The 4.4% figure has a specific denominator
Simner and colleagues identified 22 people among 500 participants with at least one tested form of synesthesia. The participants came from the communities of Edinburgh and Glasgow universities. Reports were followed up with consistency tests rather than counted from a simple yes/no question alone.
That produces 4.4% in this sample. The recruitment was opportunistic, so the result should not be presented as a precise count of everyone in the world.
Letters and numbers are a narrower question
For the comparison between samples, the researchers defined grapheme–color synesthesia as experiencing colors for both letters and numerals. Seven of the 500 university participants met that definition: 1.4%.
A separate study of 1,190 visitors to London’s Science Museum found thirteen cases: 1.1%. Those visitors included children, and recruitment took place in June–August 2004. This part of the research tested the grapheme–color variant rather than every form.
The two figures therefore do not contradict 4.4%. They concern a narrower experience, and the museum result also describes a different participant group. The explorer keeps those choices visible rather than averaging them into a new percentage.
Why reports were checked
An association reported once can have several explanations. The university study compared consistency over time with control participants, including surprise retesting. The museum study used a consistency task and, for participants aged twelve and over, a questionnaire.
These procedures matter when comparing an estimate from tested participants with a survey that simply asks whether people have unusual associations. The number depends on the definition and the measurement as well as on the people recruited.
One person can have several forms
The seven university participants with colors for letters and numbers are included in the 22 with any tested form. Adding their percentages would count those people twice. Likewise, adding the two study sample sizes would not create a larger sample tested for all forms.
Every download row retains its own count, denominator and definition. For another example of how a testing rule changes the answer, compare the perfect-pitch study groups.
Good questions
How common is synesthesia?
Simner and colleagues identified 22 people with at least one tested form among 500 participants from Edinburgh and Glasgow university communities: 4.4%. This was a convenience sample, not a worldwide census.
How rare is grapheme–color synesthesia?
For color experiences linked to both letters and numbers, the study found seven of 500 university participants, or 1.4%, and thirteen of 1,190 museum visitors, or 1.1%.
Why not add the percentages for different forms?
A participant can have more than one form. The seven university participants with letter-and-number colors are already included among the 22 people with any tested form.