used in the bootstrap test and/or calculating, reporting, and integrating the 95% CI for a participant’s BSITER score during individual classification analyses. We believe these steps are both important and necessary for improving statistical rigor in our field.