US BirthData
Observed rates from NVSS natality microdata

Why Compare With Similar Babies?

What a birth weight percentile is, why doctors disagree about adjusting it, and what this tool does that earlier charts did not

What a percentile isSame week, not same ageWhat has been done beforeThe argumentDiabetes, blood pressure and smokingThe fatherWhat this tool does differentlyWhat it cannot tell youReferences

Back to the tool Why compare?

The short version

A percentile is a rank, not a grade. Doctors have argued for thirty years about whether that rank should be worked out among all babies or among babies of similar parents. Both sides have good data. This tool does not pick a side. It lets you choose the comparison group yourself, from the full record of United States births, and it always shows you how many babies are in the group you chose.

What a percentile is

Line up 100 babies born in the same week of pregnancy, lightest to heaviest. A baby at the 30th percentile is the 30th in that line: 30 babies weigh less, 70 weigh more. A baby at the 50th percentile, the median, is in the middle. A baby at the 5th percentile is smaller than 95 of the 100.

That is all a percentile says. It does not say whether the baby is healthy. By definition, 10 out of every 100 babies are below the 10th percentile, and most of them are fine. Some babies at the 50th percentile are not. The percentile is one measurement that a doctor or midwife reads together with everything else they know about the pregnancy. This is why the tool asks how your baby's weight compares, and does not ask whether it is normal. There is barely such a thing as a normal weight; there is a place in the line.

Same week, not same age

The tool compares your baby with babies born in the same week. That matters most before 37 weeks. A baby born at 32 weeks is not a random sample of all babies at 32 weeks of pregnancy. Many are born early because something went wrong, and the same problems, high blood pressure above all, also slow a baby's growth. So the weights of babies born at 32 weeks run lighter than the weights of healthy babies still in the womb at 32 weeks.

This is why an ultrasound estimate of weight during pregnancy belongs on a different chart, one built from ultrasound measurements of babies who went on to be born at term.1 If your doctor gave you a percentile from an ultrasound, that number came from such a chart and will not match this tool. The two are answering different questions.

What has been done before

The idea that a baby's weight should be judged against babies of similar parents is not new. It has two families.

Printed curves by race, sex and parity. In 1995 Zhang and Bowes used three years of United States birth certificates to draw separate weight curves for Black and White babies, boys and girls, first babies and later babies.2 Alexander and colleagues published a national reference the next year and later split it by race, Hispanic origin and sex.3,4 The most recent version, by Aris and colleagues in 2019, uses 2017 births and offers curves by sex or by parity, but not both together and nothing else about the mother.5 These were tables. You found the printed curve that matched your baby and read off the percentile.

Customised charts. In 1992 Jason Gardosi in the United Kingdom proposed something different: instead of separate curves, a formula that adjusts the expected weight of each individual baby for the mother's height, her weight at the start of pregnancy, whether she has had a baby before, her ethnic group and the baby's sex.6 The formula, now called GROW, has been fitted to births in more than 30 countries, including the United States in 2009,7 and is built into maternity software across the United Kingdom, Australia, New Zealand and elsewhere.8,9

Ultrasound charts have gone the same way. The largest American ultrasound study, run by the National Institutes of Health, followed 2,334 healthy pregnancies and found that at 39 weeks the middle baby of a White mother was estimated at 3,505 grams, of a Hispanic mother 3,336, of an Asian mother 3,270 and of a Black mother 3,260, and recommended separate charts by race and ethnicity.10 A large international project, INTERGROWTH-21st, reached the opposite conclusion from eight countries: when mothers are healthy and well fed, babies grow so alike everywhere that one chart serves all.11,12 The World Health Organization's own charts sit between the two.13

The argument

Gardosi's case is that a small baby of a small mother is not the same as a small baby of a large mother, and that adjusting for the mother finds more of the babies who are truly not growing well.8

The other side, led by Jennifer Hutcheon, Xun Zhang and Michael Kramer in Canada and Sweden, answers with numbers. In 782,303 Swedish births, the better prediction of stillbirth that customised charts seemed to give came almost entirely from using an ultrasound-based curve for preterm weeks, not from adjusting for the mother; adjusting for her characteristics added almost nothing.14 Their reason is simple. A mother's height, weight and background explain why populations differ, but they explain only a small part of why one baby weighs what it weighs, so the best guess for any baby stays close to the overall average.15 A Scottish study of 979,912 term births found that adjusting for the mother's height and parity did not improve the prediction of stillbirth or infant death at all.16 A Cambridge study found that the stronger links reported for customised charts could be explained by preterm birth and maternal obesity rather than by the adjustment itself.17 A review that pooled 20 studies concluded that both kinds of chart find babies at risk and that neither has been shown to be better.18

The most useful study for a parent asked whether babies of short, thin or first-time mothers are "normally" or "abnormally" small. The answer was: it depends which.

Zhang and colleagues found that the smaller size of babies born to short women and to thin women was harmless: once the chart allowed for the mother's height or build, those babies were no more likely to die than others. But the smaller size of first babies was not harmless: first babies had higher mortality, and a chart that adjusted for parity would have hidden it.19 In other words, some maternal characteristics are physiology and some are not, and a chart cannot tell which by itself.

This is why the tool now offers the mother's height as a menu. It is the one adjustment both sides accept.

Diabetes, high blood pressure and smoking

Every customised chart leaves these out of its formula on purpose. Gardosi's documentation says why: the aim is to describe the weight a baby would reach under good conditions, "not to predict a pathological birthweight".9 Diabetes makes babies heavier; high blood pressure and smoking make them lighter. A chart that adjusted for them would call a baby average when the condition had changed its growth.

This tool lets you choose those conditions anyway, and it is important to understand what you get when you do. If you set diabetes to Yes, the percentile that appears describes your baby among babies whose mothers had diabetes. It is accurate. It is also a comparison with a group whose growth was, on average, altered by the condition. A 4,300 gram baby of a mother with diabetes may sit near the 85th percentile of that group and above the 95th percentile of all babies. For that reason the tool shows a red notice whenever one of these menus is set, and keeps the comparison with all babies on the screen beside it.

The difference has a name. A reference describes what happened. A standard describes what should happen under good conditions.20 This tool is a reference.

The father

No published chart and no calculator I could find uses the father's race or ethnicity. It has been studied only as a risk factor. In 1983 United States births, babies of a White mother and a Black father, and of a Black mother and a White father, weighed less on average than babies of two White parents, though the mother's race carried most of the difference.21 A later national study found that, once other family characteristics were taken into account, the father's race mattered much less than the mother's, and that the babies with the worst outcomes were those whose father was not named on the birth certificate at all.22 A 2021 study of four million births confirmed that early birth is more common when either parent is Black.23

The tool offers the father's race and ethnicity as a menu because it is on the certificate and because parents ask. It keeps "unknown or not stated" as a choice of its own, for the reason that study gave: those births are a different group, and dropping them would hide it. About one certificate in six has no father's race recorded.

What this tool does differently

The public calculators available today take fewer inputs than this one. A British calculator built from 2013 to 2014 births takes only the baby's sex and week.24 The GROW calculator takes the five Gardosi variables and returns a modelled centile.9 The NIH charts are by race and ethnicity alone.10 None of them lets you set the mother's age, height, BMI, parity, the baby's sex, both parents' race and ethnicity, and diabetes, high blood pressure and smoking at the same time, and none is built on every birth in the country.

Two other differences matter more than the number of menus. First, this tool does not model anything. Every percentile is a count of babies who were actually born with the characteristics you chose, from all 32 million singleton births in the United States from 2016 to 2024. Second, it always tells you how many. Narrowing the comparison makes it more like your pregnancy and smaller at the same time, and a percentile built on 40 babies means less than one built on 40,000. When a week has fewer than 30 matching babies the tool says so and shows nothing, rather than a number that looks precise and is not.

What it cannot tell you

It cannot tell you whether your baby is healthy. It cannot be used for an ultrasound estimate during pregnancy. It cannot say which of the differences between groups are physiology and which are the results of poor care, poverty or discrimination; the birth certificate records the outcome, not the cause. And it describes babies born in 2016 to 2024; average birth weight in the United States has been drifting down for two decades, so an older chart would place the same baby a few percentile points higher.5

If a number on this page worries you, the person to ask is the doctor or midwife who knows the pregnancy. They will read the weight together with the reason your baby was born when it was, the baby's length and head size, the placenta, and how the baby is doing. The percentile is where the conversation starts, not where it ends.

Back to the tool

References

  1. Hadlock FP, Harrist RB, Martinez-Poyer J. In utero analysis of fetal growth: a sonographic weight standard. Radiology. 1991;181(1):129–133. doi:10.1148/radiology.181.1.1887021
  2. Zhang J, Bowes WA Jr. Birth-weight-for-gestational-age patterns by race, sex, and parity in the United States population. Obstet Gynecol. 1995;86(2):200–208. doi:10.1016/0029-7844(95)00142-e
  3. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol. 1996;87(2):163–168. doi:10.1016/0029-7844(95)00386-X
  4. Alexander GR, Kogan MD, Himes JH. 1994–1996 U.S. singleton birth weight percentiles for gestational age by race, Hispanic origin, and gender. Matern Child Health J. 1999;3(4):225–231. doi:10.1023/a:1022381506823
  5. Aris IM, Kleinman KP, Belfort MB, Kaimal A, Oken E. A 2017 US reference for singleton birth weight percentiles using obstetric estimates of gestation. Pediatrics. 2019;144(1):e20190076. doi:10.1542/peds.2019-0076
  6. Gardosi J, Chang A, Kalyan B, Sahota D, Symonds EM. Customised antenatal growth charts. Lancet. 1992;339(8788):283–287. doi:10.1016/0140-6736(92)91342-6
  7. Gardosi J, Francis A. A customized standard to assess fetal growth in a US population. Am J Obstet Gynecol. 2009;201(1):25.e1–25.e7. doi:10.1016/j.ajog.2009.04.035
  8. Gardosi J, Francis A, Turner S, Williams M. Customized growth charts: rationale, validation and clinical benefits. Am J Obstet Gynecol. 2018;218(2S):S609–S618. doi:10.1016/j.ajog.2017.12.011
  9. Perinatal Institute / Gestation Network. GROW 1.5 customised centile calculator: documentation. Birmingham: Perinatal Institute; updated December 2020. Available from: https://www.gestation.net/GROW_documentation.pdf
  10. Buck Louis GM, Grewal J, Albert PS, Sciscione A, Wing DA, Grobman WA, et al. Racial/ethnic standards for fetal growth: the NICHD Fetal Growth Studies. Am J Obstet Gynecol. 2015;213(4):449.e1–449.e41. doi:10.1016/j.ajog.2015.08.032
  11. Villar J, Cheikh Ismail L, Victora CG, Ohuma EO, Bertino E, Altman DG, et al. International standards for newborn weight, length, and head circumference by gestational age and sex: the Newborn Cross-Sectional Study of the INTERGROWTH-21st Project. Lancet. 2014;384(9946):857–868. doi:10.1016/S0140-6736(14)60932-6
  12. Villar J, Papageorghiou AT, Pang R, Ohuma EO, Cheikh Ismail L, Barros FC, et al. The likeness of fetal growth and newborn size across non-isolated populations in the INTERGROWTH-21st Project. Lancet Diabetes Endocrinol. 2014;2(10):781–792. doi:10.1016/S2213-8587(14)70121-4
  13. Kiserud T, Piaggio G, Carroli G, Widmer M, Carvalho J, Neerup Jensen L, et al. The World Health Organization fetal growth charts: a multinational longitudinal study of ultrasound biometric measurements and estimated fetal weight. PLoS Med. 2017;14(1):e1002220. doi:10.1371/journal.pmed.1002220
  14. Hutcheon JA, Zhang X, Cnattingius S, Kramer MS, Platt RW. Customised birthweight percentiles: does adjusting for maternal characteristics matter? BJOG. 2008;115(11):1397–1404. doi:10.1111/j.1471-0528.2008.01870.x
  15. Hutcheon JA, Zhang X, Platt RW, Cnattingius S, Kramer MS. The case against customised birthweight standards. Paediatr Perinat Epidemiol. 2011;25(1):11–16. doi:10.1111/j.1365-3016.2010.01155.x
  16. Iliodromiti S, Mackay DF, Smith GC, Pell JP, Sattar N, Lawlor DA, et al. Customised and noncustomised birth weight centiles and prediction of stillbirth and infant mortality and morbidity: a cohort study of 979,912 term singleton pregnancies in Scotland. PLoS Med. 2017;14(1):e1002228. doi:10.1371/journal.pmed.1002228
  17. Sovio U, Smith GCS. The effect of customization and use of a fetal growth standard on the association between birthweight percentile and adverse perinatal outcome. Am J Obstet Gynecol. 2018;218(2S):S738–S744. doi:10.1016/j.ajog.2017.11.563
  18. Chiossi G, Pedroza C, Costantine MM, Truong VTT, Gargano G, Saade GR. Customized vs population-based growth charts to identify neonates at risk of adverse outcome: systematic review and Bayesian meta-analysis of observational studies. Ultrasound Obstet Gynecol. 2017;50(2):156–166. doi:10.1002/uog.17381
  19. Zhang X, Cnattingius S, Platt RW, Joseph KS, Kramer MS. Are babies born to short, primiparous, or thin mothers "normally" or "abnormally" small? J Pediatr. 2007;150(6):603–607. doi:10.1016/j.jpeds.2007.01.048
  20. Grantz KL, Hediger ML, Liu D, Buck Louis GM. Fetal growth standards: the NICHD fetal growth study approach in context with INTERGROWTH-21st and the World Health Organization Multicentre Growth Reference Study. Am J Obstet Gynecol. 2018;218(2S):S641–S655.e28. doi:10.1016/j.ajog.2017.11.593
  21. Migone A, Emanuel I, Mueller B, Daling J, Little RE. Gestational duration and birthweight in white, black and mixed-race babies. Paediatr Perinat Epidemiol. 1991;5(4):378–391. doi:10.1111/j.1365-3016.1991.tb00724.x
  22. Ma S. Paternal race/ethnicity and birth outcomes. Am J Public Health. 2008;98(12):2285–2292. doi:10.2105/AJPH.2007.117127
  23. Green CA, Johnson JD, Vladutiu CJ, Manuck TA. The association between maternal and paternal race and preterm birth. Am J Obstet Gynecol MFM. 2021;3(4):100353. doi:10.1016/j.ajogmf.2021.100353
  24. University of Leicester, TIMMS. Birth weight centiles calculators (MBRRACE-UK, births 2013 to 2014). Available from: https://timms.le.ac.uk/birth-weight-centiles/

All journal references checked against PubMed on 4 September 2026. Reference 5 carries an article number in place of pages, as the journal publishes it. References 9 and 24 are web documents.