Hair typing systems

Key takeaways

• Four systems dominate: André Walker (1A-4C), LOIS, FIA, and L'Oréal's 8-type geometric classification Callender 2026, which comes from two large worldwide studies De la Mettrie 2007, Loussouarn 2007.
• The Walker chart is the best known and the least scientific. A stylist built it in the 1990s to market his products, it is qualitative, and no published study has tested whether two people would give the same hair the same code Callender 2026, Gaines 2023.
• The only measured system is also the only one tested for reliability, and it struggled. Independent raters agreed only moderately on 8 groups (kappa 0.38-0.46). Merging the four curly groups into two (6 groups) raised agreement to good (kappa 0.61-0.73) Mkentane 2017.
• Every system hits the same wall: one head has several curl types. Ten hairs from a single 0.56 cm2 patch never fell into one class; each person spanned 2-3 classes Mkentane 2017.
• Race-based labels are the worst system. African, Asian and Caucasian hair overlap heavily in curl, size and shape, and measured curl classes cut across them Daniels 2023, Mkentane 2017.
• Careful observation still works. Trained observers' curl ratings found the same genetic signal as machine measurements Ho 2020, and women's self-descriptions of "fine hair" matched instrument data Bouabbache 2016.
• "Texture" means different things in different systems. In Walker it means curl pattern; in FIA and LOIS it means strand thickness or feel.
• Keep the familiar 1A-4C letters if they help you communicate, but treat thickness, density, porosity and scalp as separate things.

The systems at a glance

LOIS letters: L has sharp bends and little curve, O rolls into an O or spiral, I is straight and lies flat, S is an S-wave or curl. FIA volume: under 5 cm ponytail is thin, 5-10 cm normal, over 10 cm thick.

How the 8-type method works

• About 6 cm of virgin hair from near the root is washed and air-dried so the natural curl returns.
• Curve diameter: each strand is laid on a template of half-circles and matched to the best-fitting circle. This separates groups I-IV.
• Curl meter: for tighter curls, check whether the 6 cm strand fits inside a 0.98 cm circle (types VII-VIII) or not (V-VI).
• Waves: a 5 cm segment is shortened to 4 cm and the waves counted: up to 3 is V, 4 or more VI, up to 5 VII, 6 or more VIII. This was the least reliable and hardest step.

All of this is from Mkentane 2017. The exact curve-diameter cut-offs for I-IV come from the 2007 papers and are not reproduced here.

How the systems roughly map onto each other

The Walker to 8-type mapping follows a L'Oréal-funded review Callender 2026. It is approximate: no study has measured the same hair in two systems and checked the overlap.

Curl in populations

• In the validation group of a DNA-prediction study, European hair was 40.7% straight, 46.6% wavy and 12.7% curly (n=2,138). East and West Asian hair was 46.7% straight, 41.3% wavy and 12% curly (n=92). African hair was 94.9% curly and 5.1% wavy, but only 39 people were included Pospiech 2018, as summarized by Daniels 2023.
• In the 8-type samples, certain types were more common in particular regions, but no region mapped onto a single type Loussouarn 2007.

What people actually care about

• In a 506-person survey of women with curly hair (3A and above), curl type was the only hair attribute linked to their goals and routines Daniels and Heitmayer 2024.
• A mixed-methods study (14 interviews plus a 212-person survey) found four kinds of hair goals: aesthetic, feel, practical and emotional. These predicted hair esteem better than length or curl type did, so a useful classification should capture goals as well as fiber traits Daniels and Heitmayer 2024.
• Researchers have argued for measuring fiber shape quantitatively rather than classifying by eye Lasisi 2021.

What it means for your hair

• Curl pattern is the axis all four systems agree on. Describe yours as a main type plus a range, and say where it differs (front, crown, nape). See curl pattern.
• Strand thickness is separate from curl. See strand thickness.
• Density or volume is separate again. See hair density.
• Porosity is mostly a damage marker. See hair porosity.

Myths and facts

• "1A-4C is a scientific classification." It was created to sell products and has never been reliability-tested Callender 2026, Gaines 2023.
• "The letters A, B, C mean thickness." For types 2-4 they mean curl tightness. In some type-1 versions they describe strand thickness.
• "Measured systems are objective, so they are reliable." Even the measured 8-type system gave only fair-to-moderate agreement between trained raters until it was simplified Mkentane 2017.
• "Your hair has one type." Sampled hairs from one person spread across 2-3 classes Mkentane 2017.
• "Hair type equals ethnicity." Measured curl classes cut across self-identified groups Mkentane 2017, Daniels 2023.

Practical points

1. Judge curl on washed, product-free, air-dried hair (see taking a good hair photo for how styling and wetness distort curl).
2. Accept that two experts may disagree by a subtype. Family level (straight, wavy, curly, coily) is the part to trust.
3. Do not fold thickness, density or porosity into the curl letter.
4. For more on describing your own hair, see how to work out your hair type at home.

Bottom line

Hair typing systems are handy shorthand, not measurements. The popular chart has no reliability data, the one measured system needed simplifying to become reliable, and every head holds several curl types. Use the letters to find products and styling advice, and keep your own observations in charge.