Choices: Small Shifts, Big Stakes
Interpreting consumer choices through business impact, not just significance tests
In survey research, the common reflex is clear. Run the numbers. Check the p-value. Did the difference clear the 95% threshold? Was it “real”, or just noise? That reflex - the data-reflex - works well in many areas of research. Trackers, ATUs, segmentations, and driver analyses often rely on reporting, supported by means or percentages from survey data. In these contexts, significance testing is standard and expected. But conjoint results are different.