Evaluation and Action Research Methodologies
Two of the source's qualitative methodologies put the researcher inside the situation being studied. One redistributes authority over the findings; the other gives up control of the variables. Both trade rigour of control for relevance.
Two entries that put the researcher inside the situation
The supplied teaching source lists seven research methodologies. Evaluation and action sit at the end of that list, and they share a feature the other five do not have: in both, the researcher is part of the setting being studied rather than an observer of it.
That brevity is the reason this page spends more time on consequences than on description. Each definition contains a clause that changes how the research must be run, and neither clause is developed anywhere else in the supplied material. The full set is mapped in Research Methodologies: An Overview.
"Equal partners" and what it actually commits you to
The evaluation definition names four stages at which participants are equal partners: design, implementation, interpretation and results. Read that as a list of four places where you give up sole authority, not as a statement of good intentions.
The four partnership stages, and the authority each one transfers
Participants shape the questions
If participants are partners in design, they can change what is asked and what is left out. A design fixed before they are involved and then shown to them for comment is consultation, not partnership.
Participants share the running of it
Partnership at this stage means participants have a role in how data is gathered and from whom, not merely a role as the people it is gathered from.
Participants can contest your reading
This is the hardest of the four. If participants disagree with what you conclude from the data, partnership means the disagreement has to be resolved or reported, not overruled.
Participants have standing in the output
Partnership in results raises questions of what gets published, in whose words, and who may veto. Settle those before collection, not after a difficult finding appears.
For a project manager researching their own organisation, the fourth stage is where the methodology bites hardest. Your participants are usually colleagues, sometimes reports, occasionally your sponsor. Equal partnership in results means agreeing in advance what happens when the finding is unwelcome.
Evaluation inherits the weaknesses of descriptive research
The source classifies evaluation as a form of descriptive research. That classification carries the descriptive entry's warnings with it, and those warnings are explicit.
The source says descriptive methodology relies on observations recorded and subsequently analysed, depends on human responses and observations, and "can be distorted by selective observation and biased interviews". An evaluation study is therefore exposed to both distortions by definition. The detail is in Descriptive Research Methodology.
Partnership makes this sharper rather than safer. Participants who helped design the study have their own stake in what it finds, and the researcher who works alongside them has a stake too. Neither of those is a reason to avoid the methodology; both are reasons to document how observations were selected and how interview material was handled.
There is also a definitional loose end worth knowing about. The source's terminology table defines evaluative research as seeking "to test policy/management outcomes against benchmarks", but the methodology entry for evaluation does not mention benchmarks at all. If you are evaluating against a standard, say which of the two framings you are working to. Both are set out in Descriptive, Explanatory and Evaluative Research.
Action research and the loss of control over variables
The action research definition has three parts: the real world, a specific problem in a particular situation, and little or no control over independent variables. The third part is the constraint that determines everything you may claim.
What the constraint costs you
- You cannot attribute an outcome to your intervention with confidence, because other conditions moved too
- You cannot replicate the study, since the situation was particular
- You cannot generalise to other settings without a separate argument for why they are comparable
- You cannot pre-specify everything, because the situation will change under you
What the constraint buys you
- Access to a problem as it actually presents, not a simplified version of it
- Data from the setting where the decision will be made
- A result that has already been tested against real constraints
- Findings the organisation can act on, because they came from its own situation
This is the mirror image of the source's experimental entry, which requires that every relevant condition is isolated and controlled. The same table therefore contains both an idealised control condition and a plain statement that real-world work does not have it. That contrast is worked through in Correlation and Experimental Methodologies.
The honest response is not to apologise for the constraint but to state it, early and precisely. Name the conditions that moved during your study, say when they moved, and describe what that does to your conclusions. A limitations paragraph written from that list will do more for the credibility of the work than any methodological vocabulary.
Choosing between them
EVALUATION AND ACTION RESEARCH SIDE BY SIDE
| Evaluation | Action | |
|---|---|---|
| Type in the source | Qualitative | Qualitative |
| What it is designed for | Complex social issues | A specific problem in a particular situation |
| Relationship to participants | Equal partners across design, implementation, interpretation and results | Not specified by the source |
| Stated constraint | Not specified beyond its descriptive classification | Little or no control over independent variables |
| Classified as | A form of descriptive research | Research carried out in the real world |
| Strongest claim available | An account of the issue and its outcomes, arrived at jointly | An account of what happened when this problem was addressed here |
Every cell reflects what the supplied source states or leaves unstated. Where a cell says "not specified", that is a gap in the source, not a property of the methodology as treated in the wider literature.
Which entry describes your study
The wider selection procedure, including the source's own advice on how to choose, is in Choosing a Research Methodology.
Setting either one up inside your own organisation
A working sequence
Write down who the participants are and what standing they will have
Name the group, and state which of the four partnership stages they will actually hold authority over. Vague answers here become disputes later.
Agree the rules for disagreement before you collect anything
Decide now what happens if participants read the data differently from you: joint resolution, or both readings reported. Put it in writing.
Record the baseline conditions
List the conditions in the setting at the start — staffing, scope, priorities, external pressures. You cannot control them, but you can date them and show what changed.
Keep a dated log of changes you did not choose
Every reorganisation, budget move or personnel change that lands during the study belongs in this log. It is the raw material for your limitations section.
Separate your practitioner role from your researcher role in the record
When you act to solve the problem, note it as an intervention. When you observe, note it as data. The two are easy to blur when you are doing both.
Run the ethics screening before, not after
The source asks whether the project needs access to confidential or personal information, whether the questions respect cultural norms, and whether collection methods are unbiased and ethical. Colleagues as participants raise all three.
The source does not prescribe this sequence. It is assembled from the constraints its own two definitions impose, plus the screening questions it gives elsewhere; see Step 6: Seeking Ethics Clearance.
How to tell you have got it right
Evidence a reviewer can check
- You can name at least one thing participants changed about the design
- There is a written rule for what happens when interpretations conflict
- The conditions you could not control are listed and dated, not summarised as "various factors"
- Your claims are about this situation, unless you have argued separately for transfer
- The distinction between your actions as practitioner and your observations as researcher is visible in the record
- Nothing in the conclusion asserts that your intervention caused an outcome
What to carry forward
- Evaluation and action research are both labelled qualitative, and both place the researcher inside the situation being studied.
- Equal partnership spans design, implementation, interpretation and results. Claim it only where participants held real authority at those stages.
- Evaluation is classified as descriptive research, so it carries the descriptive entry's exposure to selective observation and biased interviews.
- Little or no control over independent variables is a statement about the setting. Log the conditions that moved and let the limitations section do its work.
- The supplied source gives no procedure, no cycle and no partnership protocol for either methodology; take those from a cited methods text and say which.
Frequently asked questions
What is the difference between evaluation research and action research?
In the supplied source, evaluation is descriptive research designed for complex social issues in which participants are equal partners in design, implementation, interpretation and results. Action research is carried out in the real world to address a specific problem in a particular situation, with little or no control over independent variables. Evaluation is defined by the partnership; action research is defined by the loss of control.
Does "equal partners" mean participants can veto my findings?
The source does not say. It states that participants are equal partners in design, implementation, interpretation and results, but gives no protocol for resolving disagreement. Because the source is silent, you must settle the question yourself before collection begins and record what you agreed.
Can action research support a causal claim?
Not on its own. The definition concedes little or no control over independent variables, which means other conditions were free to move while you intervened. You can report what happened and what you observed about why, but attributing the outcome to your intervention requires the kind of control the source's experimental entry describes.
I am researching my own team. Is that a problem?
It is a condition to manage, not a disqualification, and both methodologies assume proximity to the setting. The practical issues are the descriptive entry's warning about selective observation and biased interviews, and the ethics questions the source raises about confidential information and unbiased collection. Keep the practitioner role and the researcher role separately recorded.
How many action research cycles should I run?
The supplied source says nothing about cycles or iterations. It defines action research in a single sentence and offers no procedure. Any cycle structure you use must come from a methods text you cite, and your methodology chapter should say which one.
References and source attribution
- Clarke, R. J. 2005, Research methods and methodologies, pp. 35–45 — the work the supplied teaching source adapts for its table of seven research methodologies.
- Bryman, A. 2016, Social Research Methods, 5th ed., Oxford University Press, Oxford.
- O'Leary, Z. 2017, The Essential Guide to Doing Your Research Project, 3rd ed., Sage Publications, London.
- An edited volume of master's degree research projects in project management, 2018, academic publisher, 378 pp. — the source of the illustrative research design referred to on this page.
- The supplied teaching source: consolidated week 2 and week 3 teaching notes and slide material on research methods, used here as the basis for all statements attributed to "the source".
Suggested questions for Ask KEVOS
- Draft a partnership agreement covering design, interpretation and publication for an evaluation study.
- Help me write the limitations paragraph for an action research study run during a reorganisation.
- Is my study evaluation research or descriptive research with a member check?
- What does the supplied source say about evaluation and action research, word for word?
- List the conditions I should log when I have no control over independent variables.