Baddeley (1966a)

Aim: To investigate how information is coded in short-term memory (STM).

Method: Baddeley conducted three experiments investigating coding in STM. This summary focuses on the first experiment. Baddeley used an independent groups design involving 41 participants. 20 participants were allocated to Condition A and 21 to Condition B. Each participant attempted to recall 24 sequences of five words in the correct order.

In Condition A, participants recalled 24 sequences:

• 12 sequences containing acoustically similar words, such as mad, man, mat, cap and can

• 12 sequences containing acoustically dissimilar words, such as cow, day, bar, few and pen

In Condition B, participants recalled 24 sequences:

• 12 sequences containing semantically similar words, such as big, long, broad, great and large

• 12 sequences containing semantically dissimilar words, such as old, deep, late, safe, and hot

Results: The mean percentages of sequences correctly recalled were: 

 Baddeley found that:

• recall was substantially worse for acoustically similar words than for acoustically dissimilar words

• there was only a small difference in recall between semantically similar and semantically dissimilar words

Conclusion: Baddeley concluded that information in STM is coded acoustically. Acoustically similar words were much more difficult to recall because their similar sounds caused confusion within STM. Semantic similarity had a much smaller effect, suggesting that word meanings were less important when information was recalled immediately.