Misinformation and Memory: Study Notes

Misinformation and Memory

Introduction

• Eyewitness memory is malleable and subject to alteration over time.
• Misinformation effect: Errors in recalling event details due to exposure to false information.
• People genuinely believe the misinformation details they recall.
• Typical misinformation study involves witnessing a simulated event, followed by exposure to false information (suggestive questioning, erroneous details from another witness).
• Exposure to misinformation can lead to distortions in memory for genuinely experienced objects or people.
• People may recall seeing objects that did not appear in the original event.
• Researchers can persuade people to recall completely fictitious people or experiences.

Creating False Memories

• Recipe for creating false memories:
  • Convincing the individual that the false event is plausible.
  • Leading the individual to believe the false event was personally experienced.
  • Creating a false memory rich in detail through false feedback or manipulations of information.
• Memory distortion studies often focus on impersonal events or personal ones that are chronologically distant.
• It remains to be shown that suggestive techniques could affect memory for a recently experienced, personally relevant, highly stressful event.
• Laboratory simulations may not activate neurobiological systems in the same manner as realistic events.
• The study assesses whether human memory for recently experienced, personally relevant, high-stress events would be altered by exposure to suggestive misinformation.
• The study was conducted in U.S. military personnel enrolled in Survival School training.
• Survival School is an ideal setting for testing hypotheses about the real-world impact of misinformation for personally relevant, highly stressful events because of realistic stress and homogenous group of participants.
• Hypotheses:
  • Exposure to misinformation at an individual level would result in false memories related to a specific stressful event (mock interrogation).
  • Group exposure to misinformation would result in false memories about an event that participants experienced together as a group.

Methods

• Participants: 861 active duty military personnel (649 male, 192 female).
• Mean age: 26 (SD=5).
• Participants were randomly assigned to one of four experimental groups:
  • Control Group (N=158): No misinformation.
  • Misinfo-Questionnaire group (N=372): Post-event misinformation embedded in the memory questionnaire.
  • Misinfo-Photo group (N=85): Photographic misinformation during mock captivity.
  • Misinfo-Video group (N=246): Videotape concerning a specific event experienced as a group. Sub-groups (N=81, 90, 75) were exposed to different versions of the misinformation-videotape.

Targeted Events for Memory Assessment and Misinformation

• Survival training includes wilderness evasion, followed by mock-captivity in a prisoner of war camp (POWC).
• Stressors are modeled from the experiences of actual military personnel who have been prisoners of war.
• The experience is highly stressful.

Interrogation Stress

• Approximately 12 hours after being placed in the POWC, participants individually experience a highly stressful interrogation.
• Students are unable to cue other students as to interrogator identity.
• The student is required to face the instructor and maintain eye contact.
• The student must adopt a height that is less than that of the instructor.
• Failure to comply results in physical punishment.
• Interrogator asks questions and physically confronts the student if he or she does not appear to be answering the questions.
• Physical confrontation includes facial slaps, abdominal punches, walling, and stress positions.
• Interrogations are approximately 30 minutes in duration.
• The interrogation stress is intense and elicits alterations in psychological and neurobiological indices.
• Norepinephrine and cortisol levels are elevated.
• Gonadal hormone suppression is observed.
• Psychological symptoms of dissociation are also observed.
• Alterations in cognitive function are measured by standardized psychological testing (Rey Osterrieth Complex Figure [ROCF]).
• Following interrogation stress exposure, all participants were exposed to the stress of isolation, as well as sleep and food deprivation for approximately the next 36 hours.
• After several hours of isolation stress, all participants were exposed as a group to an event called the “Propaganda Speech.”

Propaganda Speech

• Participants listen to a “Commandant” of the camp give a lecture extolling the virtues of a non-US friendly political system.
• The Propaganda Speech lasts approximately 30 minutes.

Assessment of Memory

• Approximately 4 hours after release from the mock POW camp (36 hours after exposure to Interrogation stress and approximately 20 hours after exposure to the Propaganda Speech), participants in the No-Misinformation Control Group completed a non-leading questionnaire.
• Participants were told the focus was the interrogation and not to guess.
• The initial section of the questionnaire asked participants to select descriptive characteristics of the interrogator.
  • 11 descriptive categories, based on a consensus assessment from the Instructor Cadre at the Survival School.
  • Categories: Sex/Gender, Race, Height, Build, Hair Length, Hair Color, Face Shape, Facial Hair, Eye Color, Ear Appearance, and Teeth.
• The second section of the questionnaire had open-ended questions about emotionally neutral items and items that would likely carry emotional valence for military personnel.
• The third section was a target-absent photo-spread of nine former Survival School instructor/interrogators.
• Participants were instructed to indicate whether his or her interrogator was depicted in the photo-spread.

Misinformation Conditions

• Misinformation-Questionnaire Group received identical instructions followed by a questionnaire containing misinformation and leading questions.
• Misinformation-Photo Group was exposed to a photograph of a Caucasian male ('Foil') while in individual isolation cells, followed by the non-leading questionnaire and the photo-spread.
• Misinformation-Video Group was exposed to misinformation in the form of a video about the Propaganda Speech, followed by a non-leading questionnaire, then completed the self-report portion of the Clinician Administered Dissociative Symptom Scale (CADSS).
• Three eight-minute videotapes were created about the Propaganda Speech; two contained misinformation.
• In the misinformation videos, the survival school staff wore automatic weapons and/or carried Rocket Propelled Grenade devices (RPGs).

Data Analysis

• The 11 descriptive categories were derived from a standard list of adjectives used in military debriefing assessments of eyewitness statements.
• Responses to the open-ended and leading questionnaires were manually reviewed and scored by two separate teams to permit a ‘blind’ assessment.
• Responses of participants in the No-Misinformation Control Group were coded for whether subjects responded to the question (0=no response; 1= response) and whether their answer was correct (0=incorrect; 1=correct; 2= I don't know/I don't remember).
• Responses of subjects in the Misinformation-Questionnaire Group, the Misinformation-Photo Group and the Misinformation-Video Group were similarly coded to indicate whether subjects responded to the misinformation and endorsed recalling such information, or whether they denied experiencing the suggested stimuli, or did not remember the information (0=no endorsement; 1=endorsed misinformation; 2= I don't know/I don't remember).
• Chi-square analyses were used to examine whether the distributions of endorsements to questions were significantly greater in participants exposed to misinformation.
• Sign tests were used [for the Misinformation-Questionnaire Group] in order to assess whether there was a relationship between the endorsement of the different types of misinformation (neutral/non-neutral).
• General Linear Model Univariate Analyses of Variance were used to determine whether and to what degree endorsements of false information were significantly greater between the three comparison groups [Misinformation-Video Group].

Results

• 861 participants completed the initial section of the questionnaire (description of the interrogator).
• All 861 subjects correctly reported the Gender of their Interrogator.
• 836/861 (97%) of participants correctly described the Race of their interrogator, and 26 (3%) were mistaken.
• 621/861 (72%) correctly described their interrogator's Height, and 474/861 (55%) correctly described their interrogator's Build.
• Participants were incorrect in the majority of categories related to head/face appearance, except for Teeth and Hair Length.
• Compared to those not exposed to misinformation, participants exposed to misinformation were more likely to endorse false memories.

Misinformation Questionnaire Condition (Group Two)

• Exposure to misinformation increased false memory for both neutral and non-neutral items.
• Neutral items (Glasses: No misinformation, 4/158 (2.5%), Misinformation, 74/372 (20%), {\chi^2 = 2.6, df=1, p < 0.001}; Telephone: No misinformation, 16/158 (10%), Misinformation: 365/372 (98%), {\chi^2 = 4.2, df=1, p < 0.001}).
• Non-neutral items (Military Uniform: No misinformation, 35/158 (22%), Misinformation, 316/372 (85%), {\chi^2 = 1.9, df=1, p < 0.001}; Weapons: No misinformation, 5/158 (3%), Misinformation, 100/372 (27%), {\chi^2 = 3.9, df=1, p < 0.001}).

Misinformation-Photo Condition (Group Three)

• Exposure to misinformation in photograph form resulted in a significant increase in false positive eyewitness identifications.
• Eyewitness Identification False Positive endorsement rate: No misinformation, 84/158 (53%); Misinformation, 77/85 (91%); {\chi^2 = 4.5, df=1, p < 0.001}.
• Foil selection: No misinformation, 13/84 (15%); Misinformation, 65/77 (84%); {\chi^2 = 7.6, df=1, p < 0.001}.

Misinformation-Videotape Condition (Group Four)

• Impact of misinformation on the endorsement of weapons present during the Propaganda Speech.
• No misinformation, familiar staff in video: 4/75 (5%); Misinformation, unfamiliar staff in video: 10/90 (11%); Misinformation, familiar staff in video: 42/81 (51%); {\chi^2 = 5.9; df= 2; p < 0.001}.
• General Linear Model Univariate Analysis of Variance using endorsements as the dependent variable and Group as the independent variable (i.e., the ‘between subject’ factor) indicated that there was a significant between-subjects effect ({F (1,2)= 38; p<0.001}).
• Post-hoc comparisons indicated that this effect was due to higher endorsement rates in the group exposed to the misinformation video containing familiar staff members.

Dissociation

• No differences in stress-induced dissociation symptoms were observed between the four groups.
• No significant relationships were observed between dissociation scores and vulnerability to misinformation.

Discussion

• Data confirm that human memory for realistic, recently experienced stressful events is subject to substantial error and modification by exposure to misinformation.
• False memory endorsements about non-trivial items (i.e., weapons) were observed in at least 27% of participants, with higher rates for uniforms or human faces.
• Altering memory in military personnel trained to resist propaganda extends the applicability of false memory research and suggests these observations should be taken seriously by professionals working with victims of traumatic stress.
• Real-world victims of stressful events may be exposed to misinformation through various sources.
• Misinformation presented at a group level was effective at creating false memories.
• Misinformation paired with trusted/familiar individuals is more readily accepted.
• Eyewitness identifications had a surprisingly high false positive rate of approximately 50%.
• Participants were generally correct in describing characteristics observable at a distance (race, gender, height, build) but incorrect in more discriminating characteristics (facial hair, eye color, face shape).
• Errors in describing build/height tended towards describing interrogators as larger/taller.

Limitations

• Time constraints prevent debriefing to explore whether memory alterations represent altered ‘beliefs’ or ‘remembering’.
• Inability to assess the impact of various misinformation types in each participant.
• Food and sleep deprivation may have influenced memory accuracy/vulnerability to misinformation.
• False memory endorsement rates may not reflect the general population due to sample homogeneity.
• Future research should use open-ended, free-recall procedures.

Implications

• Physicians and psychotherapists should videotape evaluations and use non-leading, open-ended interviewing techniques.
• Law enforcement professionals should take great care in interviewing methods and control the level of exposure interviewees have to information that might alter their memory.
• Event memory for events that have “just occurred” are also vulnerable to misinformation.