The science of intelligence
What is working memory?
Short answer
Working memory is the system that holds a small amount of information in mind while you use it. Baddeley and Hitch modeled it in 1974 as a control system plus separate stores for verbal and visual material. It holds about four chunks by Cowan’s estimate, and it is strongly related to fluid reasoning without being the same thing.
Want your own number? The test is 25 visual puzzles, about 12 minutes, no sign-up.
The Baddeley and Hitch model
In 1974 Alan Baddeley and Graham Hitch proposed a three-component model of working memory (Baddeley, 2000). It has become the standard way to describe the mind's short-lived workspace, and a fourth part was added later.
| Component | What it does |
|---|---|
| Central executive | A limited-capacity system for attention-demanding control |
| Phonological loop | Holds verbal information, which fades quickly unless refreshed by silent rehearsal |
| Visuospatial sketchpad | Holds visual and spatial information, which turn out to be separable |
| Episodic buffer (added 2000) | A limited store that binds information from the other parts and long-term memory |
The central executive and phonological loop are described in a 2025 review by the model's authors, which also notes that the sketchpad has separable visual and spatial components (Hitch et al., 2025). Baddeley proposed the episodic buffer in 2000 (Baddeley, 2000).
How much it holds: seven or four?
George Miller's 1956 paper argued that the span of immediate memory severely limits how much information we can take in and remember, and that grouping input into chunks stretches this bottleneck (Miller, 1956). His "magical number seven, plus or minus two" became one of the best known numbers in psychology.
Nelson Cowan later noted that Miller's seven was meant more as a rough estimate and rhetorical device than a real capacity limit, and brought together evidence for a central capacity limit averaging about four chunks (Cowan, 2001). The difference matters less than the shared point: the workspace is small, and chunking (seeing a phone number as three groups rather than ten digits) is how people get more into it.
Try one · Pattern Recognition
Which tile comes next?
Tap the option you think is right, or show the answer.
Show the answer
The missing tile shows four outlined crosses (option 4).
How to solve it
- Compare neighboring tiles and name what changes each time.
- From tile to tile, one more cross is added.
- The fill cycles outline, hatched and solid, then starts again.
- Everything else stays the same from tile to tile.
- The last tile shows three solid crosses. Apply each change once more.
- So the missing tile shows four outlined crosses.
That was 1 puzzle. The full test has 25.
Working memory and intelligence
In 2005 Ackerman, Beier and Boyle meta-analyzed 86 samples and found an average true-score correlation of .479 between working memory and general intelligence, which they called substantially less than unity (Ackerman et al., 2005). Kane, Hambrick and Conway replied with a reanalysis of 14 latent-variable data sets covering more than 3,100 young adults. They found a median correlation of .72 between working memory capacity and fluid reasoning, or about 50% shared variance (Kane et al., 2005).
The two sides are closer than they sound. Kane and colleagues agreed that working memory capacity is not the same thing as fluid intelligence; they argued the link is stronger than Ackerman's team indicated (Kane et al., 2005). Part of the gap comes from what each compared: general intelligence in one case, fluid reasoning in the other. A 2023 network analysis of the Woodcock-Johnson IV norm sample (ages 9 to 19) went further and suggested a working memory and attentional control complex may be central to how the broad abilities fit together (McGrew et al., 2023). See CHC theory for where short-term working memory (Gwm) sits among the broad abilities.
You can see it in reasoning puzzles
Hard reasoning items often demand working memory because several rules must be held at once while you check the options. Compare matrix puzzles at easy and hard levels and you will feel the load rise. The working memory glossary entry has tips for managing it on a timed test.
Does working memory training work?
If working memory and reasoning are so closely linked, training one might seem a route to the other. The evidence says no. A 2013 meta-analysis of 23 studies found reliable short-term gains on working memory tasks themselves but no convincing evidence that the gains generalize to other skills such as verbal and nonverbal ability (Melby-Lervåg & Hulme, 2013). A larger 2016 analysis of 87 publications and 145 comparisons found no convincing evidence of improvement on far-transfer measures such as nonverbal ability when training was compared with a treated control group, and concluded the effects are short-term and specific (Melby-Lervåg et al., 2016).
More on what does and does not move measured intelligence is in can you raise your intelligence. Working memory also overlaps with the control skills covered in executive function.
Sources
- Baddeley, A. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4(11), 417-423. doi.org/10.1016/S1364-6613(00)01538-2
- Hitch, G. J., Allen, R. J., & Baddeley, A. D. (2025). The multicomponent model of working memory fifty years on. Quarterly Journal of Experimental Psychology, 78(2), 222-239. doi.org/10.1177/17470218241290909
- Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63(2), 81-97. doi.org/10.1037/h0043158
- Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87-114. doi.org/10.1017/S0140525X01003922
- Ackerman, P. L., Beier, M. E., & Boyle, M. O. (2005). Working memory and intelligence: The same or different constructs? Psychological Bulletin, 131(1), 30-60. doi.org/10.1037/0033-2909.131.1.30
- Kane, M. J., Hambrick, D. Z., & Conway, A. R. A. (2005). Working memory capacity and fluid intelligence are strongly related constructs: Comment on Ackerman, Beier, and Boyle (2005). Psychological Bulletin, 131(1), 66-71. doi.org/10.1037/0033-2909.131.1.66
- McGrew, K. S., Schneider, W. J., Decker, S. L., & Bulut, O. (2023). A psychometric network analysis of CHC intelligence measures. Journal of Intelligence, 11(1), 19. doi.org/10.3390/jintelligence11010019
- Melby-Lervåg, M., & Hulme, C. (2013). Is working memory training effective? A meta-analytic review. Developmental Psychology, 49(2), 270-291. doi.org/10.1037/a0028228
- Melby-Lervåg, M., Redick, T. S., & Hulme, C. (2016). Working memory training does not improve performance on measures of intelligence or other measures of "far transfer." Perspectives on Psychological Science, 11(4), 512-534. doi.org/10.1177/1745691616635612