Sixteen Strokes a Day, Measured on the Wrong Heads

Every anti-breakage claim on a shampoo bottle rests on a laboratory test, and every laboratory test rests on an assumption about what real hair goes through in a real bathroom.

In 2018 a team went and measured it, because nobody had. They asked 231 women how often they comb, filmed ten of them doing it, and then attached single fibers in a tress to a load cell to record what each hair actually experiences.

The women were Asian and Caucasian. All 231 of them, and all ten of the filmed ones.

What the benchmark says

The numbers are worth having, because they are more specific than anything else in the field.

People comb an average of 1.7 times a day, with a range from none to five. That works out at about sixteen comb strokes per head per day, plus or minus 2.3. Combing speed in life ran between 22 and 35 centimeters per second across hair types, which is why the laboratory protocol settled on about 25.

Then the load cell. The team defined an event as any single fiber taking a load of one gram or more, and ran thirty combing sets, each of ten strokes with tangling in between.

Dry combing on virgin hair: an event probability of 0.2 and an average load of 1.7 grams. Wet combing on virgin Asian hair: 0.1. Wet combing on virgin Caucasian hair: 0.47 — nearly five times as often, on the same protocol, which is itself an argument that hair type changes the answer.

Wet-combed bleached hair: a probability of 1. Every single set. Conditioner brought that back below 0.1.

Their conclusion is the part the industry should have found uncomfortable. During combing, individual fibers may not experience any significant loads and are unlikely to experience repetitive loads above ten grams. And, in their words, the data indicates the need for a significant rethink of methods used for product evaluation and claim substantiation.

A group set out to find the real-world load, discovered that standard laboratory tests are probably pulling much harder than a person does, and said so.

Why that exclusion is the wrong one to make

Of all the hair types you could leave out of a study of combing damage, Afro-textured hair is the one where combing damage has been shown to matter most.

The demonstration is already on this site: in combed natural hair, length plateaus within about a year of a haircut because breakage catches up with growth, and a controlled comparison of frequent and occasional braiders found cuticle damage rising in a linear trend with the number of combing strokes — with 480 strokes doing more surface damage than frequent braiding did. The fiber itself is not defective. The handling is what sets the ceiling. I went through all of that in the piece on steady state.

So the sequence is: the population in which combing is the demonstrated mechanism is absent from the dataset that describes what combing does. Claims are substantiated on a protocol whose real-world calibration came from hair that behaves differently, and then printed on products sold into the market where the mechanism actually bites.

That is not fraud. Nobody is lying on a bottle. It is a measurement that stopped one population short, and a whole industry standing on top of it.

The same shape, in the pathology lab

There is a version of this that is not about shopping at all.

When a dermatologist cannot tell what is causing hair loss, the next step is a punch biopsy: a four-millimeter core of scalp, sectioned, with the follicles counted. The count is then read against a reference range, and the reference range decides whether the answer is “normal density” or “reduced density.”

Sperling published the check on that in 1999. He counted follicles in scalp biopsies from 22 African American and 12 white patients with clinically healthy scalps, at three levels through the specimen, and compared them against previously reported data drawn from predominantly white patients.

Total hair density, total terminal follicles and total terminal anagen hairs were all significantly lower in the African American specimens, at p below 0.001.

His conclusion is stated as a clinical warning rather than an observation: data previously collected from white patients may not provide adequate guidance when evaluating scalp biopsy specimens from an African American patient, and could lead to an incorrect diagnosis.

That was published in the Archives of Dermatology in 1999, from a sample of thirty-four people. It has not, as far as I can find, been superseded by anything larger.

A 2015 case-control study of hair care practices and hair shaft parameters in African American and Caucasian women found significant differences in growth rate, density, diameter, number of broken hairs and scalp blood flow — and, notably, no significant difference in hair cycle parameters. The cycle is the same. What is measured off the cycle is not. That paper does not state its sample size in the abstract, and does not say in the abstract which group was higher on each biological measure, so I am reporting the p-values and not inventing directions for them.

What is actually well measured

It is worth saying that the fiber science, where it has been done, is good.

A 2022 study measured hair porosity properly — nitrogen sorption porosimetry, the same physics used to characterize catalysts, applied to the pore surface area inside the cortex — alongside single-fiber tensile testing. Damage showed up as a statistically significant decrease in break stress and toughness, and coconut-based hair oil reversed both. The proposed mechanism is that the oil blocks diffusion pathways in the endocuticle and the matrix of the cortical cells, which is a specific and testable claim rather than a nourishing one.

And in 2002, a group demonstrated that a ceramide binds to relaxed African-American hair, using radioactivity microimaging and secondary ion mass spectrometry to show where it went, then showed less breakage when a shampoo containing it was used. That one comes with a caution: the breakage test was an in-house brushing method with no published validation I could locate, and the paper does not state how many fibers or heads were involved.

Both are real work. Neither is the thing that is missing. What is missing is the plain descriptive dataset — how often, how fast, how hard, on tightly coiled hair — that would let anyone calibrate a test to it.

The bottle’s percentage, and whose hair it came from

When a bottle says it reduces breakage by some percentage, the percentage came from a tress on a machine, and the machine was set to a force profile derived from combing that nobody in the underlying study had hair like yours.

That does not make the product useless. Conditioner taking wet-combing events on bleached hair from certain to almost never is a large, real, measured effect, and it will not have stopped being real because the sample was narrow.

It does mean the number on the front is softer than it looks, and it means something more specific for the reader this section is written for: the one hair type where the published evidence says grooming force sets how long your hair gets is the hair type that the grooming-force benchmark left out. Until somebody films thirty women with coils and puts their fibers on a load cell, everybody is extrapolating — the brands, the trichologists, and me.

This article is educational and is not a substitute for care from someone who can look at your skin. Anything new, changing, painful or spreading should be assessed rather than researched.