Working memory
Do working memory tests secretly rely on long-term memory?
Open access · cc by · source: Europe PMC
Memory for word order, locations and colors showed no sign of leaning on long-term memory, but memory for object-word pairs did, so some popular working memory tasks are contaminated and others are not.
Study at a glance
- Design
- Human experiment — Seven within-subject experiments alternating two memory tasks in mini-blocks with fresh materials each block; accuracy across trials within a block indexes proactive interference; Bayesian mixed models
- N
- Separate samples per experiment: 50 each in 1a, 1b and 3, 49 in 2a, 60 in 2b (online, English speakers); 33 in Exp 4 and 29 in Exp 5 (in-lab German-speaking students)
- Population
- Young adults aged 18-35: English native speakers via Prolific and University of Zurich students
- Outcome
- Change in recall accuracy (or reproduction error) across trials within a mini-block for serial recall, probed recall, object-word pairs, spatial and color reproduction; protrusion errors in Exp 4
Structured fields used in claim comparison tables when every cited study has a complete layer.
Key findings
Accuracy on the object-word pairs task reliably declined across trials within a mini-block in every experiment that used it, showing proactive interference. Serial recall, spatial location memory and color reproduction showed evidence against such a decline, and probed recall showed interference only in one experiment (2a) that the authors judge a false positive after three later experiments found evidence against it. In Experiment 4, words from earlier trials were not chosen increasingly often as a block went on.
Methodology
The researchers used proactive interference, the tendency for older memories to disrupt newer ones, as a fingerprint of episodic long-term memory, since working memory is thought to be protected from it. Participants did mini-blocks of four (or eight) trials of one task with the same small set of words or objects, then switched to a different task and fresh materials, which should release any interference. Across seven experiments they compared serial recall and probed recall of word lists with an object-word pairing task, a spatial location task, and a color reproduction task, including a version with slow five-second-per-word presentation.
Limitations
The conclusion rests on the assumption that working memory contents are immune to proactive interference, which some researchers dispute. Absence of interference does not prove that no long-term trace was formed; the authors note the Hebb repetition effect shows traces exist, so their account (a 'flexible gate' that decides when to use long-term memory) is an interpretation rather than something directly tested. Samples were young adults, mostly tested online, and results for some tasks were inconsistent across experiments, which the authors attribute to sampling error.
How this study connects
Role on claims
Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.
Not yet placed on a claim. This paper has study layers, but no concept page yet cites it as support, challenge, or qualifier.
Related papers in this topic
Same topic cluster — not a recommendation engine.