Ackil, J.K. & Zaragoza, M.S. (1998). Memorial consequences of forced confabulation: Age differences in susceptibility to false memories. Developmental Psychology, 34, 1358-1372.
 

Study concerns cases in which children are pressured to confabulate answers and whether they will later remember this confabulated information as if it were real.  They distinguish this from cases where children are merely "passively exposed" to false information.

Authors give example of the Wenatchee case as a case where children were forced to confabulate answers.

Roediger, Challis & Wheeler (1993) found that adults who were forced to make up answers on a memory test later had difficulty determining what they had actually experienced versus what they had confabulated.

The authors argue that confabulations may be less influential for complex events than for the pictures used by Roediger et. al

• Participants may not be as willing to fabricate details for entire events
• It may be easier for subjects to distinguish between fabricated and real events based on their content
• Roediger et. al's results might be a function of the large number of items subjects had to keep track of.

Main purposes of the study were to

• See if confabulated details later became confused with real details
• Check for age related trends

Subjects

First graders, fourth graders and college students participated

Procedures

Subjects watched a videotape of two brothers at summer camp

Subjects were asked a set of questions

• True event questions: Asking for details about a true event
• False event questions: Asking for details about an event that did not happen

Subjects in the forced confabulation condition were required to answer all questions, even if it involved guessing.  in the free condition subjects were not forced to respond

One week later subjects were asked source monitoring questions about what had been in the film by a new investigator.  Subjects were warned that they had been asked about some non-events.
The test included four types of items

• Cofabulated Items: Items that the subjects had confabulated (made up) in response to the false event questions (thus different items for every subject)
• True items the subject had mentioned in response to the true event questions
• True items the subject had not been asked about
• Control Items: Never seen never mentioned items (subjects were yoked such that these items were items that had been confabulated by the subject they were yoked with)

For subjects in the free condition, the false items were experimenter generated...
 

• Participants were more likely to state that confabulated items were in the tape than were control items
• Younger participants were more likely to claim confabulated items were in the tape than were older participants (but they were also more likely to say control items were in the tape). A&Z claim that the lack of a significant difference on control items between 4th graders and college students means that children are more prone than adults to confuse their confabulations with reality.
• Overall subjects were pretty good at remembering that they talked about the confabulated items, but this tendency clearly increased with age, especially if one corrects for response bias.
• Forced Condition Participants made more misattributions errors than free condition participants.

To a certain extent the authors find these results surprising.  Subjects know that some of the questions could have contained false information.  In the forced confabulation case they might be expected to remember their initial response of denying that the item existed.  This should improve source memory.

However, subjects may be especially convinced by confabulated events because they have integrated it into their own systems of knowledge and beliefs (Jim's note: Some have made similar arguments for how belief perseverance occurs).