How the prediction works
The Khamis-Roche method is a regression built from the Fels Longitudinal Study, which measured children in Ohio every six months from birth to adulthood. For each half-year of age it gives four coefficients: one each for current height, weight and mid-parent height (the average of mother and father), plus a constant. Multiply, add, and you get a predicted adult height. Its errors are only slightly larger than methods that need a hand X-ray to read bone age.
We show the result as the range the method's own error supports, not a single number. A single number looks precise and is wrong for almost everyone by a few centimetres.
How to measure for the best result
- Measure barefoot, back against a wall, heels together, looking straight ahead. Lay a book flat on the head and mark the wall.
- Measure in the morning. Everyone is slightly shorter by evening as the discs in the spine compress.
- Use measured parent heights, not remembered ones. People tend to round their own height up.
What the range means
| Method | Inputs | Typical 90% error |
|---|---|---|
| Khamis-Roche (this page) | Age, height, weight, parents | ±5.3 cm boys · ±4.3 cm girls |
| Mid-parental target | Parents only | ±8.5 cm |
| Bone-age methods | Hand X-ray + measurements | Slightly narrower, needs a doctor |
Genes explain most of the difference in adult height between people: twin studies put heritability at roughly 69–84% for men and 53–78% for women. The rest is environment: sleep, nutrition and health through the growing years. That is the part you control, and it decides whether you finish near the top or the bottom of your range.
Limits
Khamis-Roche was built on white American children without growth disorders. It is less reliable for children with medical conditions that affect growth, very early or very late puberty, or ancestry that differs strongly from the reference group. If a child is growing unusually slowly, a doctor can measure bone age and look for causes.
