How the Hair Pattern Assessment Works — Methodology

On this page 9 sections

1. What this assessment is (and is not)
2. What we intentionally do not do
3. The Andre Walker typing system: a shared vocabulary, not an objective test
4. Most scalps carry more than one pattern
5. Alternatives, and what is actually measurable
6. Where curl comes from: the follicle, not ethnicity
7. The limits of any visual assessment
8. Why we still use the vocabulary anyway
9. Sources

The short version: curl-pattern typing is a shared vocabulary, not a measurement. It is genuinely useful for describing what hair looks like and for talking to a clinician or stylist. It is not an objective biological test, and no visual assessment — including ours — can substitute for an in-person clinical examination.

This page explains the reasoning behind our Hair Pattern Assessment. It is written the way we write everything on this site: we say what is supported by evidence, we say plainly what is not known or not measurable, and we do not dress up a rough visual guess as a precise result.

What this assessment is (and is not)

The Hair Pattern Assessment helps you put a common name to the shape of your hair — roughly straight, wavy, curly, or coily — using the vocabulary most people and most clinics already use. That is its whole job.

It is not:

• a biological or medical classification,
• a diagnosisDiagnosis A distinct step in the procedure, attributable to a named person. of any hair or scalp condition,
• a measurement of hair density, strand caliber, porosity, elasticity, or donor capacity, or
• an assessment of whether you are a candidate for a hair transplantHair transplantation Surgical relocation of follicles from a donor area to a recipient area..

Those last points require a clinician looking at your scalp in person, often with magnification. A curl label describes the shape of the fibre you can see; it says nothing reliable about the numbers a surgeon actually plans around.

What we intentionally do not do

This tool does not:

• diagnose disease
• predict a surgery outcome
• predict donor quality
• estimate transectionTransection Cutting through a follicle during extraction, damaging or destroying the graft.
• estimate graft survival
• infer ethnicity
• judge candidacy for a procedure

Every one of these is a clinical judgement made from an in-person examination — not from a photo, a questionnaire, or a curl label. Drawing that line is deliberate: it is the difference between a tool that informs you and one that pretends to decide for you.

The Andre Walker typing system: a shared vocabulary, not an objective test

What it is

Most “what’s my hair type?” charts descend from the system popularized by stylist Andre Walker in the 1990s. It sorts hair into four broad families, each with A/B/C sub-steps:

• Type 1 — straight (1A–1C)
• Type 2 — wavy (2A–2C)
• Type 3 — curly (3A loose curls → 3C corkscrews)
• Type 4 — coily / kinky (4A defined coils → 4C tightest coils)

The curly-to-coily range most relevant to textured hair runs roughly from 3A through 4C. As a common language, this works well — it lets two people picture roughly the same thing when they say “3C” or “4B.”

Two honest caveats about its origins: it was created in the context of marketing a hair-care line, not as a research instrument, and it has been openly criticized for implying a hierarchy that treats straighter hair as a baseline. We use the numbers only as descriptive shorthand, with no implication that any pattern is better or worse than another.

Why it is subjective

A board-certified dermatologist has described the system on the record as “built on visual observation — as in, what hair looks like to the human eye, not on biology,” with “no biological or scientific basis” as a classification (Dr. Oyetewa Asempa, Baylor College of Medicine). It sorts by appearance, not by any measured property of the hair.

Because it relies on eyeballing, different people — and even the same person on a different day, in different light, with hair at a different length or dampness — can reasonably land on different labels for the same head of hair.

We looked for, and did not find, a peer-reviewed study that formally measures inter-rater reliability for the Andre Walker system specifically. We will not invent a number we cannot cite. What the literature does show is that visual scalp-and-hair scales are hard to apply consistently even for trained clinicians: a reliability study of the Hamilton–Norwood scale — a different, clinician-facing system for pattern hair lossAndrogenetic alopecia Genetically driven progressive hair loss following a recognised pattern. — reported an intraclass correlation of only about 0.65 and concluded that its reliability was “unsatisfactory even in the hands of expert appraisers” (Guarrera et al., 2009). That does not measure the Walker chart, but it is a fair warning: if experts disagree when classifying hair by eye with a clinical scale, a consumer curl chart should be treated as approximate, not exact.

Most scalps carry more than one pattern

A single label almost always oversimplifies. It is normal — closer to the rule than the exception — for hair to be shaped differently in different regions of the same scalp: the crown, the mid-scalp, the temples/edges, and the nape can each read as a different “type.” Sun, friction, heat and chemical processing can loosen or alter pattern over time, and the nape is often tighter than a sun-exposed crown.

This is one reason a one-letter answer can mislead. If you use this tool, it is more accurate to describe your hair as a range (for example, “mostly 4A with looser 3C at the edges”) than to force a single verdict. That is exactly why our result is always a range, never a single category.

Alternatives, and what is actually measurable

Curl “type” is not the only way to describe hair, and it is worth separating consumer labels from parameters that dermatology can actually measure.

The L.O.I.S. system

The L.O.I.S. system classifies a single strand by the shape it naturally forms: L (bends and sharp angles), O (curls/coils into an O), I (straight), S (waves). It adds two further descriptors — strand size (fine / medium / coarse, compared against a sewing thread) and strand texture (silky, cottony, spongy, wiry). Like the Walker system it allows combinations, precisely because one head of hair often shows more than one pattern. It is still a descriptive, visual scheme — but it captures thickness and surface texture that a simple 1–4 curl number ignores.

Measurable fibre geometry

Distinct from consumer “type” labels, hair-fibre shape has genuinely quantifiable parameters that researchers measure directly:

• Curvature / curl diameter — how tightly the fibre bends,
• Cross-sectional shape (ellipticity) — curlier fibres tend to have a more flattened, elliptical cross-section, straighter fibres a rounder one, and
• Curl descriptors such as curve diameter, a curl index, and the number of waves and twists along a strand.

Using measurements like these, researchers proposed a more objective scheme that sorts hair worldwide into eight groups rather than four, deliberately moving away from ethnic labels toward measurable features (Loussouarn et al. and De la Mettrie et al., 2007). These are laboratory measurements, not something a phone photo or a home mirror can produce — which is the reason we are careful to call our tool an assessment of appearance, not a measurement of the fibre.

Where curl comes from: the follicle, not ethnicity

A common misconception is that curl pattern is a property of ethnicity. The better-supported view is that the shape of the hair is set by the shape and biology of the follicle that grows it, and that this tracks with curl pattern independently of ethnic origin.

• Thibaut et al. dissected and cultured hair follicles and found that curly follicles are physically bent — curved at the root with a “retrocurvatureFollicular curvature The curved path a follicle takes beneath the skin, pronounced in tightly curled hair.” at the bulb, and asymmetric in how the growing cells divide. Crucially, follicles kept that curvature even when grown in vitro. Their conclusion: “Hair curliness is programmed from the bulb” (British Journal of Dermatology, 2005).
• A later study by the same group confirmed that this bulb “retrocurvature” was present in curly hair “independent of ethnic origin,” that curlier fibres were more elliptical in cross-section, and that the fibre behaves like a “shape-memory material” whose curl is built in at the level of the follicle’s growing cells (International Journal of Dermatology, 2007).

In plain terms: your curl comes from the geometry of the follicle under your skin, not from a category on a chart. Two people from very different backgrounds can grow the same-shaped hair, and one person can grow more than one shape.

A note on sourcing. An earlier working reference for this project labelled the second study “J Struct Biol 2006.” The PubMed record at that identifier is in fact Thibaut, Barbarat, Leroy & Bernard, International Journal of Dermatology, 2007. We cite it as it actually appears rather than repeat an inaccurate label — the same standard we hold every claim on this site to.

The limits of any visual assessment

Be clear-eyed about what a look-and-label tool cannot do. This assessment cannot:

• Measure the things that matter clinically — hair density (hairs per cm²), strand caliber, follicular-unit density, degree of miniaturization, or donor-area capacity. These require in-person examination, often with a densitometer or trichoscope.
• Diagnose anything. It cannot detect or rule out pattern hair loss, telogen effluvium, scarring alopeciaScarring alopecia Alopecia in which the follicle is destroyed and replaced by scar tissue., or scalp disease. Sudden changes in texture or pattern can occasionally signal a medical issue and should be checked by a professional, not by a chart.
• Predict how your hair will behave in response to a product, climate, stress, or shifts in your own biology.
• Give a stable, permanent answer. Pattern can shift with age, hormones, pregnancy, medication, and heat/chemical processing, so any label is a snapshot of today.
• Judge candidacy or plan a procedure. Nothing about a curl label determines whether a hair transplant is appropriate, how many grafts are needed, or what result to expect. Those are clinical decisions made from an in-person assessment.
• Escape the subjectivity of the input. The result is only as good as the glance behind it — lighting, length, dryness, product, and styling all change the apparent pattern.

Known limitations of this assessment

This assessment is at its least reliable when the natural pattern is hidden or altered. Treat the result with extra caution — or wait for a better look — when:

• the hair is chemically straightened or relaxed,
• the hair is in braids, twists, or a protective style,
• the hair is very short,
• the hair is heavily weighed down with product or moisturiser,
• more than one texture is present across the scalp, or
• you are judging it quickly, in poor light, or from memory.

Almost no curl-typing tool states these out loud. We do, because a result you can trust is worth more than one that hides its own blind spots.

What this tool does: puts a common, descriptive name to the visible shape of your hair.

What it does not do: it is not a medical test, not a diagnosis, and not a measurement. It cannot assess density, caliber, donor capacity, or transplant candidacy. Most scalps show more than one pattern, so treat the result as an approximate range, not a verdict. Curl shape is set by your follicles, not by ethnicity, and it can change over your lifetime. For anything involving hair loss or a possible procedure, rely on an in-person clinical examination — not this tool.

Why we still use the vocabulary anyway

If the labels are subjective, why offer the tool at all? Because a shared vocabulary is genuinely useful — as long as everyone knows it is a vocabulary and not a measurement.

• It helps you communicate. “My crown is around 4A and my edges are looser, maybe 3C” gives a clinician or stylist a fast, roughly-shared mental picture to start from.
• It sets a common starting point for discussing expectations, care, and — where relevant — surgical planning, when both sides use the same words.
• It is a conversation-opener, not a conclusion. The right use of a curl label is to begin a discussion that a professional then refines with real observation and measurement.

We include the tool for that reason and no other: to give you and your clinician a shared word to start from. Everything that actually determines a treatment decision comes from the in-person examination, not from the label.

Sources

Peer-reviewed literature

• Thibaut S, Gaillard O, Bouhanna P, Cannell DW, Bernard BA. “Human hair shape is programmed from the bulb.” British Journal of Dermatology. 2005;152(4):632–638. PubMed
• Thibaut S, Barbarat P, Leroy F, Bernard BA. “Human hair keratin network and curvature.” International Journal of Dermatology. 2007;46 Suppl 1:7–10. PubMed — the record at the identifier an earlier reference labelled “J Struct Biol 2006,” cited here per its actual publication.
• Loussouarn G, Garcel AL, Lozano I, et al. “Worldwide diversity of hair curliness: a new method of assessment.” International Journal of Dermatology. 2007;46 Suppl 1:2–6. PubMed
• De la Mettrie R, Saint-Léger D, Loussouarn G, et al. “Shape variability and classification of human hair: a worldwide approach.” Human Biology. 2007;79(3):265–281. PubMed
• Guarrera M, Cardo P, Arrigo P, Rebora A. “Reliability of Hamilton-Norwood classification.” International Journal of Trichology. 2009;1(2):120–122. PMC — cited as an analogy only; it concerns a different, clinician-facing scale, not the Andre Walker system.

Sourcing caveats, stated openly

• No inter-rater reliability figure exists for the Andre Walker system itself, as far as we could find. We therefore describe its subjectivity qualitatively and use the Hamilton–Norwood reliability study only as a clearly-labelled analogy.
• The “varies across the scalp” and consumer-system descriptions rely partly on secondary sources (dermatologist interviews, curl-community references) rather than primary research, because that is where these particular points are documented. We have not attached primary-research citations to claims that only secondary sources support.