Since the early 1970s, Loftus and her colleagues have been studying the role of post-event suggestion in the creation of false memories, a finding now referred to as the “misinformation effect” (Loftus, 1997, p. 70). In the first experiment of an early study (Loftus & Palmer, 1974), participants viewed brief scenes (5 to 30 s) of an automobile accident, after which they received a questionnaire, where the test question, varied between-subjects, was: “About how fast were the cars going when they…each other?” (p. 586). In the ellipses, the experimental words were “smashed,” “collided,” “bumped,” “hit,” or “contacted” (p. 586). The results indicated that the 5 words respectively were associated with faster to slower speeds (mean speed estimates declined from 40.5 s when “smashed” was used to 31.8 s when “contacted” was used).
To discriminate between the interpretations that participants’ uncertainty caused the words to bias their responses vs. that the words caused a change in the participants’ memories, in a second experiment, after viewing an accident scene, participants, as in the first experiment, received a questionnaire, but a third of the participants received the word “hit,” a third received the word “smashed,” and a third did not receive the question. Participants returned a week later to answer other questions about the accident scene, and this time the key question was: “Did you see any broken glass?” (p. 587). As in the first experiment, participants’ who received the word “smashed” estimated faster speeds than those who received the word “hit,” but the most important result was that 32% of the 50 participants who had received the word “smashed” falsely remembered glass, compared to 14% and 12% of the 50 participants who had received the word “hit” and the 50 who did not receive the speed-estimation question.
Thus the second interpretation that the words altered participants’ memory representations of the scene was supported.
In summary, the above research provided evidence that the independent variable, word in a key sentence, influenced both estimates of speed (dependent variables in both experiments) and reported memories of broken glass in tapes of accidents that depicted no broken glass (the dichotomous dependent variable in the second experiment). The research, along with similar studies reported below, has practical implications regarding both the accuracy of eyewitness testimony and how we should interpret less formal accusations people make against others. For example, whether one is serving as a juror or hearing from a friend an informal accusation against another person, we might want to require evidence other than memories to conclude an accusation is true.
The study raised no ethical concerns because participants were not misinformed and the questions were impersonal and not of an emotionally arousing nature. However, as discussed below, the research has been one of those in a category that has been criticized for undermining confidence in eyewitness testimony and also has led to other research using the same post-event suggestion technique to create false personal memories that also has been criticized as potentially harmful to participants. Along with other research, also discussed below, the study provided evidence extending the well-supported general theory of memory being a constructive process (Brown & Craik, 2000). Also, while the individual differences found in the reported research did not allow assessing the roles of nature and nurture, other research, discussed below, has suggested such differences are attributable to interactions of genetic predispositions with experiences.
Memories Created by Post-event Suggestion
Since the time of the experiment described above, the technique of providing participants with post-event suggestions has been used in many experiments, providing supporting evidence that post-event information becomes integrated into memories. Ten months after an El Al airline plane crashed into an apartment building in Amsterdam, Cromberg, Waneaar, and Van Koppen (1996, as cited in Schacter, 2001) asked Dutch university students questions about what they remembered, including the key question: “Did you see the television film of the moment the plane hit the apartment building?” (p. 112).
After 55% of the respondents said yes, the researchers conducted a follow-up study, where the percentage rose to 66%, and most of these participants also remembered one or more details they had seen – in a film that did not exist. Note that the question provided more than suggestion, since by use of the phrasing “the television film,” as opposed to “a television film,” a legitimate inference was that there actually was such a film, probably accounting for the higher percentage of memory alterations than in Loftus and Palmer (1974) and other studies. Similarly, after participants viewed a videotape of a robbery at a convenience store (Higham, 1998, as cited in Schacter, 2002), those given the false information that the attendant at the store wore a white apron were more likely than other participants to remember the apron two days after viewing the tape.
The Nature of Human Memory
Loftus and Palmer’s study (1974) and other studies using the same technique extended a theory of the nature of human memory suggested by Freud (1897, as cited in Bowers & Farvolden, 1996). When Freud began interpreting many of his patients’ recovery of “memories” of early “repressed” sexual experiences with their fathers as fantasies, he provided an example of what has become the major theory of the nature of human memory (Brown & Craik, 2000). This theory is that the human brain is structured (by nature) so that our memory representations are not nor are they intended to be replicas of experiences. The format of information in long-term memory generally is of meaning or gist, making memory distortions inevitable.
Bartlett (1932, as cited in Roedigger & McDermott, 2000) provided major evidence in the development of the theory that memory is a constructive process. After participants heard a story about events in a culture different from their own (e.g., The War of the Ghosts), their recall became more and more distorted over time by experiences typical of their own culture. Also, we remember meaning but usually cannot discriminate between different ways of expressing meaning.
In a classic study, Bransford, Barclay & Franks (1972, as cited in Hunt & Ellis, 2004) demonstrated that accuracy, not similarity, resulted in false memories of previously presented sentences. For example, when a sentence the participants heard was “Three turtles rested on a floating log and a fish swam underneath it,” they were likely to falsely remember the sentence “Three turtles rested on a floating log and a fish swam beneath them” (p. 338). Obviously, people are able to remember verbatim information, such as lines in a play, but, in general, memory is not a camera-like process (and earlier descriptions of “flashback” or camera-like memories of learning about a shocking historical event have been disconfirmed, Schooler & Eich, 2000).
Loftus and Palmer (1974) and other researchers cited above extended this theory by providing evidence that after a memory representation has been formed, false information can alter the representation. Loftus (1979/1996) frequently served as an expert witness, presenting findings on the influence of suggestion on memory, as well as experimental findings of false recognition memory of perpetrators of a crime and conditions under which such false memories are most likely to occur (e.g., brief exposure to the person, poor lighting conditions, etc.). Early critics of the ethics of Loftus and others providing evidence and testimony that had cast doubt on eyewitness testimony should note that a case in 2002 (Loftus) was the 100th conviction that had been overturned based on DNA evidence, a procedure that hadn’t been developed when the people were convicted.
Clearly, considerations other than finding the truth should not impose limits on research in either the hard or social sciences (indeed, expectancy effects, beyond the scope of this paper, have led to such methods as double-blind experiments where neither participants nor data collectors know whether a medication or a placebo had been used). Despite knowing about the damage suffered by people who had spent years in prison prior to being released on the basis of DNA evidence, scandals involving serious accusations that later turn out to be based on memories created by suggestion continue unabated.
Planting Memories of Sexual Abuse
The 1980s and 1990s might be characterized respectively as the decades of planting false memories of sexual abuse in preschool children and in women receiving therapy. After one mother reported that her preschool son had been abused by staff at the McMartin Preschool, accusations “spread like wildfire” of bizarre sexual abuse and rituals at preschools across the country (Schacter, 2002, p. ). It turned out that the first mother reporting abuse suffered from both schizophrenia and alcoholism and had a record of having made previous false accusations (Pokorney & Jackson, 1995).
After spending millions of dollars and after more than 5 years of trials against the McMartin family and others working at the school, there wasn’t a single conviction and jurors reported that tapes of the children being interviewed (individually and for over an hour) convinced them that the children were being forced by interviewers to “remember” the events being charged by the prosecution (Garven, Wood, & Malpass, 2000). When copies of the tapes became available, Garven, Wood, Malpass, & Shaw (1998) noted that interviewers were repeating outrageous questions (e.g., “Are you going to be stupid, or are you going to be smart and help us here?”, p. 349) until the children “remembered” events.
In experiments using, of course, much milder and briefer forms of suggestion, percentages up to more than a majority of preschoolers, depending on the experiment, who were given suggestions that false events occurred, not only agreed the events occurred but actually formed memories of the false event occurring. For example, a week after preschoolers had been misinformed that “the big kids” reported that a visitor to the class did “bad things,” children reported remembering the event that had been suggested the previous week (Garven, Wood, & Malpass, 2000).
In the 1990s, there were reports across the country that women, mainly those receiving “therapy,” were “recovering” memories of abuse, typically by their fathers, that they had never before remembered. Indeed, therapists, some with actual credentials, not only admitted they strongly suggested abuse to their clients, but also theorized that strong suggestion was necessary to help clients remember previously unconscious experiences of abuse and that “recovering” such memories was necessary if clients were to recover from whatever symptoms caused them to seek therapy (Bowers & Farvolden, 1996).
In many follow-up experiments, again using milder forms of suggestion, about 25% of participants would form memories of just about any suggested event, including impossible ones, such as memories of events occurring the first few days after birth (Loftus, 1997). While researchers used the findings to argue that it was unethical to use suggestion to elicit “recovered” memories of abuse, therapists argued the experiments undermined the pain of clients who had experienced abuse and also that it was unethical to plant false memories in participants. Researchers argued that it was false reports that undermined the pain of those who actually been sexually abused.
There are strong ethical arguments against the use of suggestive forms of therapy, many of those who had “recovered” their memories while in therapy themselves later came to believe that their memories were false ones attributable to therapists “planting” the memories (Loftus, 2002, p. 42), and there have been no reports of harm from participants in experiments where they learned in debriefing that their “memories” had been planted. Nonetheless, following the decade of accusations against parents, large numbers of young men began reporting “recovered” memories of sexual abuse by Catholic clergy (Loftus, 2002). Perhaps ironically, evidence described below has not only suggested that most people who had experienced different forms of seriously traumatic events were well-adjusted, but also that therapy encouraging people to remember and focus on trauma actually is harmful.
Individual Differences in Memory Distortions
Most of the findings on individual differences have been related to positive and negative memory distortions. There have been consistent findings that most people’s memories of themselves (e.g., Bahrick, Hall, & Berger, 1996) and others perceived as like themselves (e.g., White, Coppola, & Multunas, 2008) are distorted in a positive direction and negative information becomes inaccessible more quickly than positive information (e.g., Walker, Vogl, & Thompson, 1997). Indeed, while most adolescents are more likely than their parents to report feeling “awkward” or “lonely” and less likely to report feeling “happy” or “proud” (Larsen & Richards, 1994, as cited in White, Coppola, & Multunas, 2008), when adults have been asked to report their most positive memories (e.g., their happiest or proudest memories), they often report memories of adolescent experiences, but when asked to report their most negative memories, they are most likely to report recent events (e.g., Rubin & Berntsen, 2003).
The most interesting finding has been that the minorities of people who distort memories in the opposite direction have also been less happy and less well-adjusted than most people. For example, people with poor or even no memories of their concentration camp experiences during the Holocaust were as well-adjusted as a matched group who had never been in a concentration camp (Kaminer & Lavie, 1996, as cited in White, Coppola, & Multunas, 2008), when widows have been asked to report the levels of grief they had reported more than 4 years prior to a report they gave during their first year of bereavement, most distortions were underestimates of prior reports and those with these distortions also were better adjusted than those who overestimated prior grief (Safer, Bonanno, & Field, 2001), and in research using college students, poor memory for negative information was associated with shorter reaction times in generating happy than sad memories and with fewer diary entries of experiencing sad events and feelings (Boden & Baumaster, 1997, and Cutler, Larsen, & Bunce, 1996, both as cited in White, Coppola, & Multunas, 2008).
The relationships between memory and adjustment reported above might suggest that most people have genetic predispositions to focus attention on positive events and that poor adjustment may be a result of an interaction of an opposite genetic predisposition with experience. In other words, whether the latter genetic predisposition is manifested depends upon the number and magnitude of negative experiences a person encounters.
- Bahrick, H. P., Hall, L.K., & Berger, S. A. (1996). Accuracy and distortion on memory for high school grades. Psychological Science, 7, 265-271.
- Bowers, K.S., & Farvolden, P. (1996). Revisiting a century-old slip: From suggestion disavowed to the truth repressed. Psychological Bulletin, 119, 355-380.
- Brown, S.C., & & Craik, F.I.M. (2000). Encoding and retrieval of information. In E. Tulving and F.I.M. Craik (Eds.), The Oxford handbook of memory (pp. 93-107). New York: Oxford University Press.
- Garven, S., Wood, J.M., & Malpass, R.S. (2000). Allegations of wrongdoing: The effect of reinforcement on children’s mundane and fantastic claims. Journal of Applied Psychology, 85, 38-49.
- Garven, S., Wood, J.M., Malpass, R.S., & Shaw, J.S. (1998). More than suggestion: The effect of interviewing techniques in the McMartin preschool case. Journal of Applied Psychology, 83, 347-359.
- Hunt, R.R., & Ellis, H.C. (2004). Fundamentals of cognitive psychology. New York: McGraw-Hill.
- Loftus, E.E. (1979/1996). Eyewitness testimony. Cambridge, MA: Harvard University Press.
- Loftus, E.E. (1997). Creating false memories. Scientific American, 377, 20-26.
- Loftus, E.E. (2002). Memory faults and fixes. Issues in Science and Technology, 18, 41-50.
- Loftus, E.E., & Palmer, J.C. (1974). Reconstruction of automobile destruction: An example of the inteersection between language and memory. Journal of Verbal Learning and Verbal Behavior, 13, 585-589.