Warning against warnings: alerted subjects may perform worse : misinformation, involvement and warning as determinants of witness testimony

The article presents experiments exploring the memory misinformation effect. Subjects heard a recording and afterwards read a description of it, which included, in the misled group, some details inconsistent with the recording; finally thay answered questions about the recording. The aim of the research was to replicate the tainted truth effect, consisting in poor memory functioning of non-misled warned subjects and to check whether a subject's involvement in the issue moderates this effect. Highly involved subjects were more resistant to the misinformation effect than those lowly involved. In the case of highly involved participants, warning was effective in reducing the misinformation effect, but it also caused more errors in the case of non-misled subjects. Thus, warning witnesses about nonexising discrepancies between what they saw/heard and what they were told, might lead to less accurate testimony.