People may behave differently when they know they are taking part in research. A participant might read instructions more carefully, persist with a difficult task or follow a workplace procedure they would usually shorten. The Hawthorne effect is commonly used to describe changes associated with being observed or studied, although the term covers several possible mechanisms rather than one predictable response.

For a researcher, the practical question is what the study situation may have changed and whether that change matters to the conclusion. Calling something a Hawthorne effect does not establish its cause, size or direction.

The familiar story is less settled than it sounds

The name comes from research at Western Electric’s Hawthorne Works near Chicago in the 1920s and 1930s. A widely repeated account says productivity rose whenever lighting changed because workers received attention. Steven Levitt and John List’s reanalysis of the illumination experiments challenged that simple reading of the original data.

The broader question of research participation remains relevant. A systematic review by McCambridge, Witton and Elbourne found evidence that participation can affect behaviour, while emphasising uncertainty about the mechanisms, conditions and magnitude. That is a more defensible basis for practice than treating every observation as distorted by the same fixed effect.

Identify what may be different in your study

A fictional team researching a workplace scheduling tool might notice that participants carefully review every assignment while being watched. In ordinary use, they may be interrupted, under time pressure or already familiar with the schedule. The research setting has changed more than the presence of an observer, so the team should not attribute the difference to attention alone.

Examine the task, incentives, environment and introduction. Is the participant trying to demonstrate competence? Does the task encourage unusually thorough reading? Has the team explained what it hopes to improve? These questions connect a general concern to features of the actual study that can be reviewed.

Moderator bias concerns the influence of the researcher’s conduct, while response bias concerns how answers are produced. They can overlap with participation effects, but the distinctions help identify an appropriate response.

Reduce avoidable influence and preserve the limits

Use realistic tasks, neutral instructions and a clear explanation that difficulties with the design are useful to understand. Record interventions and relevant differences from ordinary use. Giving participants time to settle may help the session, but there is no universal warm-up period after which observation becomes unaffected.

Unmoderated testing removes a live researcher’s reactions, yet participants still follow study instructions and may know they are recorded. It should not be presented as a guaranteed cure. Compare findings with other appropriately collected evidence where the decision requires it, including behaviour outside assigned tasks.

Keep consent and privacy arrangements intact. Concealing observation to obtain supposedly pure behaviour introduces a different and serious problem; it is not a routine methodological fix. A useful report describes the study conditions and explains how they may limit transfer to ordinary use, rather than asserting that the Hawthorne effect has been eliminated.