Nonlinear Effect Amplification: Differential Susceptibility of Verbal Overshadowing as a Function of Time to Interference

Verbalizing visual memories can interfere with later accurate recall. Whereas changes in the magnitude of this verbal overshadowing effect (VOE) as a function of delay have been reported, no study has systematically investigated multiple shorter nonimmediate delays. Does VOE happen when verbalization occurs 5-min postencoding? 10 min? 15 min? We show in a preregistered study involving 4,501 American adults randomly assigned to different timing paradigms, that the effect size of VOE at 5 or 10 min is nearly zero, with a stable and significant inhibitory effect from 15 to 20 min. We further investigate this nonlinearity in a second study of 3,174 individuals showing a distinct nonlinear "effect amplification " sometime between 12 and 14 min. This apparent critical period after stimulus onset where susceptibility to verbal interference dramatically increases may help explain potential difficulties replicating VOE. More importantly, it suggests the possibility that the 12-14 min period may represent a critical window for other interference paradigms as well. General Audience Summary Eyewitness memory after seeing a crime committed can become corrupted by numerous sources. One source of corruption is describing the memory using words, called verbal overshadowing. Here, we show in two large preregistered studies that the effect of verbalization after seeing a crime occur nonlinearly amplifies from a null effect at short time intervals to a stable effect around 14 min after seeing the crime. This nonlinear effect occurs somewhere in the 12-14 min window. Future studies should investigate whether this 12-14 min window where effect sizes nonlinearly amplify is seen around other memory paradigms.