What is evaluation
Evaluation is the systematic and objective assessment of an ongoing or completed project, programme or policy, including its design, implementation and results. The aim of evaluation is to determine the relevance and fulfilment of objectives, efficiency, effectiveness, impact and sustainability of a given intervention.
OECD/DAC definition: Evaluation is "the systematic and objective assessment of an ongoing or completed activity, project, programme or policy, its design, implementation and results, with the aim of determining the relevance and fulfilment of objectives, efficiency, effectiveness, impact and sustainability."
Why is evaluation important?
- Accountability – evaluation provides evidence of how resources were used and what results were achieved
- Learning – policymakers, programme managers and implementers can all learn from evaluation findings
- Improvement – evaluation identifies the strengths and weaknesses of an intervention and provides recommendations for improvement
- Evidence-based decision-making – evaluation findings support informed decisions about future interventions
- Transparency – publishing evaluation reports increases the transparency of public spending
Evaluation vs. monitoring vs. audit
It is important to distinguish evaluation from related but different activities:
- Monitoring is an ongoing process of collecting data on the implementation of a project/programme. Unlike evaluation, it is continuous and focuses primarily on inputs, activities and outputs.
- Audit focuses on verifying compliance with rules and regulations. It assesses whether funds were used in accordance with legal and contractual requirements.
- Evaluation goes further – it assesses relevance, effectiveness and impact. It seeks to understand why the intended results were or were not achieved.
Types of evaluation
By timing
- Ex ante evaluation – carried out before a programme/project begins. It assesses the relevance of the proposed intervention, the quality of the intervention logic, and the realism of objectives and indicators.
- Mid-term evaluation – carried out during implementation. It assesses the progress of implementation and allows the direction of the programme to be adjusted.
- Ex post evaluation – carried out after a programme/project is completed. It assesses the results achieved, impacts and sustainability.
By who carries it out
- Internal evaluation (self-assessment) – carried out by staff of the organisation implementing the programme
- External evaluation – carried out by an independent external evaluator or evaluation team
- Mixed evaluation – a combination of internal and external evaluation
- Peer review – carried out by colleagues from similar organisations
By purpose
- Formative evaluation – focused on improving a programme during its implementation
- Summative evaluation – focused on the overall assessment of a programme and decisions about its future
- Impact evaluation – focused on measuring the long-term changes caused by a programme
- Process evaluation – assesses how a programme was implemented
- Needs assessment – identifies the needs of a target group before an intervention is designed
Special types
- Thematic evaluation – focuses on a specific theme across several programmes (e.g. gender equality, sustainability)
- Cluster evaluation – assessment of a group of related projects
- Meta-evaluation – evaluation of the quality of other evaluations
- Evaluability assessment – assesses whether a programme is ready for evaluation
Evaluation methods
Quantitative methods
- Randomised controlled trials (RCTs) – the "gold standard" in impact evaluation. Random assignment to a treatment and a control group.
- Quasi-experimental methods – use comparison groups without random selection (difference-in-differences, regression discontinuity, propensity score matching)
- Surveys – structured collection of quantitative data from a large number of respondents
- Statistical analysis of secondary data – analysis of existing administrative or statistical data
- Cost-benefit analysis (CBA) – comparison of the costs and benefits of an intervention in monetary terms
- Cost-effectiveness analysis (CEA) – comparison of the costs of different alternatives for achieving the same result
Qualitative methods
- In-depth interviews – semi-structured or unstructured interviews with key actors
- Focus groups – group discussions with programme participants or stakeholders
- Case studies – in-depth analysis of specific cases or projects
- Observation – direct observation of programme implementation in practice
- Document analysis – systematic review of project documentation, reports and other materials
- Most Significant Change (MSC) – a participatory method for collecting and selecting stories about the most significant changes
- Outcome Harvesting – identification and verification of outcomes to which a programme has contributed
Mixed methods
Combining quantitative and qualitative methods is considered best practice in evaluation. Quantitative data provide information about the scale of change and causality, while qualitative data help to understand the context, processes and experiences of the actors involved.
Tip: The choice of evaluation method depends on the evaluation questions, available resources, time frame and context. There is no universally best method – the key is to choose the approach that best answers the evaluation questions.
Participatory approaches
- Participatory evaluation – active involvement of stakeholders in the evaluation process
- Empowerment evaluation – evaluation focused on strengthening the capacities of communities
- Developmental evaluation – ongoing evaluation of innovative programmes in complex environments
- Utilization-focused evaluation – evaluation designed to maximise the use of findings by primary users
The evaluation cycle
The evaluation cycle usually involves the following phases:
1. Planning the evaluation
- Defining the purpose and scope of the evaluation
- Formulating the evaluation questions
- Identifying stakeholders
- Setting the budget and timetable
2. Terms of Reference (ToR)
- Context and background of the programme
- Purpose and objectives of the evaluation
- Evaluation questions and criteria
- Methodological requirements
- Expected outputs and timetable
- Qualification requirements for the evaluator
3. Selecting the evaluator
The selection process usually involves public procurement or approaching established evaluators. Key criteria: expertise in the relevant field, methodological competence, independence and the absence of any conflict of interest.
4. Inception (the initial phase)
- Refining the evaluation design and methodology
- Preparing the inception report
- Preparing the data collection tools
5. Data collection
Carrying out field research according to the approved methodology – interviews, surveys, document analysis and so on.
6. Analysis and interpretation
Systematic analysis of the data collected, triangulation of findings and formulation of conclusions.
7. Reporting and presenting the results
- Preparing the draft evaluation report
- Comments from the commissioner and stakeholders
- Finalising the report with recommendations
- Presenting the findings
8. Use of findings
Implementing the recommendations, the management response, and tracking the implementation of the recommendations.
Evaluation standards
OECD/DAC evaluation criteria
Revised in 2019, these criteria form the basis of most evaluations in international development:
- Relevance – Is the intervention meaningful and does it respond to the needs of the target groups?
- Coherence – Is the intervention consistent with other interventions in the field? (a new criterion since 2019)
- Effectiveness – Did the intervention achieve its objectives?
- Efficiency – Were resources used economically?
- Impact – What changes did the intervention bring about?
- Sustainability – Are the results sustainable even after the intervention ends?
Ethical standards
- Independence and impartiality of the evaluator
- Confidentiality and protection of respondents' data
- Informed consent of participants
- Transparency of the process and findings
- Cultural sensitivity
- The "do no harm" principle
International evaluation societies and standards
- American Evaluation Association (AEA) – Program Evaluation Standards
- European Evaluation Society (EES) – European evaluation standards
- UNEG – Norms and Standards for Evaluation in the UN system
- OECD/DAC Network on Development Evaluation – quality standards for the evaluation of development cooperation
Evaluation in public administration
Evaluation in public administration in Slovakia is developing mainly in the context of drawing on EU funds and initiatives to increase the efficiency of public spending.
Evaluation of ESIF and EU funds
The European Structural and Investment Funds (ESIF) require systematic evaluation of operational programmes. In Slovakia, the coordination of evaluations is ensured by the Central Coordinating Body (CCB) at the Ministry of Investments, Regional Development and Informatisation of the Slovak Republic (MIRRI).
- Ex ante evaluation of operational programmes
- Mid-term evaluations of implementation
- Ex post evaluations of results and impacts
- Evaluation plan for the Slovakia Programme 2021–2027
Value for money
The Value for Money Unit (ÚHP) at the Ministry of Finance of the Slovak Republic introduced systematic spending reviews that assess the efficiency of spending across ministries. By 2026, more than 39 evaluation reports had been prepared.
Ex post evaluation of regulations
Since 2022, an obligation to carry out ex post evaluation of regulations has been in place in Slovakia, based on the Unified Methodology for Assessing Selected Impacts, in line with the Recovery and Resilience Plan of the Slovak Republic.
Challenge: According to the OECD, Slovakia should strengthen regulatory oversight and make more systematic use of ex ante and ex post evaluation of the impacts of regulations.
Evaluation in the non-governmental sector
Non-governmental organisations in Slovakia use evaluations to assess the effectiveness of their programmes and projects, to increase accountability towards donors and the public, and to learn and improve.
Main areas of evaluation in NGOs
- Social inclusion and education – programmes for marginalised Roma communities, inclusive education
- Transparency and anti-corruption – evaluation of anti-corruption measures and transparency
- Civil society and participation – involving citizens in decision-making
- Social innovation – new approaches to solving social problems
- Human rights – monitoring and evaluation of respect for human rights
Key organisations
- SGI – Slovak Governance Institute – evaluations in the fields of public administration, education and social inclusion
- Transparency International Slovensko – evaluation of transparency and anti-corruption measures
- MESA10 – analytical and evaluation activities in the field of education
- Centre for the Research of Ethnicity and Culture (CVEK) – evaluations in the field of ethnic minorities
- Open Society Foundation – support for an evaluation culture in the non-governmental sector
Evaluation in international development
The Slovak Republic has been a member of the OECD/DAC since 2013 and is systematically building its system of development cooperation under the SlovakAid brand.
SlovakAid and evaluation
The Slovak Agency for International Development Cooperation (SAIDC) is responsible for the monitoring and evaluation of bilateral and trilateral development projects. The Ministry of Foreign and European Affairs (MZVaEZ) ensures independent evaluations of programmes.
OECD/DAC Peer Review
Slovakia underwent DAC peer reviews in 2019 and 2025. The 2025 review confirmed progress in strengthening the development cooperation system, recognised technical assistance as a strength of the Slovak Republic, and recommended, among other things:
- Increasing the level of ODA
- Concentrating resources on a smaller number of programme initiatives
- Continuing regular independent evaluations
The Public Finance for Development Programme
Implemented since 2009 in cooperation with UNDP, it focuses on sharing the Slovak Republic's experience in public finance management with partner countries.
Medium-Term Strategy for Development Cooperation of the Slovak Republic 2025–2030
The strategy maintains the existing thematic priorities and introduces a new "Green Programme" to support environmental sustainability.
Evaluation in Slovakia – history and context
An evaluation culture in Slovakia began to develop in the 1990s, particularly in the context of preparation for EU accession and with the support of international organisations. Key milestones:
- 1990s – The first evaluation activities in the context of the PHARE programmes and EU accession
- 2001 – The founding of SGI (Slovak Governance Institute) as one of the first organisations systematically dedicated to evaluation in Slovakia
- 2003 – The start of building the Slovak Republic's development cooperation system (SlovakAid)
- 2004 – Slovakia's accession to the EU, the start of drawing on structural funds with evaluation requirements
- 2013 – Slovakia's accession to the OECD/DAC
- 2016 – The creation of the Value for Money Unit at the Ministry of Finance of the Slovak Republic
- 2019 – The first OECD/DAC Peer Review for the Slovak Republic
- 2022 – The introduction of an obligation to carry out ex post evaluation of regulations
- 2025 – The second OECD/DAC Peer Review, the new Medium-Term Strategy for Development Cooperation 2025–2030
Note: Despite this progress, the evaluation culture in Slovakia remains relatively less developed than in Western European countries. Systematically building evaluation capacities and demand for high-quality evaluations is a key challenge.
The use of AI in evaluation
Artificial intelligence (AI) – particularly machine learning, natural language processing and generative models – is rapidly changing the way evaluations are carried out. AI can streamline time-consuming tasks, process large volumes of both qualitative and quantitative data, and reveal patterns that traditional methods might not capture. At the same time, it brings new risks and ethical questions that a responsible evaluator must actively manage.
Context: AI in evaluation is not a replacement for the evaluator, but a tool that extends their possibilities. The quality of findings still depends on professional judgement, contextual understanding and a critical assessment of the AI outputs.
Areas of AI use in evaluation
- Data collection and preparation – automatic transcription of interviews and focus groups, extraction of data from documents and forms, cleaning and structuring of large datasets
- Natural language processing (NLP) – automated coding of qualitative data, sentiment analysis, thematic analysis of open-ended survey responses, text classification
- Analysis of large volumes of text – systematic literature reviews, analysis of hundreds of project reports or media outputs, identification of recurring themes across extensive sets of documents
- Machine learning in impact evaluation – predictive modelling, identification of subgroups with different effects (heterogeneity of impacts), modelling the counterfactual and finding patterns in extensive administrative data
- Generative AI – summarisation of extensive source material, suggestions for evaluation questions and indicators, support for writing and structuring reports, generating visualisations and first drafts of texts
- Visualisation and reporting – automated creation of charts, interactive dashboards and clear summaries for different target groups
NLP and the analysis of qualitative data
The most significant benefit of AI so far appears in working with qualitative data, which is traditionally very demanding to process. Natural language processing tools can support:
- Automated (assisted) coding – suggesting coding schemes and preliminary assignment of codes to text segments, which the evaluator then verifies and adjusts
- Thematic analysis – identifying recurring themes and the relationships between them in extensive sets of interviews or responses
- Sentiment and tone analysis – assessing the attitudes and emotions of respondents in large volumes of text data
- Summarisation – summarising long transcripts and documents while preserving the key findings
Case study – World Bank IEG: The World Bank's Independent Evaluation Group uses advanced content analysis combining NLP, machine learning (both supervised and unsupervised learning) and knowledge graphs to analyse large volumes of project documents. This approach enables theory-driven evaluation at scale – systematically processing thousands of documents that manual analysis could not handle.
Ethical aspects and risks
The use of AI in evaluation brings significant risks that must be actively managed:
- Bias – AI models can reproduce and amplify the biases contained in their training data, which may disadvantage certain groups
- Hallucinations and accuracy – generative models can produce convincing-sounding but untrue statements; every output must be verified against primary sources
- Data protection and privacy – sensitive respondent data must not be entered into public tools without proper anonymisation and consent
- Transparency and explainability – the evaluator must be able to explain how findings were obtained, including the role of AI in the process
- Human oversight (human-in-the-loop) – the final responsibility for conclusions and recommendations remains with the evaluator, not the tool
- Reproducibility – the outputs of generative models may differ when repeated; it is necessary to document the tools, versions and procedures used
The AIDEM framework: At the conference of the American Evaluation Association (AEA 2025), the six-step AIDEM framework for ethical and responsible decision-making about the use of AI in evaluation was presented. The framework guides the evaluator from clarifying the purpose and assessing the suitability of AI, through identifying risks (data protection, bias) and ensuring human oversight, to transparently documenting and communicating the way AI was used.
Practical recommendations for evaluators
- Use AI as an assistant, not as an authority – critically verify every output and triangulate it with other sources
- Clarify in advance which tasks AI is suitable for (summarisation, preliminary coding, data cleaning) and which it is not (final evaluative judgements)
- Protect respondents' data – anonymise sensitive information and use tools with appropriate privacy safeguards
- Document the tools, versions, prompts and procedures used, in the interest of transparency and reproducibility
- Be aware of the risk of bias and actively assess it, especially with vulnerable groups
- Invest in capacity building – understanding both the possibilities and the limits of AI is a new key competence for the evaluator
Glossary of terms
- Baseline – the situation before the start of an intervention, against which changes are measured
- Beneficiary – the target group for which an intervention is intended
- Counterfactual – what would have happened without the intervention
- DAC – Development Assistance Committee (the OECD committee for development assistance)
- Donor – the provider of funds
- Evaluation criteria – the perspectives according to which an intervention is assessed
- Evaluation question – a specific question that the evaluation is to answer
- Indicator – a measurable piece of data that signals a change
- Intervention logic – the chain of assumed causal relationships: inputs → activities → outputs → outcomes → impact
- Logical framework (logframe) – a matrix showing the intervention logic, indicators, sources of verification and assumptions
- ODA – Official Development Assistance
- Stakeholder – an interested party
- Theory of Change – an explanation of how and why an intervention is expected to lead to the desired change
- ToR (Terms of Reference) – the terms of reference for an evaluation
- Triangulation – verification of findings from multiple sources and methods
Useful resources
International resources
- OECD/DAC Network on Development Evaluation – standards, guidelines and a database of evaluations
- UNEG (United Nations Evaluation Group) – norms and standards for evaluation in the UN system
- Better Evaluation – a comprehensive resource on evaluation methods and approaches
- European Evaluation Society (EES) – the European evaluation community
- J-PAL (Abdul Latif Jameel Poverty Action Lab) – a leader in randomised evaluations
Slovak resources
- Partnership Agreement – ESIF Evaluation – a database of ESIF evaluation reports
- Value for Money Unit (Ministry of Finance of the Slovak Republic) – spending reviews
- SlovakAid – evaluations in the field of development cooperation
- MIRRI SR – evaluations of operational programmes and EU funds
- SGI – Slovak Governance Institute – research and evaluation of public policies