How do you operationalise a cognitive process in a psychology study?

To operationalise a cognitive process, set out tasks you can observe and rules for scoring them that represent the process you want to study. A score provides evidence about the construct, meaning the process or ability being investigated; it is not the construct itself.

Suppose the target is verbal working memory, defined here as keeping verbal information available while using it or carrying out other mental operations. The following are two proposed tasks, not results from an existing experiment.

Two ways to measure it

If someone gives four of five words in their correct positions, that trial scores $4/5=0.80$. Specify in advance how omissions, repeated words and spelling errors are treated. Also decide how to handle trials where a participant appears to skip sentence processing; otherwise a high recall score may reflect ignoring half the task.

Task B follows the logic of complex-span measures, which combine remembering items with another mental task. Daneman and Carpenter's original 1980 reading-span task combined reading sentences with remembering their final words. The proposed task above uses the same broad idea; it does not reproduce their procedure or scoring system.

Evaluate the operational definitions

Construct validity asks whether a task measures the process it is meant to measure. Task B covers our definition of both holding and using information more directly. However, weak performance might reflect language difficulties rather than limited verbal working memory. Task A offers a simpler comparison but cannot, alone, justify broad conclusions about complex reasoning.

Reliability concerns how consistently a task measures something. Keep presentation timing, instructions and scoring consistent, and use several equivalent lists. Consistent measurement is desirable, but a task that is biased in the same way each time can still have poor validity.

Alternative explanations: Check familiarity with the reading material, whether participants can see or hear it clearly, fatigue and strategies. Counterbalance task order by having different participants do the tasks in different orders, and use different item lists to reduce practice effects. A five-item task could also be too easy or too hard for a particular group, hiding differences between participants.

Agreement across suitably different measures strengthens an interpretation. Disagreement can be informative too: it may reveal that the tasks measure different parts of the process. Keep your conclusions within what the measures support, rather than turning one task score into a claim about a person's overall intelligence.

Build on this comparison with the cognitive-model and measurement essay in the Psychology cognitive-approach questions.