Do more colours make a crowd of dots look bigger?
Arrays made of many different colours looked more numerous than one-colour arrays, while colour helped counting only when same-coloured items were grouped together.
Source
The influence of increasing color variety on numerosity estimation and counting
Study at a glance
- Design
- Human experiment — Two within-subject psychophysics experiments manipulating colour variety (1, 4 or 8 colours), spatial arrangement (clustered vs random) and numerosity (13-41 dots).
- N
- N=60 · 30 undergraduates in Experiment 1 (estimation) and a new group of 30 in Experiment 2 (counting).
- Population
- Chinese university students aged 18-23 with normal colour vision.
- Outcome
- Experiment 1: estimation error (estimate minus true number) after a brief display; Experiment 2: counting response time and error rate.
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
In Experiment 1, 30 students saw arrays of coloured circles for just 300 ms and said aloud how many there were; arrays had one, four or eight colours, arranged either in same-colour clusters or randomly mixed. In Experiment 2, a new group of 30 students counted similar arrays as quickly and accurately as possible, with the display staying on until they finished. Both experiments used five target quantities between 13 and 41 plus filler quantities to stop people guessing.
What they found
In the estimation task, eight-colour arrays were judged more numerous than single-colour arrays (estimates up by roughly 0.77-4.12%), regardless of whether colours were clustered or mixed; four colours did not differ reliably from one colour. In the counting task, clustered colours made counting of the largest arrays faster (about 7.05% faster with eight colours) and more accurate, whereas randomly mixed colours slowed counting without changing accuracy. The authors interpret this as colour variety acting as a cue for quantity under distributed attention, and colour clusters helping to guide focused attention during counting.
The limits
What it doesn't show
Each experiment had only 30 students from one university, and the gender balance differed sharply between the two groups. The size of the estimation bias was small, and the study does not pin down where it starts (four colours had no reliable effect) or whether it reflects numerosity itself versus perceived density. The attention-based explanation is inferred from behaviour rather than measured directly, and the arrays were simple dots in fixed grid positions, so the effect may differ with overlapping or real-world objects.
Key terms
- Numerosity estimation
- Judging approximately how many items are present at a glance, without counting them one by one.
- Distributed attention
- Spreading attention across a whole scene to extract global properties such as average size or overall number.
- Focused attention
- Attending to items one at a time or in small groups, as needed for serial counting.
- Central tendency effect
- The bias for judgments to drift toward the middle of the range of stimuli, so small values are overestimated and large ones underestimated.
- Within-subject design
- A design in which each participant experiences every condition, so comparisons are made within the same people.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
Which arrays were estimated as most numerous?
Common questions
Why would more colours make a set look larger?
The authors suggest people use diversity as a shortcut for quantity: a richer-looking collection is inferred to contain more things, especially when more colours exist than attention and working memory can track at once.
Why did clustering matter for counting but not for estimation?
Counting uses focused attention that moves through the display, so colour clusters act like dividers that segment it; quick estimation takes in the whole display at once, so arrangement mattered little.
Does colour variety make estimates worse?
Not uniformly. It pushed estimates upward, which made small arrays more overestimated but reduced the usual underestimation of large arrays, and it did not change precision.
More on Numerical cognition