Questioning short-term memory and its measurement: Why digit span measures long-term associative learning

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Neuroscience

Background:

• Traditional models of verbal short-term memory attribute performance differences to item characteristics.
• Previous experience is typically controlled by exclusion or random permutations.

Purpose of the Study:

• To investigate the basis for superior short-term memory recall of digit sequences compared to other verbal materials.
• To challenge the assumption that inherent item characteristics drive memory performance.

Main Methods:

• Four experiments were conducted to compare recall of digit sequences versus other verbal materials.
• Corpus-based computational analysis was used to assess sequence frequency in natural language.
• Associative learning processes were examined in relation to linguistic experience.

Main Results:

• The advantage for digit sequences is not due to inherent digit properties but their higher frequency in natural language sequences.
• More frequent digit sequences are recalled better in short-term serial recall tasks.
• Performance is linked to associative learning based on linguistic experience.

Conclusions:

• Short-term memory performance is influenced by long-term memory processes and learned linguistic regularities.
• The findings question the general use of digit span as a cognitive measure.
• Learned sequence frequency, not item distinctiveness, explains digit span advantages in memory.