Loftus & Palmer (1974) Effect of Leading Questions (Edexcel A-Level Psychology): Revision Notes
Loftus & Palmer (1974) Effect of Leading Questions
Background
Elizabeth Loftus and John Palmer conducted groundbreaking research into how memory can be distorted by post-event information. Their work built upon earlier studies showing that recollection of events can be altered by information received after the event occurred. This research focuses specifically on eyewitness testimony and the role that language, particularly leading questions, plays in memory accuracy.
The study examined memory formation in general contexts rather than under specific high-stress situations, making it highly applicable to criminal investigations and court proceedings. The research has had lasting influence on both cognitive psychology theory and practical law enforcement procedures over the past five decades.
This research is considered one of the most influential studies in cognitive psychology, particularly in understanding how memory is not simply a recording of events but an active reconstructive process that can be altered by external factors.
Theorist, title, year
Elizabeth Loftus and John Palmer - Reconstruction of automobile destruction: An example of the interaction between language and memory (1974)
Participants
Experiment One: 45 students were recruited for the initial investigation.
Experiment Two: 150 students participated in the follow-up study.
Both samples consisted of students, suggesting they were likely young adults with limited driving experience compared to the general population. This demographic choice would later become a consideration when evaluating the study's generalisability.
Aim
The research investigated whether the wording of questions could affect witnesses' memory of events. Specifically, the experimenters sought to determine if leading questions (questions phrased in a way that suggests a particular answer) would influence participants' estimates of vehicle speed when recalling a traffic accident.
Procedure
Experiment one
Participants viewed seven short film clips depicting traffic accidents. Each clip lasted between 5 and 30 seconds. After watching each film, participants completed a questionnaire asking them to provide an account of the accident they had witnessed and answer specific questions about it.
The critical manipulation involved a single question about vehicle speed. All participants received identical questionnaires except for one word in this critical question. Participants were divided into five groups of nine people each, with each group receiving a different verb in the question:
- "About how fast were the cars going when they smashed into each other?"
- "About how fast were the cars going when they collided into each other?"
- "About how fast were the cars going when they bumped into each other?"
- "About how fast were the cars going when they hit each other?"
- "About how fast were the cars going when they contacted each other?"
The critical question was randomly positioned among other questions in the questionnaire to reduce the likelihood that participants would identify it as the focus of the study, thereby minimising demand characteristics.
Importantly, the researchers knew the actual speeds of the vehicles in most film clips, allowing them to assess whether the verb used influenced estimates independently of the actual speed.
Experiment two
A total of 150 participants watched a film lasting less than one minute that showed a multiple-vehicle collision. The actual accident within the film lasted four seconds. Following the film, participants completed a questionnaire describing the accident and answering related questions.
The participants were divided into three groups of 50:
- Group 1: Asked "About how fast were the cars going when they smashed into each other?"
- Group 2: Asked "About how fast were the cars going when they hit each other?"
- Group 3: Control group - not asked about vehicle speed at all
One week later, all 150 participants returned to answer ten additional questions about the accident without rewatching the film clip. The critical question asked: "Did you see any broken glass?" Participants responded with 'Yes' or 'No'.
Notably, there was no broken glass visible in the original film clip.
Findings
Experiment one results
The verb used in the critical question had a measurable effect on speed estimates:
| Verb used | Mean speed estimate (mph) |
|---|---|
| Smashed | 40.5 |
| Collided | 39.3 |
| Bumped | 38.1 |
| Hit | 34.0 |
| Contacted | 31.8 |
Participants who were asked about cars that "smashed" gave the highest speed estimate (40.5 mph), whilst those asked about cars that "contacted" gave the lowest estimate (31.8 mph). This represents a difference of almost 9 mph between the two conditions, demonstrating that word choice alone influenced participants' numerical estimates.
Experiment two results
The responses to the broken glass question one week after viewing the film revealed:
| Answer to critical question | Verb condition 'Smashed' | Verb condition 'Hit' | Control (no verb) |
|---|---|---|---|
| Yes (saw broken glass) | 16 | 7 | 6 |
| No (no broken glass) | 34 | 43 | 44 |
Although most participants correctly identified that no broken glass was present in the film clip, a chi-square test confirmed that the results were statistically significant. Specifically, 32% of participants in the "smashed" condition reported seeing broken glass, compared to only 14% in the "hit" condition and 12% in the control group.
This finding is particularly important because participants did not actually see any broken glass—the verb used in the initial question appeared to have influenced their memory of the event even a week later. This suggests genuine memory alteration rather than simply affecting their immediate response.
Conclusion
The researchers concluded that the wording of questions can affect witnesses' responses in two distinct ways:
First interpretation: Response Bias
Participants may have been uncertain about the exact speed and used the verb as a contextual cue to inform their estimate. The verb created a response bias, influencing their decision-making process without necessarily altering their underlying memory of the event.
Second interpretation: Memory Alteration
The wording of the question may have actually changed participants' memory of the accident, causing them to recall it as more severe than it actually was. This represents genuine memory alteration rather than simply influencing their response strategy.
The findings from experiment two provide stronger support for the second interpretation. The fact that participants reported seeing broken glass (which was not present) a week after viewing the film suggests that the leading question had integrated into their memory of the event.
Reconstructive hypothesis
Loftus and Palmer proposed that memory operates according to a reconstructive hypothesis. This theory states that when we recall an event, we combine two sources of information: the original information obtained from witnessing the event, and additional information supplied after the event. Over time, these two sources of information merge together, making it difficult to distinguish which details came from the original experience and which came from later sources. This process creates an overall reconstructed memory of the event.
The study demonstrates that leading questions constitute a form of post-event information that can influence eyewitness testimony, with clear implications for legal proceedings. This challenges the traditional view of memory as a passive recording device and instead supports the idea that memory is an active reconstructive process.
Evaluation: Strengths
High level of control
Both experiments were conducted in laboratory settings with standardised procedures. All participants viewed identical film clips and received the same questionnaire (except for the manipulated verb). This high degree of control allows the study to be replicated easily to verify the reliability of the findings. The consistent methodology reduces the influence of confounding variables and increases confidence that the verb manipulation directly caused the observed differences in speed estimates.
The experimental design used was highly rigorous, with careful attention paid to controlling extraneous variables. This methodological strength means that researchers can be confident about the cause-and-effect relationship between the independent variable (verb used) and the dependent variables (speed estimates and false memory reports).
Embedded critical question
The critical question about speed was randomly positioned among other questions in the questionnaire. This methodological decision prevented participants from guessing the study's true aim and reduced demand characteristics. By disguising the research focus, the experimenters minimised the likelihood that participants' responses were due to the research methodology rather than representing genuine responses to the question wording.
Knowledge of actual speeds
The researchers knew the actual speeds of vehicles in most film clips. This allowed them to assess whether estimates were influenced by the verb independently of the true speed. They determined with certainty that the verb used, rather than the actual speed of the cars, had the primary effect on participants' estimates. This strengthens the validity of their conclusion about the effect of leading questions.
Quantitative data
The use of numerical speed estimates in experiment one and yes/no responses in experiment two produced quantitative data that could be objectively analysed. This eliminated subjective interpretation of results and reduced researcher bias. The data could be subjected to statistical testing (such as the chi-square test used in experiment two) to determine whether differences between conditions were statistically significant, making the findings more reliable and objective.
Evaluation: Weaknesses
Low ecological validity
The emotional context of watching film clips in a laboratory differs substantially from witnessing a real accident. Participants viewing staged accidents in comfortable classroom settings are unlikely to experience the same emotional strain, stress, or shock that genuine eyewitnesses face. Real witnesses might process and retain information differently under emotional duress, meaning the findings may not generalise to actual eyewitness situations. The validity of applying these results to real-world legal contexts is therefore questionable.
Ecological validity refers to the extent to which research findings can be generalised to real-world settings. The artificial nature of laboratory experiments often means that behaviours observed in the lab may not accurately reflect how people behave in natural environments.
Artificial task environment
Beyond the emotional differences, the entire situation lacks realism. Participants knew they were watching films as part of a research study, which may have affected their attention levels and memory encoding processes. In genuine accident situations, witnesses have different motivations, expectations, and attention patterns. The artificiality of the laboratory setting limits the extent to which we can generalise these findings to actual courtroom testimony scenarios.
Limited sample characteristics
Both experiments used student participants who were likely young adults with limited driving experience. This restricted sample may have made participants more susceptible to the influence of the verb because they lacked sufficient real-world experience with vehicle speeds to make confident independent judgements. A sample with more diverse driving experience might show different results. The findings may therefore only represent the views of this particular demographic rather than being applicable to the wider population, including experienced drivers or older witnesses who might feature as eyewitnesses in real criminal cases.
The use of a non-representative sample is a significant limitation. Students are not only likely to have limited driving experience but may also differ from the general population in terms of age, education level, and cognitive abilities. This limits the generalisability of the findings.
Reductionist approach
The study focused exclusively on one variable (question wording) that might influence memory of an event. However, numerous other factors affect eyewitness memory, including the emotional state of the witness, environmental conditions during the event (lighting, weather, visibility), individual differences in memory capacity, the time elapsed since the event, and the presence of weapons or violence. By isolating only question wording, the study provides a somewhat simplified view of the complex factors that influence human memory. This reductionist approach may not fully capture how memory operates in real-world situations where multiple variables interact simultaneously.
Real-world application
The findings from this research have had practical implications for the criminal justice system. The Devlin Committee investigated the use of eyewitness testimony in court proceedings after discovering that many individuals had been convicted of serious crimes based on eyewitness testimony alone. Their investigation revealed that such testimony could be unreliable.
Practical Impact on Legal Procedures
As a consequence of research like Loftus and Palmer's work, police and legal professionals now receive guidance on minimising the use of leading questions during interviews and investigations. This represents a substantial change in policing techniques compared to practices common in the 20th century, with procedures designed to ensure more accurate accounts are obtained from witnesses.
These changes remain in effect today, demonstrating the lasting practical value of this research in improving the accuracy and reliability of witness testimony in criminal proceedings.
Remember!
Key Points to Remember:
-
Verb manipulation: Changing a single verb in a question (smashed vs contacted) produced a 9 mph difference in speed estimates, demonstrating the power of language in influencing memory.
-
Two experiments, two effects: Experiment one showed leading questions affect numerical estimates; experiment two demonstrated they can alter actual memory content (false memory of broken glass).
-
Reconstructive hypothesis: Memory is not a recording but a reconstruction combining original information and post-event information, which merge over time.
-
High control, low validity: Laboratory conditions allowed precise testing and replication, but the artificial setting and student sample limit how well findings generalise to real eyewitness situations.
-
Practical impact: Research findings led to changes in police interview techniques to reduce the use of leading questions and improve the accuracy of witness testimony.