Early studies: Pickles, eggs, strawberry ice cream

In the initial published studies of false memory consequences, our research group gave subjects false memories for one of two food-related childhood events, getting sick after eating either dill pickles or hard-boiled eggs (Bernstein, Laney, Morris, & Loftus, 2005b). We chose getting sick after eating a particular food because this is a nicely concrete event that could reasonably happen at almost any time in childhood, and because there is an established literature demonstrating the psychological consequences that can occur when people genuinely get sick after eating a particular type offood (e.g., Gustavson, Garcia, Hankins, & Rusiniak, 1974;Broberg & Bernstein, 1987).We chose pickles and hard-boiled eggs because they were foods that we thought most kids would have had some experience with, but not everyday experience.

These initial studies, and indeed most of the studies described in this chapter, use the false feedback procedure for implanting false memories.This procedure involves three key phases, though with a host of variations (see also Laney & Loftus, 2010). In the first phase, subjects come into the lab and are given a cover story and a set of questionnaires to complete. The questionnaires all revolve around a common theme that supports the cover story but also provides key premanipulation measures for the specific study. In the second phase, which normally occurs approximately a week after the first phase (though in some studies happens as little as ten minutes later), subjects are given the false feedback manipulation. This takes the form of a supposedly computer-generated feedback profile that gives subjects information about events that happened in their childhoods. The content of this profile is determined not by subjects’ answers to the phase 1 questionnaires (as they are told), but instead by random assignment. In particular, experimental (but not control) subjects are told that they had a specific experience as a child. In the studies described here, this is normally a specific experience with a particular food (e.g., you got sick eating a hard-boiled egg). Subjects are asked to read their profile and, in most studies, to answer some questions about it to ensure that they do actually read and understand it. In the third phase of the study, which normally immediately follows the second phase, subjects are given an additional set of questionnaires that assess whether the manipulation has produced changes in subjects’ confidence that the suggested event occurred, as well as measuring the false memory consequences of interest in the particular study.

How do we decide that someone has developed a false belief or memory? In most false feedback, false memory studies, subjects are said to have developed false memories (termed “believers”) if they meet three specific criteria: (1) their premanipulation confidence that the critical event happened is low (that is, they do not have arguably “true” memories), (2) that confidence increases after the manipulation, and (3) they report a specific “memory” or a less specific “belief” for the event at the end of the study (see Morris, Laney, Bernstein, & Loftus, 2006). Often, the key comparison for determining whether false memories have repercussions is between these “believers” and control subjects.

In the initial false memory consequence study (Bernstein et al., 2005b, Experiment 1), subjects first completed just one questionnaire (asking about their preferences for various foods) and then were told that their data would be analyzed by a special computer system. A few minutes later, subjects were given a feedback profile that the computer had supposedly generated. This profile contained three filler items for all subjects (that as young children they had disliked spinach and enjoyed fried foods and chocolate-covered almonds) and for experimental subjects also contained the suggestion that they had once gotten sick after eating either dill pickles or hard-boiled eggs (depending on condition). Subjects were then asked to briefly elaborate on their critical item (controls elaborated on a filler item) and then to complete a further set of questionnaires designed to assess whether their confidence that they had gotten sick on pickles or eggs had increased and whether these altered beliefs might have consequences (detailed later).

The methodology for Experiment 2 was similar, except that (a) subjects completed the confidence measure (called the Food History Inventory) and three filler questionnaires at premanipulation, (b) the delay between phase 1 and phase 2 was a week instead of a few minutes, (c) the food preferences questionnaire was used as a postmanipulation confidence measure instead, and (d) we used an additional postmanipulation questionnaire called the Memory or Belief form, on which they were instructed to judge their experiences of three different events (including their critical egg or pickle event) as specific memories, less specific beliefs that the event occurred, or neither of these.

In Experiment 1, subjects given the pickle feedback were more confident than egg feedback subjects or controls that they had indeed gotten sick eating pickles, though egg feedback subjects were no more confident than pickle feedback subjects that they had gotten sick after eating hard-boiled eggs. In Experiment 2, both types of feedback produced significant increases in confidence in their respective subjects from pre- to postmanipulation. In addition, the combination of confidence change and a “memory” or “belief” response on the Memory or Belief form became the criteria for labeling subjects as having false memories (being “believers”) for their critical sickness event. Specifically, 25% of pickle feedback subjects and 31% of egg feedback subjects believed their feedback, reporting increased confidence that they had been sick after eating the food and reporting a specific memory or belief that they had been sick.

To determine whether subjects’ new false memories were consequential, they were given a Party Behavior Questionnaire, on which they were asked how likely they were to consume a variety of foods (including dill pickle spears and salted hard-boiled eggs, as well as related items—pickle slices and egg salad finger sandwiches) in a backyard party situation. In Experiment 2 subjects were also asked to rate their preference for a list of foods, including both critical foods and several closely related foods. Experiment 1 did not produce significant differences in expressed desire to eat pickles or hard-boiled eggs at a party between people who were or were not exposed to false feedback about getting sick, but this may be because those who actually believed the feedback were lumped in with others who did not. That is, this study was not able to distinguish between those who believed versus did not believe the false feedback. This comparison was possible in Experiment 2, and we found that those who believed their feedback were indeed less interested in eating the food that they now believed they had gotten sick after eating. And these effects even carried over to other, closely related, foods like egg salad.

That is, false memories did seem to be consequential for those who developed them, in the same way that true memories can be consequential. So looking for evidence of consequentiality appears to be another dead-end in the search for categorical differences between true and false memories. Nonetheless, there is still much more to learn about the consequences of false memories and what they can tell us about memory processes more broadly. For details regarding the food items, manipulations, proportions of “believing” subjects, and basic results regarding two key kinds of false memory repercussions (preference consequences and action consequences) for each of the studies described in this chapter, see Table 9.1.

Once we had established that false memories, like true memories, could be consequential for those who possessed them, we looked at several other types of false memories and consequences. First we gave subjects false memories for getting sick on chocolate cake and potato chips, using essentially the same methodology as that of Bernstein et al. (2005b, Experiment 2) described earlier (Laney, Morris, Bernstein, & Loftus, 2004). Although we were able to plant false memories for getting sick on both of these foods, we did not see any consequences of these false memories. That is, although 24% of subjects falsely believed that they had once gotten sick after eating chocolate cake, these cake believers did not demonstrate any reduction in preference for cake or willingness to eat cake, relative to nonbelievers or nonexposed subjects. We also successfully convinced 49% of subjects they had once gotten sick after eating potato chips, but again, these believers were no less interested in eating potato chips than nonbelievers or nonexposed subjects. (Note that this doesn’t undermine the previous claim that false memories cannot be distinguished from true memories on the basis of their consequentiality—in this study, true memories were also not distinguishable from other groups on the basis of their consequentiality.)

(continued)

So why do false beliefs about some foods have consequences but false beliefs about other foods do not? After some thought, we hypothesized that most college students have lots of positive experiences with cake and chips. They like to eat cake and chips and know full well that they like to eat cake and chips. When faced with (new) knowledge that they had once become sick after eating one of these common and preferred foods, subjects in the study may have concluded that, even if they once got sick, it clearly hadn’t affected their preference for these foods or willingness to eat them. After all, they had plenty of evidence from their own memories that they liked to eat these foods—much more than they likely had evidence about their feelings for rarer foods like dill pickles and hard-boiled eggs.This suggests that false memories may not be the next great diet fad (though see Bernstein, Pernat, & Loftus, 2011).

We next turned this bit of post-hoc reasoning into a hypothesis for a new study (Bernstein, Laney, Morris, & Loftus, 2005a). Specifically, we hypothesized that we would be able to give subjects false memories for getting sick after eating either a very common food, chocolate chip cookies, or a somewhat less common food, strawberry ice cream, but that we would only find false memory consequences for the less common food.

The procedures for the study were similar to those used by Bernstein et al. (2005b, Experiment 2). Subjects completed questionnaires and were told that their data would be analyzed by a special computer that would produce profiles before they returned a week later. All subjects’ profiles told them that as young children they had liked bananas and hated spinach and that they had been happy when a classmate had brought sweets to school. Subjects in the experimental conditions were also told that they had once gotten sick after eating either strawberry ice cream or chocolate chip cookies. In fact we were really successful at giving people false memories only for getting sick from eating strawberry ice cream (18% of these subjects met our “believer” criteria).These people also demonstrated avoidance of strawberry ice cream, as hypothesized. No one avoided the more common food, chocolate chip cookies. Experiment 2 further demonstrated that the recency of eating the food did not, by itself, predict whether people would adopt false memories.