Criminological psychology (option) classic study: Loftus and Palmer (1974)

Experiment 1

Aim: to test whether the wording of a question affects estimates of speed (a leading question suggests the answer wanted).

• 45 students in groups watched 7 films of traffic accidents (5 to 30 seconds), then wrote an account and answered questions.
• The critical question: 'About how fast were the cars going when they ___ each other?' with the verb smashed, collided, bumped, hit or contacted (9 participants per verb; independent groups).
• Four films were staged crashes at known speeds (20, 30, 40 and 40 mph).

Pearson WPS03 summary of studies

Experiment 2

• 150 participants watched a film of a multiple car crash (under 1 minute; the crash lasted 4 seconds).
• Smashed group (50), hit group (50) and a control group (50) not asked about speed.
• One week later, without seeing the film again, they answered 'Did you see any broken glass?' There was no broken glass in the film.
• Mean speed: smashed 10.46 mph, hit 8.00 mph.

Answers to 'Did you see any broken glass?'

Conclusions

Leading questions affect answers. Two explanations: response bias (uncertain participants use the verb as a clue) and memory change: the verb alters the memory representation, so more severe verbs make the accident seem more severe. Experiment 2 supports memory change: a week later, 'smashed' participants were more likely to 'remember' broken glass. Loftus and Palmer proposed that information from the event and information supplied afterwards integrate into one memory.

Evaluation

• Strengths: laboratory control and standardised films and questions give reliability and allow cause and effect; a control group in experiment 2; clear practical application (police avoiding leading questions; cognitive interview).
• Weaknesses: films lack the emotion and stakes of real accidents (low ecological validity; Yuille and Cutshall found real witnesses resisted leading questions); student samples; people are poor at judging speed anyway; participants may have guessed the aim (demand characteristics). Ethics: low risk; mild deception about the purpose.

Worked example

Calculate chi-squared for the broken glass data (smashed, hit, control) and say whether the difference is significant at p ≤ .05 (two-tailed).

1. Each column total is 50 and 29 said yes out of 150, so expected 'yes' = 50 x 29 / 150 = 9.67 and expected 'no' = 40.33 in each group.
2. χ² = Σ (O - E)² / E = 7.78.
3. df = (2 - 1) x (3 - 1) = 2.
4. Critical value (df 2, two-tailed, p ≤ .05) = 5.99. 7.78 ≥ 5.99, so significant: reject the null hypothesis (it is not significant at p ≤ .01, where the critical value is 9.21).

Answer: χ² = 7.78, df 2, significant at p ≤ .05.

Check yourself

1. In experiment 1, the mean for 'smashed' was 40.5 mph and for 'contacted' 31.8 mph. What is the difference? Give your answer to 1 decimal place.
   Show answer

   8.7 mph. 40.5 - 31.8 = 8.7 mph.

2. What percentage of the 'smashed' group said they saw broken glass (16 of 50)?
   Show answer

   32 %. 16 / 50 x 100 = 32%.

3. Using the broken glass table (yes: 16, 7, 6; no: 34, 43, 44), calculate χ² to 2 decimal places.
   Show answer

   7.78. Expected yes = 9.67, expected no = 40.33 in each group; χ² = 7.78.

4. What was the independent variable in experiment 1?
   • The speed estimate
   • The number of films
   • Whether participants saw broken glass
   • The verb used in the critical question
   Show answer

   The verb used in the critical question. The verb was manipulated.

5. Which result best supports the idea that the leading question changed the memory itself?
   • Smashed gave higher speed estimates than hit in experiment 1
   • More 'smashed' participants reported broken glass a week later
   • Participants were poor at judging known speeds
   • The films lasted 5 to 30 seconds
   Show answer

   More 'smashed' participants reported broken glass a week later. A false detail recalled later suggests the memory changed, not just the answer.

6. Why was a control group used in experiment 2?
   • To increase the number of verbs
   • To show the film twice
   • To measure speed accurately
   • To give a baseline of how many people reported broken glass without a leading question about speed
   Show answer

   To give a baseline of how many people reported broken glass without a leading question about speed. A baseline for comparison.

7. Explain why Loftus and Palmer's (1974) findings may lack ecological validity.
   Show answer

   Participants watched short film clips in a laboratory, which lacks the shock, emotion and consequences of witnessing a real accident, so real witnesses may not be affected by leading questions in the same way.. Method plus what is missing.

Practise with instant feedback Print this lesson

Sources

• Pearson Edexcel International Advanced Subsidiary/Advanced Level in Psychology specification, Issue 6 (November 2024, first assessment January 2026)
• Pearson IAL Psychology summary of studies: WPS03
• Pearson WPS03/01 question paper, January 2022 (formulae and critical values tables)

Last updated 8 November 2026. Spotted a mistake? Tell us.