Saturday, May 8, 2010
This post is going to involve a lot of blathering on about abstract components of learning science research and some of my simplifications might be downright wrong and/or controversial. Read at your own risk ;)
I would like to apologize for the below ridiculous figure that I created to describe (in a simplified fashion) what the field of learning science research "looks like" for a class on "Design Perspectives of HCI." It's as much to structure my own understanding as anyone else's.
Stated as simply as I can: learning science research consists of choosing a perspective (granularity), a research method perspective (i.e., a mode of theory construction), and an experimental design/approach to create theory and evaluate products that work. None of the choices in these three levels necessitates the choice of another (that is, assuming this simplified model is correct, there's 8 different ways to "do" research). Although, this number '8' gets confusing when you realize you can use more than one research method in a single study. Anyways, sociocultural research can be performed with quantitative or qualitative research methods, using psychological experimental designs or design-based research designs. This does not mean there are not particular preferences for certain methodological approaches in certain perspectives. It just means it's possible to do a whole mish-mash of study designs.
Typically, you might see this ability to use multiple methods/approaches/perspectives used as a sort of secondary way to bolster the work of a preferred or primary approach to inquiry. But often people have strong beliefs about which methods and approaches produce end-results on their own, and which methods are merely supplementary. (In my community you often see quantitative researchers using design-based and qualitative methods as a way of test-piloting interventions to inform a final experimental/controlled study).
The figure above tries to visually show that psychological experimental design and design-based research share some shallow common design features, but that in the end, they are two completely different beasts. On the opposite end, oftentimes, "straw man" distinctions are made between the two approaches, such as in the [*ahem* slightly modified] table below:
So, just how are classical psychological experimental design and design-based research related? It seems that the design-based research approach (1) emphasizes preserving complexity (and not simplifying) in the situation and (2) prefers an interpretive (and not statistically-causal) reasoning for modifying the intervention. On the other hand, psychological experimentation typically must simplify the situation (i.e., control variables) in order to make the strong statistically-causal explanations they prefer. Whether or not the study takes place in the real classroom setting is not what makes the research experimental or design-based. It is instead the emphasis on complexity and interpretation of results.
There is another typical feature of design-based research that I'm currently ignoring, because I am not sure it is necessary to design-based research; I fail to believe that iteratively changing the intervention design, before the "end" of the study, without statistical causal analysis is a required component for design-based research. Maybe if I spun it positively it would be more acceptable?
This work sprung from my current training in a cognitive-psychological-experiment community, and how it seems to me that members of that community sometimes don't understand it, leading them to slightly look down upon it. I think a lot can be gained from design-based research, even without statistically-causal claims (the literature sometimes points to Ann Brown's reciprocal teaching, or even the FIRST Robotics League as examples of non-psychological-experiment successes). That is, design-based research can be a valid method of theory construction and inquiry, even without the help of causal statistics to "back it up"; however, I'm still not sure how to convince the rest of the community of this.
0 comments:
Post a Comment