Observation research
Observation research is the systematic watching and recording of what people do, in their natural settings or in controlled ones, rather than relying on what they report. It captures behavior directly, including the habits and workarounds people never think to mention.
How it works
Observation varies along two lines. It can be naturalistic, with the researcher watching behavior as it happens without intervening, or controlled, in a lab or test setting. And the researcher can be a non-participant observer or a participant who takes part in the activity. Structured observation counts predefined behaviors; unstructured observation records whatever seems relevant in field notes.
Its main strength is validity, since people's actions often differ from their accounts. Its weaknesses are the time it takes to record and code what was seen, the chance that people act differently because they know they are being watched, and observer bias, which is why coders are kept blind to a study's hypothesis where possible.
A worked example
Store staff at an electronics retailer report few problems with a new self-service returns kiosk. Researchers who observe 40 customers at the kiosk see 23 of them ask a nearby employee for help at the same screen, a step staff had absorbed into their routine and never mentioned.
Related
Automate this in Pulse
An observer's tally of what people did, the detours and the skipped steps, can set the areas of a Pulse brief, and interviews then ask participants to account for the behaviors the observation counted.