Why "Average" Is the Wrong Yardstick for Your Baby's Milestones

7 min read日本語版あり
Audience
Parents of children 0–1 year
Target length
~1,500 words
Status
Draft v2 (translated from Japanese v2)
Original
../06_milestone_average_violence.md

Lead

"Average age for rolling over: 6 months." Your child is seven months old, and still hasn't rolled over.

The moment you read that line, something happens in your head. My child is behind. The thought returns several times that day. You reopen the book. You scroll social media looking for babies the same age. The thought interrupts you again at work.

This is not parental weakness. It is the strange gravity that the word average exerts. A number meant to summarize a whole population gets applied to a single child standing in front of you, and suddenly it looks like a passing line on an exam.

This article is about why population-level averages are often unhelpful — and sometimes actively harmful — when read against an individual baby, and how to develop language that protects you from comparing without punishing yourself.

Mean, median, and reference range are different things

A quick refresher on what those numbers actually claim.

The mean is the sum of values divided by the number of people. It is pulled by outliers — a few very early or very late achievers.

The median is the value in the middle when the population is lined up. Half of children reach this milestone earlier, half later.

The reference range is usually defined as the central band that excludes the bottom and top 5% (or 1%) of the distribution. It describes a width, not a center. The Denver Developmental Screening Test, the most widely used screening tool for decades, was deliberately built around 25th, 50th, 75th, and 90th percentiles — the design encodes a band, not a point [1].

When you read "average 6 months for rolling over," you are usually reading a median with the width stripped out. What you want to know is where my child falls within the band. With the band hidden, six months looks like a finish line that must be crossed.

Statistics describe populations, not individuals. That is not a cold observation; it is a freeing one.

The WHO study showed how wide that band is

The World Health Organization Multicentre Growth Reference Study followed 816 infants in Ghana, India, Norway, Oman, and the United States longitudinally from 4 to 24 months [2]. Six gross motor milestones — sitting without support, hands-and-knees crawling, standing with assistance, walking with assistance, standing alone, and walking alone — were observed monthly through clinic visits and caregiver reports.

The reported ranges between the 1st and 99th percentile are wider than most parents expect [2]:

Walking alone spans 9.4 months between earliest and latest typical achievers. Standing alone spans 10.0 months [2]. The same study found that roughly 4.3% of infants skip hands-and-knees crawling entirely and proceed directly to other forms of locomotion [2]. Even the order is not universal.

These findings are corroborated outside the WHO study. Adolph's long-running observational work on infant locomotion has repeatedly shown that the timing and even the form of motor milestones vary enormously between children of typical development, and that the developmental sequence is not strictly linear [3].

Compressing this much variation into "average 6 months" makes nine months of healthy variability disappear. Within that hidden width, your child almost certainly falls in the normal range.

"Below average" and "needs evaluation" are not the same

A point worth separating carefully: being behind the average is not the same as needing professional evaluation.

Being below the average is, by definition, being on the slower side of a normal range. It is not in itself a medical concern. The cutoffs that pediatricians use for screening referral live elsewhere — generally at the outer edge of the reference range, not at its center.

The US overhauled its developmental milestone checklists in 2022, switching to "the age by which 75% of children achieve" as the screening threshold [4]. The same revision pushed the listed age for walking from 12 months to 15 months and removed "crawling" from the milestone list altogether, citing the variability and inconsistent definitions documented in the literature [4].

Japanese national data show similar widths. The 2024 release of the Survey on Infant and Child Physical Development by the Cabinet Office's Children and Families Agency reports that the share of infants whose caregivers say they can roll over reaches 90% by 6–7 months of age, and the share able to crawl on hands and knees reaches 90% by 10–11 months [5]. By definition, 10% of typical infants are slower than these marks, and that does not make them outside normal.

The decision to consult a pediatrician should not be triggered by "how far behind the mean" but by "are we outside the reference range," "do other developmental indicators line up," and "does this caregiver feel something is off" [1,4].

"Behind the average, so I'm worried" is the wrong sequence. "I have a concern, so I'll ask" is the better one. Consulting is cheap; consulting unnecessarily costs little. Late consultation can cost; early consultation never does.

Not "stop comparing" — find language that lets you compare without punishing

Parenting advice often reads: "Don't compare your child to other children." The intention is right. In practice, comparison is unavoidable. You see the same-age toddler in the waiting room. A friend posts a video. The neighbor's child crawls past your stroller.

The choice is not whether you will compare. It is whether the comparison will leave a scar.

Several of parents on social media support this concern empirically. In a study of parents of children under five, social comparison on social networking sites that produced negative feelings predicted higher parental distress and lower parenting self-efficacy over time [6]. Earlier work by Coyne and colleagues showed similar associations between maternal social comparison on social networking sites and depressive symptoms, parenting stress, and relationship outcomes [7]. The act of comparing is not what hurts. The unprocessed emotional residue that comparison leaves behind is what hurts.

So instead of trying to stop the comparison reflex, consider rehearsing a few factual sentences to deploy after it happens:

These are not consolations. They are facts. Having a small set of facts memorized means that when you scroll past the neighbor's video at midnight, you have something to push back with. Arguing from information lasts longer than arguing from feeling.

There is also another move available. Shift the comparison target from "other children" to "this child a few months ago." That is a longitudinal comparison instead of a cross-sectional one. Looking back through six months of logs on an app like Memori is, in most cases, far richer reading than looking sideways at the family next door.

"Recording" and "comparing to others" are not the same feature

A small caveat to close on.

Some baby-tracking apps and services display "% of same-age babies who have achieved this." That comparison can be informative in some moments; whether to look at it is a separate decision from whether to record in the first place. Given that exposure to social comparison is independently associated with negative parenting outcomes in the literature [6,7], whether to turn on a feature like this is a legitimate object of self-care.

If you find yourself doing internal math every time a milestone number appears — how many months off is my child? — then taking a break from those displays is a reasonable choice. Keep recording. Close the comparisons. Recording's value lies in your child's own timeline, not in the comparison.

Summary

A line that reads "average 6 months for rolling over" is convenient, but as a ruler for one child its resolution is too coarse. Mean, median, and reference range are different statistical objects, and the threshold for consulting lies at the outer edge of the range or in the caregiver's own intuition — not at the mean [1,4].

You can't will yourself out of comparing. What you can do is keep a few facts on hand for the nights when comparison shows up uninvited. That works longer than "don't worry about it."

The average is a statement about a group. The child in front of you is not a group.

References

  1. Frankenburg WK, Dodds JB. The Denver developmental screening test. J Pediatr. 1967;71(2):181–191. doi:10.1016/S0022-3476(67)80070-2. PMID: 6029467.
  2. WHO Multicentre Growth Reference Study Group; de Onis M. WHO Motor Development Study: windows of achievement for six gross motor development milestones. Acta Paediatr Suppl. 2006;450:86–95. doi:10.1111/j.1651-2227.2006.tb02379.x. PMID: 16817682.
  3. Adolph KE. Learning in the development of infant locomotion. Monogr Soc Res Child Dev. 1997;62(3):I–VI, 1–158. PMID: 9394468.
  4. Zubler JM, Wiggins LD, Macias MM, et al. Evidence-Informed Milestones for Developmental Surveillance Tools. Pediatrics. 2022;149(3):e2021052138. doi:10.1542/peds.2021-052138. PMID: 35132439.
  5. Children and Families Agency, Cabinet Office, Japan. Summary Report of the 2023 Survey on Infant and Child Physical Development. 2024. https://www.cfa.go.jp/policies/boshihoken/r5-nyuuyoujityousa
  6. Glatz T, Daneback K, Sorbring E. Parents' Feelings, Distress, and Self-Efficacy in Response to Social Comparisons on Social Media. J Child Fam Stud. 2023;32:2812–2823. doi:10.1007/s10826-023-02611-2.
  7. Coyne SM, McDaniel BT, Stockdale LA. "Do you dare to compare?" Associations between maternal social comparisons on social networking sites and parenting, mental health, and romantic relationship outcomes. Comput Human Behav. 2017;70:335–340. doi:10.1016/j.chb.2016.12.081.

Related