Practical investigation: a repeated measures experiment

Planning

Example (from the spec): a dual-task experiment on the phonological loop. Participants learn word lists in silence (condition A) and while repeating 'la la la' aloud (articulatory suppression, condition B).

• Hypothesis (directional): 'Participants will recall more words out of 10 in silence than during articulatory suppression.'
• IV: silence or suppression. DV: number of words recalled correctly in 1 minute.
• Design: repeated measures (controls participant variables). Counterbalance the order and use two word lists matched for length and frequency.
• Sample: opportunity sample of 10 to 20 (justify it); informed consent, right to withdraw, debrief.
• Controls: standardised instructions, same room and timing, randomise word order.

Analysis

Example results (12 participants, words out of 10):

A: 7, 6, 8, 5, 7, 6, 8, 7, 6, 5, 7, 8

B: 4, 6, 5, 4, 5, 3, 6, 5, 7, 3, 4, 5

Mean A = 6.67, mean B = 4.75; median A = 7, median B = 5; SD A = 1.07, SD B = 1.22.

Wilcoxon: one zero difference, N = 11; positive ranks total 64.5, negative ranks total 1.5; T = 1.5. Critical value for N = 11, one-tailed p ≤ .05 = 13; T is lower, and the difference is in the predicted direction, so it is significant: reject the null hypothesis.

Present a summary table and a bar chart of the two means.

Write-up and evaluation

Procedure: enough detail to replicate (materials, timings, instructions, counterbalancing).

Results: summary table, bar chart, Wilcoxon result with T, N, critical value, significance level and decision.

Discussion: link to the working memory model (suppression occupies the articulatory control process, so rehearsal in the phonological loop is blocked); evaluate reliability (standardised, replicable), validity (artificial word lists: low ecological validity; demand characteristics in repeated measures), objectivity (counting words is objective) and experimenter effects; suggest improvements (larger, randomly chosen sample; double-blind marking).

Worked example

Using the example data, state the Wilcoxon conclusion in a sentence suitable for a report.

1. T = 1.5, N = 11, critical value 13 (one-tailed, p ≤ .05).
2. T is equal to or less than the critical value and the direction matches the hypothesis.
3. So the result is significant.

Answer: The Wilcoxon test gave T = 1.5 (N = 11), below the critical value of 13 for a one-tailed test at p ≤ .05, so participants recalled significantly more words in silence than during articulatory suppression; the null hypothesis is rejected.

Check yourself

1. Calculate the mean of condition A: 7, 6, 8, 5, 7, 6, 8, 7, 6, 5, 7, 8. Give your answer to 2 decimal places.
   Show answer

   6.67. 80 / 12 = 6.67.

2. Calculate Wilcoxon T for A = 7, 6, 8, 5, 7, 6, 8, 7, 6, 5, 7, 8 and B = 4, 6, 5, 4, 5, 3, 6, 5, 7, 3, 4, 5.
   Show answer

   1.5. Differences: 3, 0, 3, 1, 2, 3, 2, 2, -1, 2, 3, 3. Drop the zero (N = 11). Ranks: 9, 9, 1.5, 4.5, 9, 4.5, 4.5, 1.5, 4.5, 9, 9. Positive total 64.5, negative total 1.5, so T = 1.5.

3. Why is articulatory suppression expected to reduce recall of a word list?
   • It damages long-term memory
   • It uses up the visuo-spatial sketchpad
   • It makes participants obey the researcher
   • It occupies the articulatory control process, so words cannot be rehearsed in the phonological loop
   Show answer

   It occupies the articulatory control process, so words cannot be rehearsed in the phonological loop. Repeating a sound blocks the inner voice.

4. Why should the order of conditions be counterbalanced in this practical?
   • To stop practice or fatigue effects from favouring one condition
   • To increase the number of participants
   • To make the data nominal
   • To remove the need for a hypothesis
   Show answer

   To stop practice or fatigue effects from favouring one condition. Counterbalancing balances order effects across conditions.

5. Why should the two word lists be matched for word length and frequency?
   • So participants can memorise every word
   • So the test becomes two-tailed
   • So differences in recall are caused by the IV, not by one list being easier
   • So the data are qualitative
   Show answer

   So differences in recall are caused by the IV, not by one list being easier. Unequal lists would be a confounding variable.

6. Suggest one improvement to this experiment and explain why it would help.
   Show answer

   Use a larger, randomly selected sample so the results are more representative and can be generalised to the target population.. Give a specific change and a specific benefit.

Practise with instant feedback Print this lesson

Sources

• Pearson Edexcel International Advanced Subsidiary/Advanced Level in Psychology specification (issue January 2026)
• Pearson International A Level Psychology component guide: WPS01
• Pearson International A Level Psychology sample assessment materials (Wilcoxon critical values table)
• IAL Psychology revision site: Practical (cognitive)

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