Two participatory, theory-based evaluation methods that do not depend on counterfactual design. Outcome Harvesting works backward from observed outcomes; Most Significant Change works forward from participant stories. Here is when each wins and how they combine.
The most common confusion in M&E. Learn the difference between outputs, outcomes, and impact with clear examples from health, education, and food security programs.
A panel follows the same people across rounds; a repeated cross-section draws a new sample each time. The choice depends on the change you need to measure, tracking, attrition and cost.
Guías de decisión · Recolección de datos · En inglés
Digital is usually the right answer, but not always. Here is when paper still wins, when digital pays for itself, and the hybrid workflow that handles both.
A sample that estimates a level well may not detect a change, and neither classifies an area against a target. One worked case sizes the same indicator three ways.
Probability vs non-probability sampling in M&E: when each approach is valid, which method fits your context, and five common mistakes that invalidate findings.
Process indicators tell you whether the program is running as designed. Outcome indicators tell you whether it is producing the intended change. You need both. Here is how to pick the right mix, and how to avoid reporting only one while pretending to measure the other.
Five illustrative program evaluation examples, from education to social work, each showing the question, design, data and what the evaluation could and could not conclude.
Program evaluation is a structured look at whether a program works and why. This page gives the main types, six steps to follow and a short example, for nonprofit and public program staff.
Performance measures are the few numbers a program tracks to see whether it is on course. Learn how they differ from evaluation, a five-test filter for choosing them, and examples.
Qualitative, quantitative, and mixed methods are not a quality ranking. They answer different questions. Here's when to use each, how to combine them, and what integration actually looks like.
Guías de decisión · Recolección de datos · En inglés
Use pretesting to find question problems and a field pilot to test the survey process. A practical sequence for wording, translation, routing and readiness.
When to use a randomized controlled trial vs a quasi-experimental design. Feasibility, cost, rigor, and what each can actually tell you about your program's impact.
Guías de decisión · Recolección de datos · En inglés
One fictional coverage survey, calculated three ways. A worked case shows how weights, design effect and the confidence interval change what the data can support.
Reliability asks whether a measure gives the same result each time. Validity asks whether it measures the right thing. A contrast table, two fictional examples and the checks to run.
A theory of change explains how change happens. A results framework fixes what you will commit to and report. See the same program both ways and what is lost in translation.
Simple random and systematic sampling both give every listed unit an equal chance. Systematic is quicker to run, but a hidden pattern in the list can break it.
Most indicators fail SMART review because Specific and Measurable are vague. Here is how to apply the framework properly, with sector examples and the revisions that fix common mistakes.
What SMART objectives are, before and after examples for grant and program objectives in education, health and social services, and a checklist to test each one. This page covers program objectives, not personal or business goals.
A template for planning who takes part in an evaluation, when and how, with a stakeholder map worksheet. Includes a filled example that follows one evaluation from design to dissemination.
A result can be statistically significant and still too small to matter, or large and not significant. A worked case shows how to read the p-value, the effect size and the interval together.
Two fictional villages share the same average spending and very different households. A worked case shows which summary to report, and what weights and small samples change.
Guías de decisión · Recolección de datos · En inglés
The three most common M&E data collection methods, compared. Surveys tell you how many, interviews tell you why, focus groups tell you what people agree on.
Validity, Reliability, Timeliness, Precision, Integrity. The five dimensions every DQA tests against, with the specific failures each one catches and how to fix them.