Consulting

International sampling consultant for household surveys

UN agencies, development banks and national statistics offices bring me in as the sampling expert on a household or population survey. I design the sample, build the weights, and stand behind the precision of the estimates. I spent 20 years at Statistics Canada, Westat, UNICEF and NORC, and now work as an independent international consultant through Samplics LLC, remote or in-country, in English or French.

Taking engagements for Q4 2026/Q1 2027.

Household surveys I've worked on

  • Population-based HIV Impact Assessment (PHIA). Designed the sampling and estimation strategy for national household biomarker surveys in 8 countries: Cameroon, Côte d'Ivoire, Malawi, Namibia, Tanzania, Uganda, Zambia and Zimbabwe. Multi-stage probability samples, with estimates at national and subnational levels.
  • Feed the Future, Haiti (USAID). Designed the sample and estimation for about 3,600 households across three corridors; trained in-country partners and supervised fieldwork.
  • RISE midline, Burkina Faso and Niger. Analysis of a two-stage cluster survey of 100 villages and 2,492 households, including correcting the baseline weights.
  • Canadian Community Health Survey (CCHS). Developed sampling methodology for Statistics Canada's national health surveillance survey.
  • National Crime Victimization Survey (NCVS). Model-based state and large-county crime rates from 15 years of survey data, released by the Bureau of Justice Statistics in 2015.
  • WHO/UNICEF immunization estimates. Led annual production for 195 countries, combining DHS and MICS household-survey microdata with administrative reporting.
  • Reproduced in open code. The published estimates of the Ethiopia DHS, Nigeria's GHS-Panel (LSMS), India's labour force and time use surveys, and NHANES. See the case studies.

Working with UN agencies and development partners

I've seen how these consultancies run from both sides. As UNICEF staff, I led the WHO/UNICEF immunization estimates for 195 countries and co-authored the WHO Vaccination Coverage Cluster Surveys Reference Manual (2018). As a consultant, I've delivered work for:

  • UNSD: sampling and small area estimation training in Senegal, Ethiopia and Vietnam
  • SADC (funded by the World Bank): a capacity assessment of NBS Tanzania
  • the East African Community: training at INSBU Burundi
  • the World Bank: an applied sampling manual and training for Rwanda's national statistics institute (NISR)

I work to your ToR: an inception report, deliverables tied to payment, a methodology annex your reviewers can audit, and code that stays with the national statistics office after the contract ends.

What I can do for your survey

  • Before fieldwork: frame review, stratification, sample size and allocation, and expected precision under realistic response rates.
  • After fieldwork: base weights, nonresponse adjustment, calibration, and standard errors (Taylor or replicate weights), delivered as code you can rerun.
  • When the sample is too thin: small area estimation for district-level or county-level figures.
  • On a deadline: sampling sections for proposals and bids, with me as named personnel.
  • For your team: training in sampling, weighting and small area estimation, in English or French.

Each comes as a fixed-scope engagement with a fixed timeline (see how I work).

Who hires me

National statistics offices, ministries of health, multilateral and bilateral programmes, foundations, research institutes, and survey firms that need a senior sampling statistician for one project rather than a full-time hire.

Common questions

Are you available for international consultancies?
Yes, as an individual consultant or as the named sampling expert on a firm's team. Remote by default, in-country for design workshops, training or fieldwork oversight.
Do you work in francophone countries?
Yes. I work fully in French, including writing reports. Senegal and Burundi are among the national statistics offices I've trained, and I've run small area estimation training for national statistics offices across French-speaking Africa.
Can you join our bid as the sampling statistician?
Yes. I write the sampling and statistical methodology sections and provide named-personnel credentials, typically within 3 to 5 days.
What software do you use?
Python and R. Your team gets documented code it can run, not a spreadsheet of final numbers. If your current process runs on SAS, SPSS or Stata, I can port it and check it against your published tables.