Average height in the US: measured adult reference

What is the average height in the US?

For US adults 20+ in NHANES August 2021–August 2023, NCHS table 7 publishes mean heights of 175.1 cm for males and 161.2 cm for females; table 8 separately publishes 68.9 and 63.5 inches.

Source: CDC National Center for Health Statistics, Anthropometric Reference Data, tables 7 and 8, printed pages 8–9 (published June 2025; DOI: 10.15620/cdc/174595).

The reference covers the US civilian noninstitutionalized population. These are weighted survey estimates from measured heights, not self-reports or a 2026 survey. Male and female follow the source table labels.

Table 7 publishes P5–P95 cut points of 162.2–187.2 cm for males 20+ and 149.6–172.5 cm for females 20+; these are distribution cut points, not confidence intervals or recommended heights.

Source: NCHS table 7, printed page 8. The interactive reference below preserves table 7 centimeters and table 8 inches separately.

Mean

The weighted arithmetic average.

Median / P50

The middle cut point in the weighted height distribution.

P5–P95

Cut points at the 5th and 95th percentiles; a span describing roughly the middle 90% of the distribution.

Mean SE

Standard error measures uncertainty in the estimated mean. It is not standard deviation or the spread of individual heights.

Showing all adults 20+ · original NCHS table 7 (cm). Unit switching reads the corresponding published table.

Published percentile points Mean · thick segment: P25–P75

Points show published cut points, with a connecting guide rather than a fitted distribution. They do not give an exact percentile for an individual height.

Complete published reference

All 16 sex-by-age rows are shown below in cm. The age selector above changes the chart and summary; it does not hide rows here. Use the unit buttons to read the other original table.

* P95 in both 80+ rows is unavailable because its standard error could not be computed. N counts people examined, not survey weights. All-adult rows already include the age subgroups: do not add them or combine means using these counts.

Source: NCHS report, table 7, printed page 8. Survey: August 2021–August 2023; report published June 2025.

An original arithmetic check

Why published centimeters and inches can disagree

One international inch equals exactly 2.54 cm. But a published number rounded to one decimal place has already lost information. Converting that rounded inch value need not reproduce the separately published centimeter value.

Conversion source: NIST: SI units — length. See also NIST SP 811, Appendix B: conversion factors and rounding. The bound and audit below are our derivation.

Our reproducible audit pairs 174 published statistics from tables 7 and 8. The largest absolute difference after multiplying rounded inches by 2.54 is 0.166 cm; all pairs fall within the conditional 0.177 cm independent-rounding bound.

Audit evidence: all 174 residuals (CSV), methodology and reproduction code. Original inputs: NCHS tables 7–8.

How the 0.177 cm bound works

If one underlying statistic is independently rounded to the nearest 0.1 cm and 0.1 inch, the two rounding uncertainties can add:

0.05 cm + (0.05 in × 2.54 cm/in) = 0.177 cm

Passing this bound is consistent with independent rounding. It does not prove which rounding algorithm NCHS used, validate the original measurements, or recover an unrounded statistic. Three decimal places in the residuals describe the arithmetic, not the precision of a measured height.

What was paired

The audit matches 16 sex-by-age rows and 11 statistics: mean, mean SE and nine percentile cut points. That makes 176 potential pairs; the two unavailable P95 pairs leave 174. These are 174 paired statistics, not 174 people. It uses integer thousandths of a centimeter, with an independent Python Decimal check of the same results.

For example, the largest absolute residual occurs for male age 40–49, P85: 185.0 − (72.9 × 2.54) = -0.166 cm.

Methods, provenance and limits

This is a site-authored reference and reproducible arithmetic audit of published survey summaries. It is not a new measurement study or a peer-reviewed population study. We transcribed the published cells, kept their one-decimal strings and original units, paired matching statistics, and checked the residuals. No NHANES microdata were reanalyzed.

• Population and period matter. Adults 20+ in the United States, measured August 2021–August 2023. This is not a child growth reference or a global average.
• Age groups are cross-sectional. Differences between groups do not isolate aging, generational or causal effects.
• Counts are not weights. N is unweighted, while means, SEs and percentiles are weighted. Do not use N to combine means or add the overlapping 20+ and subgroup rows.
• Cut points have limits. We do not fit a normal distribution or interpolate an exact individual percentile. P5–P95 is not a healthy range, confidence interval or recommendation.
• Missing values stay missing. NCHS withheld P95 for both 80+ groups because its standard error could not be computed; we do not substitute an estimate.

Primary data: CDC/NCHS report, tables 7–8 (PDF pages 14–15, printed pages 8–9). Reference data records the exact source title, DOI, population and field definitions; the manifest records the source PDF checksum. Download the full limitations.

Download the data and reproduce the audit

Save all six v1 files in the same directory, then run node reproduce-v1.mjs with Node.js. The script verifies five file checksums and reproduces every CSV cell. It checks the package and arithmetic; source-table transcription can be checked against the original PDF.

• Reference data ↗All 16 rows, both native units, definitions and source details.height-reference-v1.json
• Rounding audit results ↗174 matched statistics with converted values and residuals.rounding-audit-v1.csv
• Reproduction code ↗Node.js script to verify arithmetic, CSV cells and package hashes.reproduce-v1.mjs
• Methodology ↗Pairing rules, conditional bound and audit summary.methodology-v1.json
• Limitations ↗Population, interpretation and reuse boundaries.limitations-v1.json
• Provenance & checksums ↗Source PDF checksum, file hashes and licensing.manifest-v1.json

NCHS source tables are US public domain. Our audit, derivative files and reproduction code are released under CC0 1.0. Credit NCHS as the original data source and include the survey period, table, unit and statistic when reusing a value. CDC and NIST do not endorse this page.