From a claim about consumers to the evidence for it

Students arrive with strong intuitions about consumers — usually correct in direction and vague in every other respect. My courses are built around turning those intuitions into claims that can be measured, and then asking what design and what data would actually support them.

In Media and Consumers, for example, students work through the theory on social media before going to the platforms themselves — collecting comments, then checking what the theory led them to expect against what they actually find.

I try to teach established methods and newer ones side by side, so the syllabus moves as the field does. The current addition is LLM-based simulation of consumers, which students examine rather than adopt: they generate synthetic responses, hold them against real survey data from the same population, and work out where the resemblance holds and where it quietly fails — which segments get flattened, where a stereotype stands in for variation, and what a simulated respondent cannot be evidence of.

That the tool is new is not the point. The point is that a new tool answers to the same question as every other method on the syllabus: what would make this a credible measurement?