When I teach scientific writing (and all evaluators need to be able to communicate clearly verbally and in writing), I focus on the 5Cs: 
larity, 
oherence, 
onciseness, 
onsistency, and 
orrectness, I’ve written about the 5Cs in a previous blog post, so I won’t belabor them here. Suffice it to say that when I read a document that violates one (or more) of these 5Cs, I have to wonder.
Recently, I was reading a document where the author used design (first), then method, then approach. In reading the context, I think (not being able to clarify) that the author was referring to the same thing–a method and used these different words in an effort to make the reading more entertaining where all it did was cause obfuscation, violating 
larity, one of the 5Cs .
So I’ll ask you, reader. Are these different? What makes them different? Should they have been used interchangeably in the document? I went to my favorite thesaurus of evaluation terms (Scriven)
(published by Sage) to see what he had to say, if anything. Only “design” was listed and the definition said, “…process of stipulating the investigatory procedures to be followed in doing a certain evaluation…” OK–investigatory procedure.
So, I’m going to list several terms used commonly in evaluation and research. Think about what each is–design, method, approach. I’ll provide my answers next week. Let me know what you think each of the following is:
Case Study
Pretest-Posttest Control Group
Ethnography
Interpretive
Naturalistic
Random Control Trials (RCT)
Quasi-Experimental
Qualitative
Focus Group
Needs Assessment
Thanks for this useful post. In the BetterEvaluation Rainbow Framework (www.betterevaluation.org) we distinguish between methods and approaches.
Methods are individual strategies or tools that can be used in many different types of evaluation. Classic example is surveys and interviews. they are both data collection methods than can be used in almost any kind of evaluation. But there are also different sampling methods, different analysis methods that one can use.
Approaches combine multiple methods within in a very intentional framework or logic. E.g. RCTs can involve surveys and all kinds of other data collection methods. It can involve random sampling methods and it can also involve counter-factual analysis methods. But it packages these in a particular way and together they make up an approach.
As for design, we don’t define this in the Rainbow Framework. But I would say that your evaluation design is the result of the all the choices you make about methods and approaches in response to the context and purpose. Every evaluation has a different design and very rarely is a design copied exactly (unless it’s as part of a replication study).
So to answer your question, this is how I would define these items…
Case Study – approach
Pretest-Posttest Control Group – method
Ethnography – approach
Investigatory – investigatory what? this is more of a purpose
Naturalistic – naturalistic what? ditto
Random Control Trials (RCT) – approach
Quasi-Experimental – a family of methods (e.g. difference in difference, regression discontinuity, propensity score matching…)
Qualitative – a type of data, distinguished from quantitative
Focus Group – method
Needs Assessment – method