Between 2018 and 2023, seven Pacific Island countries asked thousands of households the same questions about their children and their women. The answers sit in seven reports and one regional database. This is what they say when you put them side by side, what the reports add that the database does not carry, and what they still cannot say.
Scroll. Every number opens its survey, its year and its source.
A regional picture is usually drawn as if every country were measured on the same day. It was not. Kiribati answered in 2018, Vanuatu and Nauru in 2023. Between the first and the last survey a child born at the time of the first one started school.
Why seven moments? A household survey of this size is run by a National Statistical Office once in a cycle, with partners' funding and technical support, and the Pacific countries joined the sixth MICS round one at a time between 2018 and 2023. The number then travels: from the household to the statistics office, into a survey report, then into the regional database, then into global SDG monitoring, and only then into 2050 reporting. At each step some series stop travelling.
That is the first fact any 2050 reporting line has to carry visibly: a comparison across these seven countries is a comparison of moments, and by the 2027 Outcome Report some of these moments will be nine years old. Hide the year and the picture lies politely.
Counts are from this dataset: series measured by a MICS survey as recorded on PDH.stat, by stage of the pipeline. The report stage is the survey's own tabulation; the regional stage is what PDH.stat carries; the 2050 stage is what the Blue Pacific 2050 dataflow attributes to a MICS source (checked on the same API, 28 Oct 2026: one series, Vanuatu).
Each dot is a survey; the bar shows the number of SDG series for children and women that PDH.stat attributes to it.
Read the surveys as a path. Is the birth registered? At three or four, is the child developmentally on track? Does she complete primary, lower secondary, upper secondary? At fifteen to twenty-four, can she read?
Pick a country, then a second one to compare. Each row counts how many of 1,000 children are there at that stage; switch to girls or boys where the survey published the split. Where the count drops, something between the state and the household is not holding.
Each row is an independent share of 1,000 children at that stage, taken from the published rate. The rows are not multiplied into a cohort: the survey does not follow the same children, so no such model would be honest.
Percent of the relevant age group. "On track" is published for four countries only. Years differ by country (chapter 1).
The same surveys measure three risks: children aged 5 to 17 in economic activity and household chores, children aged 1 to 14 who experienced violent discipline in the last month, and women aged 20 to 24 who were married before 18.
These are not school statistics, but they decide school statistics. Lower is better here.
Three risks, seven countries, as published. A missing bar means PDH.stat carries no MICS-sourced value for that country.
National averages are the least interesting number a survey produces. MICS also tells you who completes school: the poorest fifth of households against the richest fifth, rural against urban, girls against boys.
Each pair of dots is one country at one level of schooling. The longer the line, the more the path depends on where a child was born.
A small model, and only arithmetic. Every wealth quintile is one fifth of households, so of 1,000 children about 200 are in the poorest fifth. The survey tells us how many of them complete each level, and how many of the richest fifth do.
Move the lever: it closes the gap between the two rates by a chosen share. The number that appears is how many more children, per 1,000, would complete that level. It is not a forecast and it says nothing about how. It says how large the prize is, country by country, which is the question a Leaders' meeting asks first.
Extra completions per 1,000 children = 200 × (richest-fifth rate − poorest-fifth rate) × share of gap closed. Rates as published; quintiles assumed to hold one fifth of children each, which is how the survey defines them.
Four questions about women and girls are in the SDG line: marriage before 18, sexual violence before 18, the share of women who make their own informed decisions about their bodies, and the share of people who felt discriminated against.
Only the first is published for all seven countries. The informed-decisions indicator exists for two. That absence is itself a finding for a regional evidence base.
Women's indicators as published on PDH.stat. Grey means not published for that country.
The same seven numbers for every country, read against the median of the seven. Green is better than the median, amber is around it, red is worse. "Better" means more children registered, on track and completing, and fewer at risk.
Cards are what a Forum official reads in the car. The hover on every number is what makes them defensible.
Median of the countries that publish the value. A dash means not published on PDH.stat for that country.
PDH.stat carries 26 MICS-sourced series. The seven Survey Findings Reports carry hundreds. I read the seven (indicator lists, summary tables and key-findings pages) and took fourteen indicators the regional line does not publish: child survival, maternal care, immunisation, nutrition, early education, water, sanitation and child functioning. Each value carries the page it came from.
Then the reconciliation: for the three protection indicators both sources publish, the report value against the PDH.stat value. Mostly they agree to the decimal. Where they do not, the difference is small and explainable (age group, rounding, a later data revision), and that is exactly what a regional register must record rather than hide.
Values as printed in each report or snapshot; "SFR indicator list" is the MICS indicator table at the front of the Survey Findings Report. Blank means I did not find a national value in the pages read, not that the survey did not measure it. Nauru's exclusive breastfeeding is based on 25-49 unweighted cases, as the report marks it.
Twenty-six SDG series. Seven countries. The heat map shows what the regional database holds and what it does not. A grey cell does not always mean the survey did not ask; often it means the number is in a 400-page report and has not been carried into the regional line. Chapter 5c shows fourteen of those numbers, read from the reports.
Below, a small model: choose the indicators a 2027 Outcome Report would need, and see how many country values are missing. Each missing cell is one report extraction, one data-sharing agreement, or one survey module. That arithmetic is the work.
Every value is country data as published by the Pacific Community on PDH.stat (dataflow DF_SDG), filtered to observations whose recorded source is a MICS survey. Nothing is estimated, modelled or imputed. Gaps are shown, not filled.
Limits. Survey years span 2018 to 2023. Samoa's is a combined DHS-MICS. Small populations (Nauru, Tuvalu) carry wider sampling error than point values suggest; the survey reports publish confidence intervals that PDH.stat does not carry. The series set is what PDH.stat attributes to MICS; the surveys measure much more, and reading the Survey Findings Reports is the next step.
Purpose. A working prototype built in October 2026 to show how MICS results for children and women can be made findable, comparable and traceable across the Pacific SDG monitoring line, gaps visible. Personal and non-commercial.