A citation-grade reference on monitoring and evaluation practice — evaluation designs, indicators, data quality, MEAL systems and standards — published and maintained by Sibasi Ltd.https://monival.orgenhttps://monival.org/approaches/utilization-focused-evaluationhttps://monival.org/approaches/utilization-focused-evaluationUtilization-Focused Evaluation (UFE) is Michael Quinn Patton's approach built on a single premise: an evaluation should be judged by its usefulness to the specific people who will act on it. Every design decision — questions, methods, analysis, reporting — is negotiated with named primary intended users, for explicitly agreed intended uses, from the first day of the engagement.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/approaches/developmental-evaluationhttps://monival.org/approaches/developmental-evaluationDevelopmental evaluation (DE) is Michael Quinn Patton's approach for supporting the development of innovations in complex, uncertain environments. Instead of judging a stable programme against fixed objectives, the developmental evaluator works inside the innovating team, feeding data and evaluative thinking into decisions as the intervention itself is still being invented. DE is a distinct purpose, not formative evaluation stretched out — and not a licence to skip rigour.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/approaches/participatory-and-empowerment-evaluationhttps://monival.org/approaches/participatory-and-empowerment-evaluationParticipatory evaluation shifts control of the evaluation process toward the people the programme concerns; empowerment evaluation goes furthest, making self-determination the goal and the evaluator a critical friend. The family spans two distinct streams — practical participation for better decisions, transformative participation for shifted power — and it carries a genuine, unresolved rigour debate that this page presents with both sides cited.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/approaches/feminist-and-gender-responsive-evaluationhttps://monival.org/approaches/feminist-and-gender-responsive-evaluationGender-responsive evaluation assesses how an intervention affects gender and power relations, and conducts the evaluation itself in a way that is inclusive and fair; feminist evaluation goes further, treating inequity as structural and the evaluation as an instrument of change. Sex-disaggregated data is the floor of this practice, not its content — the defining move is analysing power, and being explicit about whose knowledge counts.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/approaches/most-significant-changehttps://monival.org/approaches/most-significant-changeMost Significant Change is a participatory monitoring technique in which stories of significant change are collected from the field and then systematically selected upward through the organisation, with the reasons for each selection recorded and fed back. The analytic engine is the selection deliberation — it forces explicit argument about what counts as valuable change. MSC finds and interrogates significance; it does not measure prevalence, and using it as if it did is the technique's standard abuse.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/approaches/outcome-mappinghttps://monival.org/approaches/outcome-mappingOutcome Mapping, developed at Canada's International Development Research Centre, reframes outcomes as changes in the behaviour, relationships, activities and actions of the boundary partners a programme works with directly — the actors it can influence but not control. Its three stages and twelve steps carry a programme from intentional design through monitoring journals to evaluation planning, and its graduated progress markers make influence observable long before impact is measurable.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/data/quality-dimensionshttps://monival.org/data/quality-dimensionsData quality in M&E is not one property but several: validity, reliability, integrity, precision and timeliness in the USAID tradition; completeness, timeliness and consistency in the WHO framing. Each dimension fails in a characteristic way, and each failure has a characteristic remedy — almost all of them applied at design time, not audit time.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/data/quality-assessmenthttps://monival.org/data/quality-assessmentA data quality assessment (DQA) is a structured exercise that checks whether reported results can be trusted: it verifies reported numbers against source records and appraises the data-management system that produced them. The standard toolkits assess a common core of dimensions — completeness, timeliness, internal and external consistency, and accuracy — and end in a costed action plan, not just a score.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/data/samplinghttps://monival.org/data/samplingThe sampling design decides what a survey can claim before a single interview happens. Probability designs — simple random, systematic, stratified, multi-stage cluster — support statements about a population; purposive and quota designs support statements about the cases selected. Sample size follows from precision, variance and disaggregation demands, and clustered designs need to be larger than the textbook formula suggests.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/data/questionnaire-designhttps://monival.org/data/questionnaire-designMost measurement error in surveys enters between the indicator definition and the question a respondent actually hears. Questionnaire design is the discipline of closing that gap: precise wording, deliberate recall periods, response formats that match the analysis plan, careful translation, and a pretest-pilot-lock sequence that treats the instrument as a versioned artefact.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/data/qualitative-methodshttps://monival.org/data/qualitative-methodsKey informant interviews, focus group discussions and observation each answer a different kind of question, and each produces analysable evidence only when selection, facilitation and documentation are as disciplined as any survey. The craft runs from purposive sampling and guide design through recording and transcription to coding and a defensible stopping rule.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/data/mixed-methodshttps://monival.org/data/mixed-methodsA mixed-methods evaluation combines quantitative and qualitative strands so that each covers the other's blind spots: numbers establish magnitude and pattern, qualitative work explains mechanism and meaning. The design choices are few — sequential explanatory, sequential exploratory, or convergent — and the discipline that separates genuine mixing from two stapled studies is integration: planned points where the strands must meet.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/indicators/types-and-levelshttps://monival.org/indicators/types-and-levelsAn indicator is a variable that provides a simple, reliable means to measure achievement or change. Indicators sit at different levels of the results chain — input, activity, output, outcome, impact — and come in distinct types whose arithmetic differs. Confusing the levels, or summing values that cannot be summed, produces numbers that look precise and mean nothing.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/indicators/quality-criteriahttps://monival.org/indicators/quality-criteriaSMART, CREAM and SPICED are competing answers to the same question: how do you tell a usable indicator from a plausible-sounding one? They embody genuinely different philosophies of measurement — SMART and CREAM privilege objectivity and verifiability, SPICED privileges the perspective of the people whose change is being measured — and the disagreement between them is methodological, not cosmetic.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/indicators/reference-sheetshttps://monival.org/indicators/reference-sheetsA performance indicator reference sheet (PIRS) is the governed document that fixes everything an indicator's number depends on: the precise definition, unit of measure, disaggregation, data source, collection method, frequency, responsible party and known limitations. Without one, an indicator's meaning lives in the heads of whoever currently reports it — and drifts silently every time those heads change.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/indicators/proxy-and-compositehttps://monival.org/indicators/proxy-and-compositeA proxy indicator measures something observable that stands in for a result that is unobservable, too slow or too costly to measure directly; a composite index bundles several indicators into one number. Both are legitimate and both are dangerous in the same way: the construction choices disappear into a value that looks like a measurement. The defence is the same for both — make the validity chain and the construction choices explicit, and publish them.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/indicators/baselines-targets-disaggregationhttps://monival.org/indicators/baselines-targets-disaggregationA target is only as meaningful as the baseline under it and the disaggregation behind it. Baselines establish where performance stands before the intervention; targets state the value sought by a date, set by methods that range from trend projection to open negotiation; disaggregation decides — at design time, not analysis time — whose outcomes will be visible separately. All three are design decisions with governance attached, and each has failure modes that surface years later as unexplainable results.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/management/project-management-and-mehttps://monival.org/management/project-management-and-meProject management and M&E measure the same project with different questions: PM asks whether the work is on scope, on schedule and within budget; M&E asks whether the work is producing the change it promised. A programme can be perfectly on schedule and achieving nothing — and unless the two functions share artefacts, data and review moments, only one of them will notice.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/management/earned-value-managementhttps://monival.org/management/earned-value-managementEarned value management prices the work actually performed and compares it with the work planned and the money spent, producing schedule and cost variances a plain budget report cannot see. It is a delivery instrument: it measures whether work is being done efficiently and on time, and says nothing — by design — about whether the work is achieving results.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/management/risk-registershttps://monival.org/management/risk-registersA risk register is the visible residue of a risk management process; kept without the process, it is a spreadsheet of anxieties. The register earns its place when each entry states a cause, an event and a consequence, carries a named owner and a live treatment, and is wired to early-warning indicators the monitoring system already collects.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/management/balanced-scorecardhttps://monival.org/management/balanced-scorecardThe Balanced Scorecard sets a small number of measures across four perspectives — financial, customer, internal process, and learning and growth — so that an organisation steers by more than its financial results. Its strategy-map extension chains objectives across the perspectives as cause-and-effect hypotheses. The critique the field has never fully answered is that those causal links are asserted far more often than they are tested.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/management/okrs-and-kpi-cascadeshttps://monival.org/management/okrs-and-kpi-cascadesOKRs pair a qualitative, ambitious objective with a few measurable key results, graded honestly at the end of a short cycle. They work as a focus-and-alignment device, not an accountability system — and the moment key results are wired to appraisal or budget consequences, Goodhart's law begins converting every measure into a worse one.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/management/mid-term-reviewshttps://monival.org/management/mid-term-reviewsA mid-term review is a formative exercise conducted while a programme can still change: its accountability weight is deliberately lower than an evaluation's, its learning weight higher, and its real product is not the report but the management response and the re-planned second half. An MTR that produces findings and no decisions has failed at its one job.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/meal/accountability-to-affected-populationshttps://monival.org/meal/accountability-to-affected-populationsAccountability to affected populations is the commitment to use power responsibly: to take account of the views of people receiving assistance, give account to them for what the organisation does, and be held to account by them. The IASC commitments and the Core Humanitarian Standard (revised 2024) turn that principle into operational requirements that run through the whole M&E cycle — participation in design, community validation of findings, and closing the loop on results.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/meal/feedback-and-complaints-mechanismshttps://monival.org/meal/feedback-and-complaints-mechanismsA feedback and complaints mechanism (FCM) is working only when the loop closes: the person who raised an issue receives a response, and the programme demonstrably changes where the feedback warrants it. Design starts from the channels communities actually prefer, separates sensitive complaints into confidential pathways from the first moment, and treats the aggregated feedback stream as monitoring data in its own right.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/meal/learning-and-adaptive-managementhttps://monival.org/meal/learning-and-adaptive-managementLearning becomes operational when it is attached to decisions: a learning agenda names the questions that matter and who will act on the answers; adaptive management shortens the cycle between evidence and adjustment; and the after-action review gives teams a repeatable ritual for extracting lessons while they are still usable. None of it requires new data so much as new discipline about what existing evidence is for.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/meal/ethics-and-consenthttps://monival.org/meal/ethics-and-consentM&E collects data from people, usually under a power asymmetry, and that fact — not the label 'research' — is what triggers ethical obligations. The shared vocabulary comes from the Belmont Report's three principles and UNEG's four; the practice comes down to consent that is genuinely informed and voluntary, fieldwork that anticipates harm, and knowing when a study crosses into territory that requires formal ethical review and national research licensing.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/meal/data-protectionhttps://monival.org/meal/data-protectionOrdinary M&E data — beneficiary registers, phone numbers, GPS points, photographs, survey responses — is personal data, and processing it engages data-protection law. In Kenya that law is the Data Protection Act, 2019, overseen by the Office of the Data Protection Commissioner; internationally funded programmes frequently also fall within the GDPR's reach. This page describes what the two instruments require; it is a description of the law, not legal advice.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/randomised-controlled-trialshttps://monival.org/methods/randomised-controlled-trialsA randomised controlled trial assigns eligible units to treatment or control by lottery, making the two groups statistically identical before the programme starts — so the difference in their later outcomes is an unbiased estimate of impact. The design's strength is internal validity; its limits are feasibility, ethics, cost, and the fact that a trial alone says little about why an effect occurred or whether it will travel.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/difference-in-differenceshttps://monival.org/methods/difference-in-differencesDifference-in-differences (DiD) estimates a programme's impact by comparing the change in outcomes over time between a group that received the intervention and one that did not. It removes fixed differences between the groups and shared trends over time — but it stands or falls on the assumption that, absent the programme, both groups would have moved in parallel.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/regression-discontinuityhttps://monival.org/methods/regression-discontinuityA regression discontinuity design (RDD) exploits an assignment rule: when a programme is given to units on one side of a cutoff on a continuous score, units just above and just below that cutoff are nearly identical except for treatment. The jump in outcomes at the cutoff is then a credible causal estimate — but only for the population near the cutoff, and only if no one can manipulate which side of it they land on.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/matching-and-propensity-scoreshttps://monival.org/methods/matching-and-propensity-scoresMatching builds a comparison group by pairing programme participants with non-participants who look the same on measured characteristics, so that outcome differences can be read as impact. Everything rests on one strong assumption: that the measured characteristics capture all the differences that matter. Where selection into the programme ran on motivation, connections or need that the data never recorded, matching reproduces the bias it was meant to remove.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/interrupted-time-serieshttps://monival.org/methods/interrupted-time-seriesInterrupted time series (ITS) evaluates an intervention by modelling a long series of outcome measurements before it began, projecting that trend forward as the counterfactual, and testing whether the observed series breaks from it — in level, in slope, or both. It is the strongest design available when a population-level intervention starts at a known time and no comparison group exists, and it runs naturally on routine data.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/synthetic-controlhttps://monival.org/methods/synthetic-controlSynthetic control evaluates an intervention that hit a single aggregate unit — one county, one country, one region — by constructing a weighted combination of untreated units that reproduces the treated unit's pre-intervention trajectory. The post-intervention gap between the real unit and its synthetic twin is the estimated effect. The method's credibility is earned almost entirely before the intervention: if the synthetic cannot track the treated unit's history, it has no claim on its counterfactual future.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/contribution-analysishttps://monival.org/methods/contribution-analysisContribution analysis, developed by John Mayne, builds a credible case that a programme contributed to observed results when no counterfactual design is feasible. It works by making the programme's theory of change explicit, testing each link against evidence, taking rival explanations seriously, and iterating until the resulting 'contribution story' is plausible enough — or shown not to be. It asks whether the programme made a difference as one cause among several, not whether it alone produced the result.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/process-tracinghttps://monival.org/methods/process-tracingProcess tracing establishes causation within a single case by spelling out the mechanism through which an intervention is claimed to have produced a result, deriving what evidence should exist if that mechanism operated — and what should exist if it did not — and then testing. Its core discipline is the probative value of evidence: four classic tests grade each piece of evidence by whether it is necessary and/or sufficient to sustain the causal claim.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/realist-evaluationhttps://monival.org/methods/realist-evaluationRealist evaluation, developed by Ray Pawson and Nick Tilley, starts from the observation that programmes do not work uniformly: they offer resources that people reason about and respond to, and those responses — the mechanisms — fire in some contexts and not others. The method builds and tests context–mechanism–outcome (CMO) configurations to answer not 'did it work?' but 'what works, for whom, in what circumstances, and why?'Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/methods/qualitative-comparative-analysishttps://monival.org/methods/qualitative-comparative-analysisQualitative comparative analysis identifies the combinations of conditions under which an outcome occurs across a medium number of cases — typically too many for case studies, too few for regression. Grounded in set theory rather than correlation, it tests which conditions are necessary or sufficient for the outcome, accepts that different recipes can produce the same result in different cases, and makes its analytic choices — especially calibration — explicit and checkable.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/reporting/evaluation-reportshttps://monival.org/reporting/evaluation-reportsA credible evaluation report is a traceable argument: every finding rests on identifiable evidence, every conclusion on findings, every recommendation on conclusions. Structure, methodology honesty and a stand-alone executive summary serve that chain — and the chain, not the prose, is what quality reviewers test first.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/reporting/dashboards-and-data-visualisationhttps://monival.org/reporting/dashboards-and-data-visualisationA dashboard is a recurring decision aid and a report figure is an argument — different jobs, different design rules, one shared obligation: the picture must not claim more than the data do. Truncated axes, dual axes and area distortion are ethics problems before they are aesthetics problems, because a chart is a factual claim its reader cannot easily audit.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/reporting/use-and-management-responsehttps://monival.org/reporting/use-and-management-responseAn evaluation is used when it changes something — a decision, a design, an understanding — and the management response is the instrument that forces the question. Every recommendation receives an explicit accept, partially-accept or reject with reasons, an owner, a deadline and tracked actions; a reasoned rejection is legitimate use, and a report with no response is the definition of shelf-ware.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/standards/uneg-norms-and-standardshttps://monival.org/standards/uneg-norms-and-standardsThe UNEG Norms and Standards for Evaluation are the United Nations system's common reference for what evaluation must be and how evaluation functions must deliver it. First adopted in 2005 and revised in 2016, they pair general norms — principles such as independence, impartiality, credibility, transparency, ethics and utility — with standards that translate those principles into institutional arrangements, management practice, competencies and conduct. Any organisation, UN or not, can declare conformity with them; what it gains is a widely recognised external anchor for evaluation quality, and what it does not gain is an enforcement mechanism.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/standards/program-evaluation-standardshttps://monival.org/standards/program-evaluation-standardsNorth American evaluation rests on two distinct instruments that are routinely conflated. The JCSEE Program Evaluation Standards (3rd edition, 2010) judge evaluations: thirty standards in five attribute groups — utility, feasibility, propriety, accuracy and evaluation accountability. The AEA Guiding Principles for Evaluators, first adopted in 1994 and most recently updated in 2025, guide evaluators: systematic inquiry, competence, integrity, respect for people, and common good and equity. One assesses the work; the other governs the worker — and rigorous practice uses both.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/standards/african-evaluation-principleshttps://monival.org/standards/african-evaluation-principlesThe African Evaluation Principles, adopted by the African Evaluation Association (AfrEA) in 2021, are the continent's own normative instrument for evaluation — the successor to the African Evaluation Guidelines of the early 2000s. Built on five principles — empowerment of Africans, technical robustness, ethical soundness, rooted in Africa, and global connectedness — they carry the Made in Africa Evaluation agenda from critique into commissioning practice: evaluation in Africa should be led by Africans, grounded in African contexts and knowledge systems, and rigorous by any standard. Around them runs a genuine, live debate about whether imported frameworks can be adapted or must be rebuilt.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/standards/evaluation-quality-assessmenthttps://monival.org/standards/evaluation-quality-assessmentEvaluation quality is assessable at three layers — the process that produced the evaluation, the report it produced, and the use anyone made of it — and each layer has its own instruments. The OECD DAC Quality Standards for Development Evaluation (2010) govern process, phase by phase; the UNEG Quality Checklist (2010) structures product review; the JCSEE evaluation accountability standards supply the warrant for meta-evaluation. A sound quality assurance system applies all three lenses in proportion to the stakes: a two-hour structured review for a routine study, a commissioned meta-evaluation for a flagship one.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/systems/me-information-systemshttps://monival.org/systems/me-information-systemsAn M&E information system is the machinery that moves data from the point of capture to the point of decision. Its design starts from the decisions the organisation must make — not from forms — and its quality is decided by data-flow architecture, shared master lists and interoperability, long before any software is chosen.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/systems/dhis2https://monival.org/systems/dhis2DHIS2 is an open-source data management platform, maintained by the HISP Centre at the University of Oslo, that anchors the routine health information systems of ministries of health across Africa and Asia — including Kenya's. For programme M&E it is three things at once: a source of national baseline and denominator data, a destination for programme reporting, and a standing argument against building yet another parallel system.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/systems/results-based-managementhttps://monival.org/systems/results-based-managementResults-based management is a management strategy, not a reporting template: all actors direct their processes and resources at agreed results, and use evidence on actual results to steer decisions. The apparatus — results chains, indicators, monitoring, evaluation — only earns its cost when performance information changes what the organisation does next.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/systems/national-me-systemshttps://monival.org/systems/national-me-systemsA national M&E system is a set of institutions, mandates, data systems and use routines through which a government monitors and evaluates its own performance — not a database purchase. Kenya's architecture pairs the National Integrated Monitoring and Evaluation System (NIMES) with a 2022 National M&E Policy and the eNIMES platform; the systems that endure anywhere are the ones whose information decision-makers actually demand.Tue, 18 Aug 2026 00:00:00 GMThttps://monival.org/systems/performance-contracting-kenyahttps://monival.org/systems/performance-contracting-kenyaPerformance contracting in Kenya is a negotiated annual agreement between the Government and each public institution, cascaded down to individual officers, evaluated and ranked at year end. The FY2025/26 guidelines mark the twenty-second consecutive annual cycle. The instrument's strengths and its pathologies — ratchet effects, soft targets, measurement gaming — both follow directly from its incentive design.Tue, 18 Aug 2026 00:00:00 GMT