It has been argued that event plausibility constrains what types of false memories and false beliefs can be created.  For instance, Pezdek, Finger and Hodge (1997) in a memory implantation study were able to create false memories for being lost in a shopping mall but not for receiving an (ahem) enema (see also Pezdek & Hodge, 1999).

However the authors argue that there's other evidence that false memories can be created for implausible events like satanic ritual abuse, alien abductions, and neonatal events.

So Mazzoni et. al have this three stage model that describes how false memories are created?

• The event must be perceived as plausible
• The person needs to develop a false belief
• The person must interpret their thoughts and fantasies as memories

This experiment had to do with witnessing a demonic possession (low plausibility) and almost choking (high plausibility).

Phase I.  First thing that happened is that during mass testing subjects rated the plausibility of various events to have happened to someone like them.  Importantly they rated the plausibility of witnessing a demonic possession and of almost choking.  To be selected they for Phase II they had to say that witnessing a demonic possession was implausible but that choking was plausible.

Next thing that happened is that subjects filled out a Life Events Inventory (LEI).  This asked them to rate how certain they were that various things happened to them or didn't happen to them.  Again one of the questions was how certain they were that they had or had not witnessed demonic possession before the age of three and how certain they were that they had or had not almost choked on something before age three.  To be selected for Phase II they had to indicate that they were pretty certain neither thing had happened to them.

Only subjects who said that witnessing a demonic possession was implausible and who said that they were pretty certain they hadn't witnessed such a possession incident were included in the rest of the study.

Phase II.  Phase II took place 3 months after Phase I.

In Phase II subjects read 12 realistic looking articles about four different topics.

For subjects in the possession condition three of the articles were about demonic possession and how common it is to witness possession and so forth.

For subjects in the almost choked condition three of the articles were about how common it is for people to almost choke and so forth.

Subjects in the control condition didn't participate in Phase II.

Phase III.  One week later, subjects in the possession condition and the almost choked condition took a fear questionnaire and they were told that their responses indicated a high likelihood that they had either (a) witnessed a possession before age 3 or (b) had almost choked before age 3.

Phase IV.  Subjects again rated the plausibility of the various events as well as completing a the LEI all over again.  So the question is, will the articles and personalized suggestion they received increase the plausibility of the suggested events and increase their certainty that the events had happened?

The plausibility manipulations (i.e. articles, personalized suggestion) increased the rated plausibility of the POSSESSION event but not the ALMOST CHOKED event.

The plausibility manipulations increased the LEI ratings for both the POSSESSION event and the ALMOST CHOKED event.

Post manipulation plausibility ratings did not significantly predict LEI ratings (odd).

Results

The plausibility manipulation (i.e. articles) increased plausibility ratings for both critical events (i.e. KIDNAPPING; POSSESSION).

Same thing happened for the LEI ratings, the plausibility manipulations increased subjects confidence that those events happened to them.

Oddly, they don't seem to report the correlation between plausibility and LEI this time?

So in this Experiment they repeated the demonic possession manipulation but used either SAME CULTURE articles in the plausibility manipulation or OTHER CULTURE articles in the plausibility manipulation.

Results

Plausibility ratings increased in both the SAME CULTURE and DIFFERENT CULTURE conditions.

Liklihood ratings increased in the SAME CULTURE but not the DIFFERENT CULTURE condition.

The relationship between these two phenomenon is not all that clear though.  For instance, post-manipulation plausibility ratings and LEI ratings do not seem to be strongly correlated.  Moreover, plausibility seemed to be influenced by both same culter and different culture manipulations while LEI scores only seemed to be influenced by the same culture ones.

Their argument is that this lack of relationship occurs because "only modest levels of plausibility are required for changes in autobiographical beliefs, p. 58."  It may be worth discussing what we think of this account.