Praxis Data Lab · Validation Record 002

Subgroups, variance, and uncertainty.

This benchmark tests sex-specific survey domains, design-based standard errors, derived 95% confidence intervals, and the comparison between men and women.

Authoritative domain comparison

Depression prevalence by sex among U.S. adults aged 20 and over, 2013–2016

Men

Within tolerance

NCHS estimate5.5%

SE 0.4 pp

Praxis result5.5%

SE 0.4 pp

Praxis 95% CI: 4.71%–6.39%

Full-precision estimate difference: 0.049 pp · SE difference: 0.029 pp

Women

Within tolerance

NCHS estimate10.4%

SE 0.6 pp

Praxis result10.4%

SE 0.6 pp

Praxis 95% CI: 9.32%–11.54%

Full-precision estimate difference: 0.028 pp · SE difference: 0.034 pp

Design-based comparisonMen versus women

Two-sided p-value< 0.001

NCHS reports that women had a significantly higher prevalence than men. Praxis reruns that comparison inside the complex survey design.

NCHS publishes both percentages and standard errors to one decimal place; ±0.05 percentage points is the rounding-equivalent acceptance boundary for each value.

Reproducible method

What Record 002 adds

Population: U.S. civilian noninstitutionalized adults aged 20 and over with a valid PHQ-9 depression score, evaluated as sex-specific survey domains
Domain definition: RIAGENDR categories 1 (Men) and 2 (Women), retained inside the full survey design
Survey cycles: 2013–2014 and 2015–2016
Variables: SEQN, RIAGENDR, RIDAGEYR, DPQ010–DPQ090, WTMEC2YR, SDMVSTRA, SDMVPSU
Variance: Taylor-series survey estimation using SDMVSTRA and SDMVPSU with nesting enabled
Confidence interval: Two-sided 95% Wald confidence interval derived from the full-precision survey estimate and standard error
Comparison: Design-based t test comparing men and women
Weight: WTMEC2YR divided by 2 to create a four-year MEC examination weight

Interpretation boundary

A second pathway—not blanket validation.

This record validates the documented sex-domain and design-based variance pathway for this analysis. It does not certify every possible subgroup, outcome, or NHANES variable.

This independent Praxis reproduction is supported by the cited public sources, documented method, declared tolerance, and machine-readable test.

Review all Validation Records