The effects of confirmatory interviewer feedback on eyewitness testimony following forcibly confabulated and accurate responses to repeated interview questions were investigated in two experiments. The first experiment showed that, relative to neutral feedback, confirmatory feedback provided after a forcibly confabulated response greatly increased the likelihood that participants would provide the same confabulated response when re-interviewed 2 days later, led participants to report these repeated confabulations with greater speed and fewer expressions of doubt, and increased the prevalence of false memories. Confirmatory interviewer feedback provided following accurate responses appeared to have more modest consequences for consistency and confidence, but ceiling effects provided little opportunity for observing potential effects. A second experiment showed that these effects of confirmatory feedback are of considerable practical significance, in that, regardless of their accuracy, responses that had earlier been reinforced with confirmatory feedback were much more likely to be judged by others as credible.

Children tend to choose an entity they cannot already label, rather than one they can, as the likely referent of a novel noun. The effect of input that contradicts this strategy on the interpretation of other novel nouns was investigated. In pre- and posttests, 4-year-olds were asked to judge whether novel nouns referred to "name-similar" familiar objects or novel objects (e.g., whether japple referred to an apple or a binder clip). During an intervening treatment phase, they were asked to pick the referents of novel nouns from pairs of familiar objects (Experiments 1 and 3) or were taught subordinate names for familiar objects (Experiment 2). Most resisted the lure of phonological similarity in the pretest but increased selection of name-similar familiar objects over novel ones in the posttest. In Experiment 3, which involved monosyllables that differed in initial phoneme from the familiar words, treatment produced this effect only when accompanied by a rhyme-sensitization procedure. Experiment 2 included two other age groups: 2-year-olds, who were less resistant to phonological similarity in the pretest and responded to the treatment like the 4-year-olds; and adults, who nearly always selected the novel objects in the pretest and posttest. For children, the impact of treatment was positively associated with ability to detect phonological similarity and negatively associated with vocabulary size.

No abstract