Child Growth Percentiles, Explained
After reading this you will know what a growth percentile and z-score actually measure, how to compute them from the WHO standards, and how to tell a normal reading from one that needs a follow-up.
What a growth percentile tells you
A growth percentile compares one child to a reference group of healthy children of the same age and sex. If a 12-month-old boy weighs 9.6 kg and lands on the 50th percentile, half of healthy boys that age weigh less than 9.6 kg and half weigh more. The 50th percentile is the median, not a target every child must hit.
Percentiles run from just above 0 to just below 100. A child on the 3rd percentile is small but still inside the normal range. A child on the 97th is large but also normal. The percentile answers one question: where does this measurement sit among healthy peers? It does not by itself say whether the child is healthy.
The tool covers three measurements for ages 0 to 5 years: weight, length or height, and head circumference. Each has its own WHO curve for each sex, so a 46 cm head at 12 months is scored against head data, not weight data.
When to use it, and when not
Use it to place a single measurement in context, to convert a clinic reading into a number you can track over time, and to read the full percentile table for one age. It is most useful when you record the same measurement every few months and watch the trend.
Do not use it for premature infants without correcting age first, and do not use it past the fifth birthday. The WHO Child Growth Standards stop at 60 completed months. Above that age, different reference charts apply. The tool is also built for individual tracking, not for screening a population.
One reading in isolation can mislead. A child measured just after a growth spurt, or right after an illness, can jump a full band in either direction. The pattern across several visits carries far more information than any single point.
The math: z-scores and the LMS method
The WHO standards are not stored as 100 separate percentile curves. They are stored as three numbers for each age and sex, called L, M and S. This is the LMS method. It lets a skewed distribution (weights are not symmetric) be turned into a standard normal z-score.
Here y is the child's measurement, M is the median for that age and sex, S is the coefficient of variation (roughly the relative spread), and L is the Box-Cox power that corrects for skew. When L = 0 the formula switches to Z = \ln(y/M) / S.
The z-score is the number of standard deviations above or below the median. A z of 0 is exactly the median. A z of +1 is above average, a z of -2 is well below. To turn a z-score into a percentile, feed it through the standard normal cumulative function \Phi:
Some familiar anchors: Z = -1.881 gives the 3rd percentile, Z = 0 gives the 50th, and Z = +1.881 gives the 97th. The WHO flags a z-score below -2 or above +2 as worth a closer look, which corresponds to below the 2.3rd or above the 97.7th percentile.
A worked example with the demo data
Boy, 12 months, weight 9.6 kg
The demo uses a male child, 12 completed months, weighing 9.6 kg. The WHO weight-for-age LMS values for a boy at 12 months are approximately L = -0.1600, M = 9.6479 kg, and S = 0.10958.
- Form the ratio y/M = 9.6 / 9.6479 = 0.99504.
- Raise it to the power L: 0.99504^{-0.16} = 1.000797.
- Subtract 1: 1.000797 - 1 = 0.000797.
- Divide by L \cdot S = -0.16 \times 0.10958 = -0.017533: Z = 0.000797 / -0.017533 = -0.0455.
- Convert to a percentile: \Phi(-0.0455) \times 100 = 48.2.
So the demo child sits on the 48th percentile, a hair below the median. That is textbook average. A weight of 9.6 kg is close to the median 9.65 kg, which is exactly why the z-score is near zero.
You can sanity-check any result. If the entered value equals M, the z-score must be 0 and the percentile 50. If the value is above M, the percentile must be above 50. Any tool that breaks this rule has a bug.
Reading the percentile curve
The chart below shows a boy's weight-for-age band from birth to 24 months. The lines are the 3rd, 50th and 97th percentiles. The demo point (12 months, 9.6 kg) falls almost on the median line.
Notice how the bands spread apart over time. At birth the 3rd-to-97th spread is about 1.8 kg. By 24 months it is about 5.6 kg. That is the coefficient of variation S at work: spread scales with the median. A 0.5 kg difference means a lot at birth and little at two years.
Watch your own z-score move
The single most instructive thing to feel is how much weight it takes to shift a percentile at a given age. Change the weight and watch the z-score and percentile respond.
Common mistakes
- Using age in weeks or years
- The standards are indexed by completed months. A child who is 12 months and 3 weeks is still 12 completed months. Entering 13 shifts the reference median upward and drops the percentile.
- Mixing length and height
- Under 2 years the standard uses length lying down. From 2 years it uses standing height, which is about 0.7 cm shorter. The WHO curves build in this step, so use the measurement method that matches the age.
- Reading a low percentile as a diagnosis
- The 3rd percentile is not a red line. Three in a hundred healthy children sit at or below it by definition. What matters is whether the child holds a band over time.
- Comparing across measurements
- A child can be on the 20th percentile for weight and the 70th for height. That is common and not a contradiction. Each measure has its own distribution.
Related tools
Growth percentiles apply to children under 5. For older bodies, other calculators fit better. The BMI Calculator gives the WHO weight category for adults. The Ideal Weight Calculator lists target ranges by several formulas. For daily energy needs, the BMR & TDEE Calculator estimates resting and total calories, and the Macro Calculator splits those calories into grams. For body shape and clinical scaling, see the Waist-to-Hip & Waist-to-Height Calculator, the Body Fat Estimator and the Body Surface Area Calculator.
Frequently asked questions
What is a good growth percentile for a baby?
There is no single good number. Any percentile from the 3rd to the 97th is within the normal range. A child steadily on the 15th is usually fine. The concern is a large drop or rise across several visits, for example falling from the 50th to the 5th over a few months.
What is the difference between a percentile and a z-score?
They carry the same information in different units. The z-score counts standard deviations from the median: 0 is the median, -2 is low, +2 is high. The percentile is that z-score turned into a 0 to 100 rank. A z of -2 equals the 2.3rd percentile.
Why does the tool ask for sex?
Boys and girls grow along different curves, especially in the first two years. At 12 months the median weight is about 9.65 kg for boys and 8.95 kg for girls, so the same 9.6 kg lands near the 48th percentile for a boy and near the 68th for a girl.
My child dropped from the 60th to the 40th percentile. Is that a problem?
A shift within one band, from the 60th to the 40th, is usually normal variation and measurement noise. Crossing two or more bands, such as the 75th down to the 15th, is the pattern that warrants a check-up. This page is educational and does not replace a clinician.
Can I use this for a premature baby?
Only after correcting the age. Subtract the weeks of prematurity from the calendar age before entering completed months. A baby born 8 weeks early at a calendar age of 4 months has a corrected age of about 2 months.
Related tools
BMI Calculator
Body mass index, WHO category and healthy weight range.
BMR & TDEE Calculator
Resting calories and daily energy needs by activity level.
Dynamic Weight Change Simulator
Realistic weight trajectory with metabolic adaptation.
Ideal Weight Calculator
Devine, Robinson, Miller and Hamwi targets plus BMI range.
Macro Calculator
Protein, carb and fat grams from your daily calories.
Body Fat Estimator (Navy Method)
Body fat percent from tape measurements only.