According to the State of Texas, Office of the State Auditor, (A Guide to Measure the Performance of Information Systems, April 1991) the following items are critical to successful implementation of a performance management system (similar issues apply to implementing a M&E system):

• Management support and involvement from all levels in the organization to counteract resistance to change associated with introduction of new policies.
• Appropriate measures that include defining the goals and objectives, identifying the measures to be used, evaluating the costs and benefits, and then implementing the cost-effective measures
• Start small and measure the processes, not the people.  Fewer measures mean less initial cost; measures can be added.  Measure how things are done, and the result, not the people.
• Provide feedback by using the results and reporting where improvements are ...

M&E is not an end in itself.  It should serve a means...a means to learning and being accountable. 

So, unless M&E processes/tools render information that is actually used by someone to further these ends, they are an unethical waste of resources.

So what does 'utilisable' M&E information actually mean?

I suspect that there are at least fourcontributing factors:

• Relevant: does the information fundamentally meet a need?  Is it perceived as important/valuable by someone (an information 'client')?
• Accurate: can the information be trusted? 
• Timely: is the information delivered (to the information 'client') within a timeframe that permits action to be taken? 
• Accessible: is the information provid ...

While the concept of organisational learning has been popularised within the international aid industry in recent years, the persistent challenge is to ground the concept.  In practical terms, what is learning?  What actual mechanisms can be employed to promote learning in a structured way?  How can learning be moved from a tacit process within individuals to a shared process among team members? 

One practical understanding of learning has been proposed by Gharajedaghi:

“Learning results from being surprised: detecting a mismatch between what was expected to happen and what actually did happen.  If one understands why the mismatch occurred (diagnosis) and is able to do things in a way that avoids a mismatch in the future (prescription), one has learned.”

The above quotation contains three practical mechanisms that an aid organisation can employ to operationalise the concept o ...

It is sometimes difficult to decide what data and/or methods should be used for M&E.  There seems to be so many perspectives and issues to consider.  I've found the following 4 points helpful to think through the range of issues, and as the basis seeking consensus among stakeholders:

• Reliability: Do the questions have the same meaning to the same respondents at different times?
• Validity: Do the answers tell us what we want to know? Do the questions have the same meaning to different respondents at the same time?
• Scalability: Can comparative analysis of the findings be carried out at all the required levels of dis-aggregation?
• Affordability: Can the method be implemented for reasonable cost?

Every now and then I observe robust discussion about the merit of quantiative vs qualitative methods for M&E.  Some people seem to feel passionately about one or the other.  The pragmatic reality that both approaches have merit in different contexts, and can complement eachother, is recognised by Patton (1997, p 266), who states that the debate about evaluation paradigms “has run out of intellectual steam".

My own experience suggests that the distinction between 'quantitative' and 'qualitative' is, in practice, a bit academic.  I will illustrate by way of example:

a)       A focus group question (i.e. a ‘qualitative method’) “how many households in this village?” would yield a ‘quantitative’ answer
b)       A survey question (i.e ...

Over the past few weeks I've been writing up an evaluation report.  Part of the evaluation involved a review of the M&E arranagements employed by four prominent international NGOs.  It was encouraging to find that all four agencies were committed (at least intellectually) to learning and improving aid effectiveness.  And it was encouraging that all had taken positive steps towards implementing M&E processes.  However, it was also disappointing to find that the efforts to implement an 'M&E system' seemed to post hoc.  They seemed to lack an overall 'information architecture' to give the M&E processes some coherence and purpose.

I may have developed an inadequate appreciation for the situation, but on reflection it seems that the focus of M&E remains very much on 'data' rathern than on 'knowledge'. 

As noted by Henri Poincaré, the great French ...

I've been periodically engaged by AusAID since 2005, working on the development of a quality framework for M&E.  In other words, a set of criteria to articulate what AusAID considers to be the basis for good quality M&E arrangements.  This framework and the associated good practice guide are now at final draft status within AusAID.

One of the 6 broad criteria for good quality M&E arrangements (and arguably the most challenging) requires that M&E practitioners ensure that M&E arrangements are both comprehensive and efficient.

On the surface, this seems perfectly reasonable.  But as with many things in life, a dig slightly below the surface reveals a surprising amount of complexity.

I'm just in the throws of writing up a report for AusAID; an evaluation of emergency response interventions in Pakistan following the earthquake in October 2005. 

The earthquake was 7.6 on the Richter scale and affected an area the size of Belgium.  An estimated 75,000 people were killed and over 3.5 million were rendered homeless...to say nothing of the widespread destruction of public infrastructure and services.

To visit the affected area is incredibly disturbing.

Part of the terms of reference of the evaluation was to examine the M&E of implementing agencies.  This has led me to reflect deeply on what is appropriate for an emergency.

Much of what was done by way of M&E during the emergency response appears to be ad hoc or informal/unstructured.  There appears to have been a focus on accountability (i.e. reporting to donors and powerful ...

There has been further discussion about my appraisal comments regarding the Cambodia rural development program discussed in my last blog.  The following is drawn from my email response, seeking to clarify the issues from my perspective...

The issues that you have raised in your feedback are a fractal of a broader debate taking placed within the field of Design, Monitoring and Evaluation.  I wholly agree with everything you have articulated in your last email, including the limitations of linear causality and the rigid ‘blue print’ (Fowler) approach to the design of social change interventions (for which the logframe is one prominent tool).  I agree with much of the critique of the logframe, including Chambers and others (e.g.Gasper, Smilie, Smutylo, Roche, Kaplan, den Heyer, Lavergne).  However, while agreeing completely with your line of argument, my conclusion is slightly different.

The heart of th ...

I've just completed an appraisal of a project design for an AusAID NGO Cooperation Agreement.  I found the design document to be well written, and in some ways innovative.

The design embraced the complexity and systemic nature of development challenges in rural Cambodia, and proposed an iterative/emergent approach to tackling these issues based on a form of action-learning cycle (GTZ's Participatory Extension Approach (PEA)). 

Although I appreciated the gist of the thinking behind the design, I raised the following questions:

• Is there a risk that adopting such an emergent/organic approach could be beyond the capacity of the implementation team on the ground?  Given the typically high workload associated with implementation, does this approach carry additional risk because of its inherent ambiguity? 
• Is the life of one Cooperation Agreement sufficient to enable iterations ...