Child Height Predictor

Adult height is one of the most heritable human traits, and it is still only about 80% genetic. That last fifth is what makes prediction imprecise.

The mid-parental method

The standard calculation, sometimes called the Tanner method:

For a boy: (father's height + mother's height + 13) ÷ 2

For a girl: (father's height + mother's height − 13) ÷ 2

The 13 cm is the average difference in adult height between men and women. Adding half of it for a boy and subtracting half for a girl is the same as taking the mid-parental average and adjusting by 6.5 cm.

For parents of 175 cm and 162 cm, the mid-parental height is 168.5 cm:

The range is the important part

The prediction carries roughly ±8.5 cm, which is a span of 17 cm — more than half a foot.

That is not a weakness of the method so much as an honest statement about how variable the outcome is. A boy predicted at 175 cm could reasonably reach anywhere from 166.5 to 183.5, and both ends of that range are entirely normal for those parents.

Two things follow. A single-number prediction is misleading wherever it is presented without the range. And a child who ends up 5 cm from the prediction has not deviated from anything — they are well inside it.

Siblings illustrate this directly. Two children of the same parents frequently differ by 10 cm as adults, because each inherits a different combination of the many hundreds of gene variants involved.

Where the other 20% comes from

Nutrition, especially in the first two years. This is the largest single environmental factor and the window is narrow. Protein, calcium, zinc, iron and vitamin D all matter, and chronic undernutrition in the first thousand days produces stunting that is not fully recovered later.

Sleep. Growth hormone is released predominantly during deep sleep. Chronic sleep restriction in children has measurable effects on growth.

Illness. Repeated infection, untreated coeliac disease, cow's milk protein allergy and chronic conditions all divert energy from growth.

Hormones. Growth hormone deficiency and hypothyroidism both reduce growth and are treatable — which is the strongest argument for tracking growth rather than assuming.

Physical activity. Supports growth generally. The belief that certain sports make children taller has it backwards: basketball and volleyball select tall children rather than producing them.

Some things widely believed to affect height do not. Weightlifting with proper form does not stunt growth. Coffee does not. Neither does sitting badly, though posture affects measured height by a centimetre or two.

The generational change

Indian average adult height has risen by several centimetres over the last few generations — a change far too rapid to be genetic.

It reflects better childhood nutrition, fewer infections, and improved maternal health. The same pattern appeared in Japan, South Korea and much of Europe during their periods of rapid development, and it is the clearest available demonstration that the environmental fifth is real.

The practical implication for the calculation: a prediction based on parents raised in poorer conditions will tend to underestimate, because the child's environment is better than the parents' was. Children of migrants to wealthier countries frequently exceed the mid-parental prediction for exactly this reason.

Other prediction methods

Bone age is the most accurate. An X-ray of the left hand and wrist shows how much growth plate cartilage remains, and combined with current height it predicts adult height to within about 2 to 3 cm. It is a medical procedure done for a clinical reason rather than curiosity.

The Khamis-Roche method uses the child's current height and weight along with the parents', and is more accurate than mid-parental alone without needing an X-ray.

Doubling the height at age two is a folk method for boys, and multiplying a girl's height at 18 months by two. It is a rough approximation with an error of ±10 cm or more.

Growth curve extrapolation — following a child's established percentile forward — is what paediatricians actually use in practice, and for a child growing steadily it is as good as anything.

What growth looks like

Growth plates close at around 15 to 16 in girls and 16 to 18 in boys, at which point height is final. Late developers are frequently the tallest in the end, because they spend longer growing before the plates fuse — the boy who is shortest at 14 has often overtaken half his class by 18.

When to raise it with a doctor

The level matters much less than the pattern.

Crossing two major percentile lines downward on the height chart.

Growing less than 4 cm a year between age three and puberty.

Height well below the mid-parental prediction, outside the range rather than merely below the midpoint.

No signs of puberty by 13 in a girl or 14 in a boy.

Height below the 3rd percentile, or a large mismatch between the height and weight percentiles.

Most short children are healthy children of short parents, and the great majority of referrals find nothing. The point of raising it is the small number where growth hormone deficiency, hypothyroidism, coeliac disease or a chronic condition is found — all of which are treatable, and all of which are treated more effectively the earlier they are identified.

Why the prediction is imprecise

Height is polygenic — influenced by many hundreds of gene variants, each contributing a fraction of a centimetre. That is why it varies continuously rather than falling into categories, and why two siblings can differ so much.

A child inherits half their variants from each parent, and which half is chance. Two children of the same parents receive different combinations, so the mid-parental average predicts the centre of the distribution of possible outcomes rather than any particular child's place in it.

The parents' own heights also reflect their environment, not only their genes. A parent who was undernourished as a child is shorter than their genes would have produced, so the mid-parental figure understates what their well-fed child will reach. This is precisely the mechanism behind the generational increase, and it means the prediction is systematically low in a family whose circumstances have improved.

Genome-wide studies now explain about 40% of the variance in adult height from common variants directly — impressive as science, and still less accurate than asking how tall the parents are.

Measuring a child properly

Small errors matter when the numbers are being tracked over time.

Under two years, measure lying down on a length board, with the head against the fixed end and the legs gently straightened. After two, measure standing, heels together, back against a wall, heels, buttocks and shoulders touching it, looking straight ahead.

The two methods differ by about 0.7 cm, so a switch between them looks like a change if it is not noted.

Same time of day, since spinal compression makes a person up to 1 cm shorter by evening. No shoes. A flat object placed on the head at right angles to the wall, rather than a hand, which tilts.

Measure every three to six months. More often than that measures noise, and the noise is what produces unnecessary worry.

What this calculator assumes

• The mid-parental method with a ±8.5 cm range, which is the standard clinical approximation.
• Parents' heights measured rather than reported. Self-reported heights are systematically overstated, which biases the prediction upward.
• Both biological parents' heights. The method does not work with one.
• Typical growth and nutrition. A child with a chronic condition or significant undernutrition will not follow it.
• This is an estimate with a wide range and is not a clinical assessment. Bone age is what a paediatric endocrinologist would use.