A growth percentile ranks one measurement against children of the same age and sex, so the answer depends on which children the chart was built from. A boy of exactly 12 months at the WHO 2006 median weight, 9.65 kg, is the 50th percentile on WHO's chart. On the US CDC 2000 chart he is the 23rd, by our arithmetic, read off the row CDC uses for every baby from 12.0 to 12.99 months; interpolating CDC's rows to his exact age makes him the 27th. CDC itself has recommended WHO's charts for children under 24 months since 2010. Singapore's Health Booklet uses neither: the file HealthHub serves today prints charts from a 2000 survey of Singapore preschool children. On any of these charts, a percentile is a position among other children rather than a diagnosis.
CDC's chart ranks WHO's median baby lower at a year old
The CDC report that set US practice, MMWR 59(RR-9) of September 2010, calls WHO's charts "growth standards, describing the growth of healthy children in optimal conditions" and CDC's "a growth reference, describing how certain children grew in a particular place and time". WHO's Multicentre Growth Reference Study measured about 8,500 children in six countries between 1997 and 2003, under conditions WHO specified.
CDC's data files list each month at its half-month point, so the 1.5 row covers babies from 1.0 to 1.99 months. On that convention, CDC's median boy for 12.0 to 12.99 months is 0.81 kg heavier than WHO's at 12 months. Interpolated to exactly 12.0 months, the gap is 0.66 kg for boys and 0.57 kg for girls. Interpolating the same way, our computation puts a baby on WHO's median weight here on CDC 2000:
| Age (months) | 3 | 6 | 9 | 12 | 18 | 24 |
|---|---|---|---|---|---|---|
| Boys | 67th | 52nd | 36th | 27th | 26th | 35th |
| Girls | 67th | 54th | 38th | 28th | 25th | 32nd |
The shift reaches the tails. WHO's −2 SD weight line for boys, the 2.3rd percentile on its own chart, falls at the 0.4th on CDC's at 12 months. The consequence, as MMWR noted, is that "fewer U.S. children will be identified as underweight using the WHO charts."
Z-scores grade a child's distance from the median
For each age and sex WHO publishes three numbers: L for the skew of the distribution, M for the median and S for its spread. A measurement X becomes a z-score through z = ((X/M)^L − 1) / (L·S), the formula in WHO's own anthro code. Read off the normal curve, the z-score gives the percentile: two scales for the same position. A z-score of 0 is the 50th percentile and −2 the 2.3rd; on our arithmetic, −3 is the 0.13th, where percentiles crowd together and z-scores stay easy to tell apart.
WHO counts populations on the z-score scale. Its Global Health Observatory defines stunting prevalence by height-for-age more than 2 standard deviations below the WHO median, among children under five. Malaysia's Ministry of Health names the weight-for-age band below −3 SD "Kurang berat badan teruk" (severely underweight) and the height-for-age band from −3 to −2 SD "Bantut" (stunted). These are categories for counting children across a population. A percentile describes a position and diagnoses nothing; a child's own clinic interprets growth over time.
The Singapore booklet we could open scores children on local charts
The Health Booklet HealthHub serves today is the Health Promotion Board's file health-booklet-2014.pdf, created in December 2014. Its charts from birth to six years, for weight, length or height, head circumference and BMI-for-age, are each credited to the "Anthropometric Study on Pre-School Children in Singapore, 2000" by the National Healthcare Group Polyclinics. Singapore paediatricians writing in the Annals, Academy of Medicine, Singapore in 2018 recommended the HPB charts. HealthHub's children's health e-service FAQ still explains readings against "our local growth charts", naming no survey year. Whether practice changed after 2014, we could not verify.
In Malaysia, a Ministry of Health guideline for private practitioners, file dated February 2025 and hosted by the Malaysian Medical Association, lists among its purposes "Penggunaan carta pertumbuhan Pertubuhan Kesihatan Sedunia (WHO) 2006", the use of WHO's 2006 charts, and names them as the charts in the child health record book. Hong Kong's Department of Health announced in April 2024 that its Maternal and Child Health Centres would move from HK1993 to the local HK2020 charts for children up to five, with deployment expected that July, although its survey found young Hong Kong children "fitted the WHO2006 reasonably well". Taiwan's Health Promotion Administration has used WHO's standard for children under five as its reference since 18 May 2009. Those are the four places we checked.
A month-by-month table scores a typical newborn below the median
WHO's instructions say its tables for constructing curves "are in days for age-based indicators", and WHO's software looks a child up by the whole day. WHO's daily median weight for boys dips just after birth and is 3.49 kg by day 7. A straight line from the birth row to the one-month row cannot follow the dip: on our arithmetic it puts the day-7 median near 3.6 kg, so a baby exactly on WHO's median looks lighter than average.
Until 7 October 2026, our baby growth percentile calculator drew that straight line through WHO's monthly table. In our check against WHO's daily tables, a baby on the median weight read as low as the 41st percentile in its second week. It read more than 5 points low on every day from day 1 to day 19, and 9.4 points off at worst. It also had no lying or standing input. The version prepared on 7 October still scores from WHO's monthly rows, calls results under three months approximate, points to WHO Anthro, which works by the day, and adds the position choice. The live page may show the earlier version until it is deployed.
Measured lying down, a two-year-old ranks several percentile points taller
WHO measures length lying down before the second birthday and height standing from it. WHO's anthro code adds 0.7 cm to a standing measurement before 731 days and subtracts 0.7 cm from a lying one after, and flags a standing measurement under 9 months. MMWR gives the difference as 0.7 to 0.8 cm, Malaysia's guideline applies 0.7 cm both ways, and the Singapore commentary puts standing height "about 0.7 cm less".
On our arithmetic with WHO's tables, leaving the correction out moves a median child by about 9 to 12 percentile points between 9 months and the second birthday. At 24 months, an uncorrected lying measurement ranks a median boy about nine points higher.
Screening lines rank children by convention rather than by outcome
When CDC recommended WHO's charts for US children under two, MMWR set screening at the 2.3rd and 97.7th percentiles, ±2 standard deviations, in place of the 5th and 95th used with CDC's charts. It called the old pair "arbitrary statistical values" and said the WHO percentiles "also are arbitrary and not based on health outcomes". The reason it gave was a count. WHO's lines on WHO's curves would identify about as many US children for follow-up, for short stature and overweight, as CDC's 5th and 95th had on CDC's curves.
View author profile → · Editorial policy
