A big part of collecting the right information is knowing what to expect, having realistic expectations about the kind of information we're going to get. And that information is information that comes from our stakeholders. So we're going to look at a concept here which we call the four levels of evaluation.

The highest level of evaluation is effectiveness. What has changed as a consequence of the project achieving its outcome? Below that we have change in behavior, change in practices, skills applied. Below that we have learning and knowledge, skills transferred, systems in operation, goods and services delivered. And at the foundation level of this we have reactions and feelings. How do our end users and other stakeholders feel about what we're doing? Are they positive? Are they supportive? Do they see it as relevant?

Now you're not going to see this kind of information every time. You see these things as time goes by.

• So, initially, in the early stages of a project you're only going to be able to get reactions and feelings. How do your end users feel about what you're doing? Is it relevant to their needs? Are they motivated? Are they positive? Are they involved? And so on.
• Over time, as time passes, we will start to see learning and knowledge, skills, systems in place, goods and services being delivered, very much equivalent to our outputs.
• If those are applied then we start to see the change in behavior- people doing things differently. And I'll give a couple of examples shortly.
• And at a much, much later stage, sometimes beyond the end of the project, we may see the effectiveness.

Let's start with reactions and feelings. That answers the question, are our stakeholders happy? Now, happy stakeholders doesn't mean the project's having an impact, but it does mean that the work we're doing is appreciated and it's understood. That means it's perceived to be relevant. Positive reactions can tell us that, at least from the stakeholder's perspective, the plan is meeting their perceived needs. Positive reactions from stakeholders also suggest that we're communicating well, that our stakeholder management plan or stakeholder engagement plan is effective.

Stakeholders can help us reflect on our methods - how we're delivering the services, not just what we're delivering, but the how. Stakeholder satisfaction is a big factor in a project being successful. Unless they see it as relevant and useful and appropriate, then you're not going to get a lot of buy-in. So tracking stakeholder satisfaction really matters.

And that data is always available to us. That's the foundation level and we will always collect that information whenever we can with each time that we're reporting.

Next we're asking, "how successful have we been in transferring knowledge and skill to the target group?" And that is necessary if we want to see any change in behavior, any change in practice, then firstly we have to transfer knowledge and skills. And that can be seen quite early in the project.

As time goes by, we'll start to see change in behavior as those skills are applied, as people start doing things differently. And here this asks the question, "are end users applying what they have learned? And how are they doing things differently?" And this very much is connected to the outcome level of the project or the project purpose or, in EU language, the specific objective, which is what the project promises to deliver by its end.

Of course, we may not see this until a much later stage of the project. So don't expect to see these things in the earlier phases.

And at the effectiveness / impact level, what we're going to see is the overall change we're aiming for. It tells us whether that project plan, the rationale, was relevant, and it represents more of a shared goal with other agencies and governments and so on.

So we're looking at the impact level here for which your project is not fully responsible, and it's quite hard to attribute impact-level success to a specific project. But it does tell us whether the plan or the intervention logic was appropriate, whether it did make a difference.

And we're looking at things here such as health and livelihood and rights., "bigger picture" things that are shared between our organization and other development partners. And we won't see this until the end of the project or possibly later beyond the project's end.

So the four levels of evaluation are:

• Reactions and feelings.
• Learning and knowledge.
• Change in behavior.
• Effectiveness.

So for example, if we are working in an agricultural project, the first thing we want is that the farmers themselves are positive about the work we're doing. Building on this, we're going to see over time skills training, new planting techniques or new crops introduced, new tools and processes learned.

And if we start to see that behavior, we're going to see people, see end users doing things differently - line-planting, post-harvest agricultural processing, and so on. And at the effectiveness level, that's going to evaluate whether we had the intended change. Did we increase livelihoods or improve nutrition?

So those four levels reflect our logical framework. The activities, the outputs, the outcome or specific objective, and the impact, the overall objective. And over time, depending on where you are in your project cycle, you will get information on more levels.

Let's consider a project which has the ultimate aim of improving health, reducing HIV among commercial sex workers. So they are our end users and the first thing we want to do is measure the reactions and feelings and we consistently do that. It reflects on our activities. Are we carrying them out in a way that is understood, appreciated, perceived as relevant? Because if they're not received well, we won't see much uptake or participation.

Now, the learning and knowledge, they're things we can test. They're very much reflecting the outputs, those immediate early results: knowledge on HIV transmission, knowledge on prevention, legal protection, rights, assertiveness training, etc. So those are outputs.

At the behavior level, we're looking at the change in behavior: more reported use of condoms, for example, and that's what the project wants to deliver. And at the effectiveness level is where we may expect to see the health impact.