Article summary
Label source weight, unknowns and coverage limits in a multi-source Mauritania +222 WhatsApp Gender and Age report.
If most of a Mauritania +222 profile sample comes from one partner, the result first describes that partner list, not a national population. Sample composition belongs before conclusions.
Show provenance weight first
| Source | Submitted | Format accepted | Profile usable |
|---|---|---|---|
| First-party forms | Count | Count | Count |
| Partner | Count | Count | Count |
| Legacy CRM | Count | Count | Count |
Do not invent statistical weighting
Without a reliable population distribution, do not multiply a partner sample by scientific-looking weights. A clear coverage limitation is better than fabricated representativeness.
Divide work from the Full Format page
The Full Format article governs field-family acceptance. This page addresses whether profile output can be aggregated and how it should be interpreted.
+222 is not a population sampling frame
A numbering plan does not prove residence, citizenship or representativeness. Foreign phones can also belong to real service customers.
Publish unknown by provenance
Calculate age, gender, format and task unknowns separately so one source’s missingness is not hidden in an average.
Combine small cells first
Stop splitting when source, region, age and gender leave too few records. Do not publish recognisable combinations.
Use controlled aggregation only
Do not use output for identity, employment, credit, insurance, minor determination or differential pricing. Person-level results stay out of ordinary sales tables.
Improve the next sample
Add provenance diversity, collection period and self-reported market rather than merely increasing one partner’s volume.
The Mauritania conclusion
The report may state what was observed in this list; without representativeness evidence, it cannot say what Mauritania users are like.
