Types of Evaluation: From Formative to Longitudinal Approaches

Every project and government scheme eventually faces a simple but demanding question: did it work? Answering it well is rarely a single judgement made at the end. Instead, evaluation happens in different forms, at different moments, and through different agencies, each designed to answer a slightly different version of “did it work?” Understanding these types of evaluation helps you read assessment reports critically and design better monitoring systems for any programme, from a municipal sanitation drive to a national flagship scheme. This post walks through the main categories: evaluation by focus, by agency and stage, and the more specialised approaches used for long-term tracking.

Evaluation by focus: formative and summative
Formative evaluation: learning while doing
Summative evaluation: judging the final result
Evaluation by agency and stage
Internal versus external evaluation
Stages of evaluation: ongoing, terminal, and ex-post
Specialised evaluations for long-term tracking
Concurrent evaluation
Longitudinal evaluation
Putting the types together

Evaluation by focus: formative and summative

The most fundamental distinction in evaluation is based on its purpose, or focus. This split was first introduced by the educational theorist Michael Scriven in 1967, who separated evaluation that guides and improves a process while it is still happening from evaluation that judges a finished product. That basic logic still organises how we think about assessment today.

Formative evaluation: learning while doing

Formative evaluation is conducted during the development or implementation of a project. Its goal is not to deliver a final verdict but to monitor progress and feed information back into the project so managers can make corrections. It is iterative, starts early in the project’s life, and is closely tied to monitoring. As one analysis of project management notes, formative evaluation provides the feedback loop that guides project change during implementation, focusing on understanding and learning rather than final scoring.

Think of a skill-development scheme that is two years into a five-year run. A mid-term review checks whether training centres are reaching their enrolment targets, whether the budget is being spent as planned, and whether trainers need more support. If enrolment is lagging in certain districts, managers can redirect resources before the problem becomes permanent. This is formative evaluation in action: it exists to improve the programme while there is still time to act.

Summative evaluation: judging the final result

Summative evaluation takes the opposite vantage point. It assesses the overall effectiveness of a programme once it is complete or has reached a major milestone. Where formative evaluation asks “how can we improve this?”, summative evaluation asks “did this achieve what it set out to achieve?” Summative evaluations assess the overall experience of a finished product against a benchmark, producing the kind of big-picture conclusion that decision-makers use to decide whether to continue, scale, or close a programme.

A common misconception is that summative means quantitative and formative means qualitative. This is not the case. Summative evaluation can use either numbers or narrative; what defines it is its timing and purpose, not its method.

The two are complementary rather than competing. A programme that relies only on summative evaluation risks discovering its failures too late to fix them. A programme that uses only formative evaluation may lack the structured accountability needed to certify its outcomes. The strongest evaluation frameworks use both together. As research on assessment puts it, formative and summative evaluations are both critical for the success of programmes, projects, and policies.

Evaluation by agency and stage

The second way to classify evaluations is by who conducts them and when. These two dimensions, agency and stage, often appear together in official guidelines and determine both the credibility and the usefulness of an evaluation.

Internal versus external evaluation

An internal evaluation is carried out by people within the implementing organisation, often by the project management team itself. When it is conducted by staff directly responsible for delivering the project, it is sometimes called a self-evaluation. Internal evaluations are usually cheaper and benefit from insiders’ deep knowledge of the programme, but they can carry a risk of bias because the evaluators have a stake in the result.

An external evaluation is conducted by someone recruited from outside, typically a consultant or an independent agency engaged by the donor or implementing organisation. According to guidance from the International Organization for Migration, an external evaluation is conducted by someone recruited externally and is considered an independent evaluation. These cost more but bring objectivity and credibility, which matters when findings will inform major funding decisions.

This is exactly why independent, third-party evaluation has become central to government practice. The Government has made evaluation of Centrally Sponsored Schemes mandatory before they come up for fresh appraisal, and has assigned the Development Monitoring and Evaluation Office (DMEO) at NITI Aayog the responsibility to conduct independent third-party evaluation of all such schemes in a time-bound manner. The independence of the evaluator is treated as a feature, not an afterthought.

Stages of evaluation: ongoing, terminal, and ex-post

Evaluations are also classified by the stage of the project at which they happen.

Ongoing evaluation takes place during implementation. It monitors live details such as schedule, budget, and quality of work, and overlaps heavily with formative evaluation. Its value lies in catching problems early. The common feature of most government programmes is that they run across long time frames and changing administrative scenarios, which makes continuous monitoring necessary to record results across the lifetime of the programme and to make mid-course corrections.

Terminal evaluation, also called final or end-of-project evaluation, is conducted when all activities are complete. It assesses whether the project met its predefined goals and objectives within the planned timeline. Final evaluations often include predictions about long-term benefits, but because they rely on short-term observations made at the moment of closure, they cannot fully capture whether project benefits actually continue after the project ends. They are based, in part, on assumptions about the future.

Ex-post evaluation closes that gap. Carried out several years after a project finishes, it measures the actual continuation of benefits rather than projected ones. It is usually an independent review of the project’s history, objectives, results, and resources, conducted to draw lessons that may be useful in future project activity and to identify procedures and techniques that did not work. The World Bank pioneered this approach decades ago, applying ex-post evaluation to projects shortly after the end of their investment phase to serve many users in many ways. The defining strength of ex-post evaluation is sustainability: it answers whether the change a project created actually lasted.

It is worth noting a related stage that sits at the very beginning. Ex-ante evaluation, conducted before a project starts, helps establish procedures and define criteria for project selection and expected results. Together, ex-ante and ex-post evaluations bookend a project, one looking forward to plan it well and the other looking back to learn from it.

Specialised evaluations for long-term tracking

Beyond these standard categories, certain programmes need approaches built specifically for tracking change over extended periods. Two stand out: concurrent and longitudinal evaluation.

Concurrent evaluation

Concurrent evaluation runs alongside a programme in real time, generating findings while the programme is still active so that course corrections can be made without waiting for it to end. It shares the spirit of formative and ongoing evaluation but is often mandated as a formal, recurring requirement in large schemes. In India’s federal set-up, where centrally sponsored schemes have sometimes been criticised for weak ownership at the state level, the presence of concurrent evaluations enables mid-course corrections and improves the effectiveness of programmes. Some flagship programmes build it directly into their design; for instance, the National Food Security Mission mandates constant monitoring and concurrent evaluation every year, with state agriculture and statistics departments responsible for carrying it out.

The rise of digitised government data has made concurrent evaluation far more practical. Administrative data forms the backbone of decentralised monitoring systems and, when combined with survey data, can enable cost-effective and time-effective concurrent evaluations. Instead of commissioning an expensive standalone survey, evaluators can read data that the programme is already generating as it operates.

Longitudinal evaluation

Longitudinal evaluation follows the same individuals, households, or cohort across multiple points in time. Unlike a cross-sectional study, which takes a single snapshot, a longitudinal study repeatedly measures the same participants, usually over months or years. By following the same people, evaluators can observe change within a group rather than just comparing different groups, which makes it useful for assessing whether there is genuine causation between an intervention and an outcome.

This design is powerful for understanding the real impact of long-term programmes. Consider a workforce-training scheme that wants to know whether its graduates actually earned higher wages. A longitudinal approach assigns each participant a tracking identifier at intake and follows them through several survey waves, perhaps at six, twelve, and twenty-four months. The key strength is that each new wave is compared to the same person’s earlier records, so a rise, stall, or drop is visible for that individual rather than buried in a group average. A famous example outside the development sector is the Framingham Heart Study, which began in 1948 and has followed generations of participants to establish links between lifestyle factors and heart disease.

The trade-off is cost and complexity. Longitudinal evaluation requires sustained funding, careful participant tracking, and strategies to deal with attrition as people drop out over the years. There is growing recognition in India’s evaluation community of the value here: a NITI Aayog compendium of essays highlights how monitoring instruments can be pivoted into evaluation instruments with a longitudinal focus, turning routine data collection into a tool for long-term impact measurement.

Putting the types together

These categories are not mutually exclusive; they are different lenses on the same activity. A single national scheme might use ongoing internal monitoring for daily management, commission a concurrent evaluation every year for accountability, undergo an independent external terminal evaluation at closure, and be revisited through an ex-post or longitudinal study years later to confirm whether its benefits endured. Each type answers a question the others cannot. Choosing the right combination, matched to the programme’s timeline, budget, and the decisions that depend on it, is what separates evaluation as a box-ticking ritual from evaluation as a genuine tool for better policy.

What do you think? If you had to evaluate a government scheme in your own city or district, which type would you prioritise first, and why? And do you think the push for mandatory independent third-party evaluation makes programmes more accountable, or does it risk overlooking the insider knowledge that only internal evaluators hold?

References

PDF 📄

Types of Evaluation: From Formative to Longitudinal Approaches

Table of Contents

Evaluation by focus: formative and summative 🔗

Formative evaluation: learning while doing 🔗

Summative evaluation: judging the final result 🔗

Evaluation by agency and stage 🔗

Internal versus external evaluation 🔗

Stages of evaluation: ongoing, terminal, and ex-post 🔗

Specialised evaluations for long-term tracking 🔗

Concurrent evaluation 🔗

Longitudinal evaluation 🔗

Putting the types together 🔗