Research Methods · Simplified
Operationalisation Laboratory
Change how a construct is measured, and watch two things move in opposite directions.
The laboratory
One construct, four ways to measure it
Academic engagement is not directly observable, so it has to be operationalised. Add and remove measures, and watch two things: what the plan reaches, and what it drags in with it.
Your measurement plan
What else this plan records
Everything below is recorded by the plan without being academic engagement.
Coverage of the five facets
Try a couple of different plans first.
What this shows
Two ways a measure goes wrong, pulling in opposite directions
An operational definition is a claim about what stands in for the construct. Every claim of that kind can fail twice over, and fixing one failure usually worsens the other.
Key idea: An operational definition is a claim, not a translation. Adding measures to reach more of the construct also adds more that is not the construct. And a facet is only ever reached as well as the best single measure reaches it. Two measures that each reach it partly give two partial views, not one good one, which is why a third measure can cost you something and buy you nothing. Even all four together leave the agentic facet only partly reached, so no plan here scores more than 9 of 10. Combining measures buys coverage. It does not buy the construct.
The five facets are a teaching model, not a fact: other accounts of engagement use three facets, or four, or argue that engagement is a property of a situation rather than of a student. The coverage figure is arithmetic over a facet list somebody wrote down. It is not a validity coefficient and must not be read as one. The measures are described generically; no published instrument or item is reproduced.
The longer version adds three constructs, six measures each, reactivity and feasibility ratings, a comparison table and a challenge on measures that change what they measure. It is at Operationalisation Laboratory in the main collection.