Children's and adults' spontaneous false memories: long-term persistence and mere-testing effects.

In studies of children's false memories of word lists, it has been found that false alarms are stable over long-term retention intervals (persistence effect), that the stability of false alarms can equal or exceed that of hits, that earlier memory tests increase the frequency of hits on later tests (true-memory inoculation effect), that earlier memory tests increase the frequency of false alarms on later tests (false-memory creation effect), and that test-induced increases in false alarms can equal or exceed increases for hits. We studied these phenomena in 6-, 8-, and 11-year-olds and in adults using short narratives about everyday objects and events. All of the phenomena were detected at all ages, but levels of spontaneous memory falsification were much higher than for word lists and patterns of developmental change were somewhat different. Important new findings were that the persistence effect and the false-memory creation effect were greatest for statements that would be regarded as factually incorrect reports of events in sworn testimony and that, like suggestive questioning, interviews that involve nonsuggestive recognition questions may nevertheless taint children's memories.