Afghanistan Third-Party Monitoring, Evaluation and Remote Surveys

Third-party monitoring (TPM) is independent verification, by an organization that neither funds nor implements a program, that assistance reached the intended people in the reported quantity and quality, without diversion or interference. Since 2021, most donors working in Afghanistan have relied on it because their own staff cannot travel there. The World Bank describes the monitoring of the Afghanistan Resilience Trust Fund as its largest third-party monitoring program.

Yes, for most questions donors ask: whether people received assistance, how much, when, whether anyone demanded a fee, and whether women were reached. We verify these through beneficiary call-backs, phone surveys, key informant interviews, partner records, geotagged evidence and satellite imagery, and call a finding verified only when two independent streams agree. Where a question needs eyes on the ground, such as construction quality, we say so and specify the field evidence required.

Women are interviewed by women. Our female interviewers work outside Afghanistan, are matched to the respondent’s language and dialect, and follow a protocol that confirms it is safe to talk before any question is asked. It matters: in September 2025, UNHCR said it could not interview the 52 percent of returnees who are women without female staff.

Every engagement starts with a do-no-harm review. Interviews begin with a safety and consent check, respondents can stop at any time, and sensitive topics run only after a written risk review. Data is encrypted in transit and at rest, identities are stored apart from answers, access is role-based, and a written destruction record closes every engagement. Nothing is stored in Afghanistan or shared with the de facto authorities.

Seven checks run between the call and the dataset. Paradata records each interview’s timing and path; daily high-frequency checks flag outliers, logic breaks, duplicates, speeding and interviewer effects; a second interviewer back-checks at least 10 percent of completed interviews and every flagged one; and consented recordings are audited. Response rates are reported to AAPOR Standard Definitions, and every limitation is stated in the methods note.

Each engagement has a written continuity plan, and the 48-hour nationwide telecommunications shutdown of September 2025 showed why. Fieldwork pauses safely, call schedules and quotas are rebuilt when networks return, and records and satellite imagery continue to be analyzed. We report any gap in coverage rather than filling it with estimates.

We interview in Pashto, Dari and 22 more Afghan languages, including Uzbeki, Turkmeni, Balochi, Pashayi, the Nuristani languages and the Pamir languages, matched to the respondent’s dialect and region. We report in English and, on request, in French, German, Spanish, Italian, Portuguese, Arabic, Turkish, Russian, Chinese, Urdu, Hindi and Hebrew.

Partners, vendors and payment channels are screened against the OFAC, UN 1988 Committee, EU and UK sanctions lists before work begins, and respondent incentives move only through channels cleared in the engagement’s sanctions review. We have no office or staff in Afghanistan and never share data with, request data from, or store data on systems controlled by the de facto authorities. We are not a law firm; your counsel keeps the final word.

Yes, for speed, never for judgment. AI helps flag unusual interviews, transcribe and translate open-ended Pashto and Dari answers, and draft thematic codes, and a trained analyst reviews every output before it counts. Client and respondent data never enter public AI tools and are never used to train models.

Evaluations are designed against the OECD DAC criteria of relevance, coherence, effectiveness, efficiency, impact and sustainability, and built around the UNEG Norms and Standards and Ethical Guidelines, ALNAP guidance on humanitarian evaluation and the Core Humanitarian Standard. Survey work follows ISO 20252:2026, AAPOR Standard Definitions and the ICC/ESOMAR Code; data handling follows the IASC data responsibility guidance and the ICRC Handbook on Data Protection in Humanitarian Action.

Yes. Independent data quality assurance is one of our core services. We re-contact a sample of respondents, test datasets for fabrication and inconsistency, reconcile reported figures against independent evidence, and grade the data on validity, integrity, precision, reliability and timeliness. You receive every finding with the evidence behind it.

With a conversation about the program, the questions your donors or board need answered, and the deadline. We then send a written feasibility note that states what can be verified remotely, what cannot, and at what confidence. We work under direct contracts, long-term agreements and framework call-offs, and every engagement is delivered in-house with one point of accountability.