HC

Loftus and Palmer (1974): Car Crash Experiment Notes

Loftus and Palmer (1974): Car Crash Experiment

Page 1: Introduction, Findings, and Implications

This section introduces the foundational study by Loftus and Palmer (1974), published in Simply Psychology by Saul McLeod, PhD, on September 9, 2025. The core experiment involved showing participants videos of car crashes and asking them to estimate speeds using various verbs such as "hit" or "smashed."

Key Finding

When stronger verbs like "smashed" were used instead of a more neutral verb like "hit," participants not only reported significantly higher speed estimates but also remembered seeing non-existent broken glass.

Implication

These findings revealed that memory is reconstructive, meaning it is not a perfect recording but can be significantly shaped and altered by subsequent information, particularly the wording of questions.

Application

The research highlighted the inherent unreliability of eyewitness testimony. It subsequently influenced how police and courts handle questioning witnesses, leading to more cautious approaches.

Criticism

Critics argue that the laboratory setting and the use of video clips, as opposed to real-life events, lack ecological validity. This raises questions about the generalizability of the results to actual accident situations.

Elizabeth Loftus's Focus

Psychologist Elizabeth Loftus has extensively researched how subsequent information can affect an eyewitness's account. Her primary focus has been on the influence of (mis)leading information, encompassing both visual imagery and the specific wording of questions in the context of eyewitness testimony.

Definition of a Leading Question

A leading question is defined as a question that explicitly or implicitly suggests what answer is desired or guides the respondent toward a particular answer.

Page 2: Memory Flexibility and Experiment One Aim

Memory Flexibility

Loftus's research indicates that memory for a witnessed event is remarkably flexible. If an individual is exposed to new information between witnessing an event and recalling it, this new information can profoundly affect what they remember. The original memory is susceptible to modification, change, or supplementation.

Introduction to Experiment One

The unreliability of eyewitness testimony and its susceptibility to leading questions is vividly illustrated by Loftus and Palmer's (1974) classic study titled "Reconstruction of Automobile Destruction."

Aim of Experiment One

The primary aim of Experiment One was to test the hypothesis that the specific language used in eyewitness testimony can alter memory. Specifically, they sought to demonstrate that leading questions could distort eyewitness testimony accounts, thereby having a confabulating effect where the recollection becomes distorted by cues embedded in the question.

Methodology for Speed Estimation

To achieve this, Loftus and Palmer (1974) asked participants to estimate the speed of motor vehicles using different forms of questions. The researchers recognized that people are generally poor at estimating vehicle speed, which might make them more open and susceptible to suggestions embedded in the questions.

Page 3: Experiment One Procedure

Critical Question

The key question posed to participants was: "About how fast were the cars going when they (smashed / collided / bumped / hit / contacted) each other?" The verb in parentheses was varied for different groups of participants.

Participants

  • Sample: Forty-five American students from the University of Washington.
  • Sampling Technique: Opportunity sample.

Experimental Design

  • Type: Laboratory experiment.
  • Conditions: Five distinct conditions, with each participant experiencing only one (an independent measures experimental design).

Stimuli and Session Details

  • Videos: Participants viewed seven short films depicting traffic accidents. These clips were sourced from driver's education and police training videos.
  • Clip Duration: The films varied in length from 5 to 30 seconds.
  • Staged Crashes: Some crashes were staged and filmed at actual speeds of 20, 30, or 40 mph.
  • Order Effects Control: Each group watched the films in a different sequence to mitigate potential order effects.
  • Session Length: The entire experimental session lasted approximately 90 minutes.

Participant Tasks and Variables

  • Initial Task: After each film, participants first provided a free-recall account of what they had witnessed.
  • Critical Question: They then answered the critical question regarding vehicle speed.
  • Independent Variable (IV): The specific verb used in the critical question (smashed, collided, bumped, hit, contacted).
  • Dependent Variable (DV): The estimated speed reported by the participants in miles per hour (mph).

Page 4: Experiment One Findings and Conclusion

Findings

  • Influence of Verb: The estimated speed consistently varied and was strongly influenced by the particular verb used in the question.
  • Verbal Implication: Each verb implicitly conveyed information about the speed, which systematically affected the participants' memory of the accident.
  • Average Speed Estimates for Each Verb:
    • "Smashed": The highest average estimate at about 40.8 mph.
    • "Collided": Followed with an average of 39.3 mph.
    • "Bumped": Averaged 38.1 mph.
    • "Hit": Averaged 34.0 mph.
    • "Contacted": The lowest average estimate at 31.8 mph.
  • Overestimation: Given that the filmed crashes were staged at actual speeds of between 20 and 40 mph, the participants generally overestimated the speeds, especially with stronger verbs.
  • Statistical Significance: Statistical analysis confirmed that the effect of the wording on speed estimates was significant.

Conclusion

Loftus and Palmer concluded that the verb used in the question conveyed an impression about the speed of the car, which in turn altered the participants' perceptions. This strongly indicates that eyewitness testimony can be biased simply by the way questions are worded.

Proposed Explanations (leading to Experiment Two)

Loftus and Palmer suggested two potential explanations for these findings:

  1. Response Bias: The different speed estimates could be due to a response bias, where the critical word influences the participant's answer without truly altering their memory of the event.
  2. Memory Alteration: Alternatively, the verb might genuinely alter the participant's memory of the event, making it seem more severe and integrating this altered perception into their stored memory. This specific possibility led to the design of Experiment Two.

Page 5: Experiment Two: The Broken Glass Manipulation

Loftus and Palmer's Two Explanations Revisited

Loftus and Palmer proposed two primary explanations for the results of Experiment One, which formed the basis for their second experiment:

  1. Response-bias factors: This explanation suggests that the misleading information (e.g., the critical verb) primarily influenced the answer a person gave. The information biased their response but did not necessarily create a false memory of the event itself. For instance, the differing speed estimates occur because words like "smashed" or "hit" merely influence or bias an individual's immediate reply.
  2. The memory representation is altered: This more profound explanation posits that the critical verb literally changes a person's perception of the accident. Stronger verbs might lead someone to perceive the accident as more serious than it actually was. This altered perception is then stored and becomes part of the individual's memory of the event, thereby changing the original memory.

Aim of Experiment Two

To definitively test which of these explanations was more accurate, Loftus and Palmer designed a second experiment. Its specific aim was to determine whether leading questions simply bias participants' immediate responses or if they actually alter long-term memory representation of the event.

Procedure of Experiment Two

  • Participants: 150 students were recruited for this experiment.
  • Stimulus: All participants watched a one-minute film. The film depicted a car driving through the countryside, followed by a brief (4-second) multiple traffic accident.
  • Initial Questioning (Independent Variable - IV): The students were questioned immediately after watching the film, divided into three groups:
    • 50 participants were asked: "How fast were the cars going when they hit each other?"
    • 50 participants were asked: "How fast were the cars going when they smashed each other?"
    • The remaining 50 participants constituted a control group and were not asked any question about the car's speed.
  • Delayed Measurement (Dependent Variable - DV): One week later, without re-watching the film, all participants answered ten questions about the accident. One of these questions was a critical, direct question, randomly placed within the list: "Did you see any broken glass? Yes or no?"
  • Crucial Detail: There was no broken glass visible in the original film whatsoever.

Page 6: Experiment Two Findings, Conclusion, and Strengths

Findings of Experiment Two

The key finding was that participants who were initially asked "how fast were the cars going when they smashed" were significantly more likely to report seeing broken glass in the follow-up questions, compared to those asked "hit" or the control group.

Conclusion of Experiment Two

These results strongly suggest that leading questions do more than just bias people's immediate answers. They can actually change how an event is remembered over time. Loftus and Palmer articulated this by arguing that memory is reconstructive: it is formed by combining original details from the event with new information introduced afterward. Over time, these disparate sources merge and become indistinguishable, resulting in a single, unified memory that can unfortunately include false details.

Confabulation

The specific process where false details are added to a memory of an event is known as confabulation. This finding has profound implications, especially for the types of questions utilized in police interviews with eyewitnesses.

Strengths of Loftus and Palmer's Experiment

  • High Level of Control: One of the most significant strengths of the study was the exceptional degree of control over confounding variables. Being a laboratory experiment, researchers could meticulously manage a range of factors, including the age of participants, the specific incident viewed, and the experimental environment.
  • Causal Inference: This stringent control ensured that these extraneous factors did not influence participants' responses. Consequently, the researchers could confidently isolate the verb condition as the sole cause for participants reevaluating and altering their memories, allowing for strong causal inferences.

Page 7: Practical Implications, Replicability, and Weaknesses

Practical Implications

  • Reconstructive Memory Hypothesis: The concept of reconstructive memory is highly valuable as a psychological explanation, particularly for its utility in developing guidelines for police questioning of witnesses and suspects.
  • Interviewing Protocols: The conclusion that leading questions can distort memory has critical implications for interviewing witnesses. This applies to police interviews conducted immediately after an event and to lawyers' questioning in court potentially much later.
  • Interview Best Practices: Interviewers, regardless of their role, are strongly advised to avoid leading questions and to carefully phrase their inquiries in a way that does not suggest a specific answer to the individual being interviewed.
  • Devlin Report (1976): This research (among other evidence from Loftus) directly influenced the Devlin Report in 1976. This report famously recommended that trial judges should instruct juries that it is generally unsafe to convict a defendant based solely on the testimony of a single eyewitness, highlighting the inherent fallibility of such evidence.

Replicability

  • Standardized Procedure: A key strength of the study is its high replicability. As a laboratory experiment following a standardized procedure, the methodology is clear and easy for other researchers to copy and repeat, bolstering the reliability of the findings.

Weaknesses

  • Low Ecological Validity / Mundane Realism: A significant limitation of the research is its lack of mundane realism and ecological validity. Participants viewed pre-recorded video clips of accidents rather than experiencing a real-life accident firsthand.
  • Reasons for Low Ecological Validity:
    • Lack of Emotional Impact: Viewing a video clip typically does not elicit the same intense emotional response or stress as witnessing a real, live accident. This reduced emotional engagement might mean participants were less likely to pay close attention and less motivated to be perfectly accurate in their judgments.
    • Lack of Context: Watching a real crash provides a much richer and more immediate context. In the experiment, participants were cued to watch the video, whereas real-life accidents are largely unexpected and sudden.
    • Experimenter Expectations: In an experimental setting, participants may anticipate being asked specific questions about what they are watching. This expectation could alter how they attend to and mentally process the film, potentially differing from how they would experience an unexpected real event.

Page 8: Further Weaknesses and Conflicting Research

Consequences and Ecological Validity (Continued)

  • Real-life Consequences: In real-life situations, the answers and testimony provided by witnesses can have serious legal consequences, which often places significant psychological pressure on the witness. This pressure is largely absent in a lab experiment.
  • Overall Conclusion on Ecological Validity: Due to these factors, it can be broadly concluded that Loftus and Palmer's laboratory experiment possessed low ecological validity. Therefore, its findings may provide limited insight into how people's memories are affected by leading questions in authentic, real-life scenarios.

Conflicting Research

  • Yuille and Cutshall (1986): A study conducted by Yuille and Cutshall (1986) directly conflicts with the findings of Loftus and Palmer. They investigated the memory of individuals who had witnessed a real armed robbery and found that misleading information did not significantly alter their memory accounts.
  • Implication of Conflicting Research: This conflicting evidence implies that misleading information might exert a greater influence in a controlled laboratory setting than in real-life, high-impact situations, further supporting the critique regarding the ecological validity of Loftus and Palmer's study.

Biased Sample

  • Use of Students: A further methodological problem with the study was the exclusive use of students as participants.
  • Lack of Representativeness: Students are generally not representative of the broader general population in several important aspects. Differences in age, memory abilities, learning habits, driving experience, and susceptibility to demand characteristics could make it difficult to generalize the findings from this specific sample.
  • Driving Experience: Importantly, students tend to be less experienced drivers and, consequently, might be less confident in their ability to accurately estimate vehicle speeds. This potential lack of confidence could have made them more susceptible to being swayed by the specific verb used in the critical question.

Demand Characteristics

  • Awareness of Experiment: Participants are inherently aware that they are part of a laboratory experiment, which can subtly or overtly influence their behavior.
  • Seeking Clues: They may actively look for clues regarding how they are expected to behave (known as demand characteristics).
  • **