About these summaries
Every chart on this page is drawn from group summaries: bins of about 160 children, fixed percentiles, histogram counts and test statistics. No individual records are published; the PHN data can be downloaded from the network to reproduce them. Diagnostics generated —.
z-scores against body size:
Each panel splits the reference children into 20 equal-sized groups (about 160 each) along one variable. Each shaded block spans its group's range on that variable and shows the middle 50% and 80% of z; wide blocks mean sparse data, such as the heaviest children. Dots are group means at the group median, and the dark line is a smoothed trend (1st to 99th percentile) with ±2 SE. For a well-fitting model the trend stays near 0 and the blocks stay level near ±0.67 and ±1.28. Hover a group for its numbers.
By sex
| Sex | Children | Mean z | ± 2 SE | SD | Above +2 | Below −2 |
|---|
Q-tests by height group:
Royston–Wright tests, the standard check for growth references. Within each height group, Z1 tests the mean (median model), Z2 the variance (σ), Z3 the skewness (λ) and Z4 the kurtosis. Values beyond ±2 are highlighted, as are overall p-values below 0.05. The Z1 test subtracts the degrees of freedom of the whole median model, including the weight-for-height and age terms, so its p-value is conservative.
| Height group | Z1 mean | Z2 variance | Z3 skewness | Z4 kurtosis | Children |
|---|
Distribution of z:
Tail rates
| Cutoff | Expected each side | Observed above | Observed below | Children: expected / above / below |
|---|
Worm plots by height group:
A worm plot is a Q-Q plot with the diagonal removed, drawn separately for six height groups. A flat worm inside the grey band means the median, spread and shape are all right for children of that size. A worm that tilts suggests the spread is off; a U or S shape suggests skewness or tail weight. Points outside the band are marked in red.
Shape comparison
Difference in AIC from the fitted model (lower is better; a difference above about 10 is meaningful). Fixing λ at 0, 1 or 2 forces log-normal, normal or area-scale tails. BCT adds a parameter for tail weight.
| Level | Fitted λ | Fitted | λ = 0 | λ = 1 | λ = 2 | BCT |
|---|
How diameter scales with height
The local height exponent, d ln(diameter) / d ln(height), from a smooth fit, with ±2 SE bands. An exponent of 1 means the aorta grows in proportion to height; adult data are often closer to 0.5. A single fixed ratio index assumes one exponent at every size.