Demand Characteristics

Transfers

Demand characteristics describe the totality of cues in an experimental setting that communicate the researcher’s hypothesis to the subject, enabling the subject to adjust their behavior to confirm (or sometimes disconfirm) that hypothesis. The concept, introduced by Martin Orne in 1962, reframes the experimental subject from a passive responder into an active interpreter of the social situation.

Key structural parallels:

• The subject as detective — Orne’s central insight is that human subjects do not simply react to experimental stimuli; they try to figure out what the experiment is about. Every element of the experimental situation — the instructions, the setting, the demeanor of the experimenter, the structure of the tasks — provides clues. Subjects assemble these clues into a hypothesis about the hypothesis, then behave in ways that confirm it. The experiment is not a one-way observation; it is a social negotiation.
• The “good subject” role — most experimental subjects adopt a cooperative stance. They want to be helpful, to produce useful data, to validate the experimenter’s time and effort. This social motive biases them toward confirming the perceived hypothesis. When a subject in a hypnosis study is asked to do something unusual, they may comply not because they are hypnotized but because the experimental context defines compliance as the expected behavior. The demand characteristics of the situation produce the result the experiment was designed to test.
• Leaky instruments — the concept transfers beyond the laboratory. Any measurement of human behavior that reveals its purpose to the measured risks contamination. Employee satisfaction surveys where management announces “we want to know if the new policy is working” are measuring compliance with the implied expectation, not satisfaction. User testing where participants are told “we designed this feature to be intuitive” will produce reports of intuitiveness. The measurement leaks its own answer.
• The experimenter as stimulus — demand characteristics are not just in the written protocol; they are embodied in the experimenter. Subtle cues — tone of voice, facial expressions, body language when the subject gives the “right” answer — shape behavior without either party being fully aware. Rosenthal’s experimenter expectancy research showed that even when protocols are standardized, the experimenter’s expectations leak through nonverbal channels.

Limits

• Not all subjects play along — Orne himself identified multiple subject roles. The “good subject” confirms the hypothesis; the “negativistic subject” deliberately disconfirms it; the “apprehensive subject” is preoccupied with self-presentation rather than the hypothesis; the “faithful subject” genuinely tries to respond naturally. The demand characteristics model predicts systematic bias, but actual subject pools contain a mixture of orientations, partially canceling out the effect. The model overpredicts the uniformity of the problem.
• Modern methodology has countermeasures — double-blind designs, cover stories, deception protocols, between-subjects designs (where no subject sees both conditions), and implicit measures (reaction times, physiological responses) all reduce demand characteristics. The model was most powerful as a critique of mid-20th-century psychology when single-blind within-subjects designs were common. Its relevance has diminished as methodology has improved, though it has not disappeared.
• The model assumes a single hypothesis to detect — in complex experimental designs with multiple conditions, interactions, and mediators, the “demand” is not a single signal but a cloud of contradictory cues. Subjects’ hypotheses about these designs are typically wrong, which paradoxically makes the design more robust to demand effects. The model is most concerning for simple, transparent designs and least concerning for complex ones.
• Generalizing to non-experimental settings is slippery — invoking “demand characteristics” to explain behavior outside the laboratory (in interviews, in therapy, in organizational surveys) stretches the concept beyond its original scope. In these settings, the “demand” is the entire social context, and calling it a confound rather than a legitimate social influence becomes a judgment call rather than a methodological diagnosis.

Expressions

• “The subjects figured out what we were testing” — the researcher’s post-hoc recognition that demand characteristics may have contaminated results
• “What is this experiment really about?” — the question subjects ask themselves (and sometimes each other) that drives hypothesis-guessing
• “Demand effects” — the shorthand in methods sections, usually invoked as a limitation rather than a central finding
• “The experimenter expectancy effect” — Rosenthal’s complementary concept, focusing on the experimenter’s role in transmitting demands
• “Social desirability bias” — the related tendency to respond in ways that present oneself favorably, which amplifies demand characteristics when the “demanded” response is also the socially desirable one
• “What did you think this study was about?” — the post-experimental inquiry (funnel debriefing) used to assess whether demand characteristics operated

Origin Story

Martin Orne introduced the concept in his 1962 paper “On the Social Psychology of the Psychological Experiment,” which reframed the experiment as a social situation rather than a neutral observation window. Orne was a hypnosis researcher who noticed that subjects in hypnosis experiments performed extraordinary feats — not because they were in altered states but because the experimental context defined those feats as expected. His “quasi-control” methodology (asking simulators to fake the experimental condition) became a standard tool for separating genuine effects from demand artifacts.

The concept intersected with Robert Rosenthal’s contemporaneous work on experimenter expectancy effects (1963-1966), creating a powerful critique of psychological methodology. Together, Orne and Rosenthal demonstrated that the psychological experiment was not a transparent window onto the mind but a social interaction contaminated by the expectations of both parties. This critique catalyzed the methodological reforms of the 1970s and 1980s, including widespread adoption of double-blind designs, standardized protocols, and deception-based cover stories.

References

• Orne, M.T. “On the Social Psychology of the Psychological Experiment” (1962) — American Psychologist, the foundational paper
• Rosenthal, R. Experimenter Effects in Behavioral Research (1966) — the complementary work on experimenter expectancy
• Weber, S.J. & Cook, T.D. “Subject Effects in Laboratory Research” (1972) — Psychological Bulletin, a review distinguishing the multiple subject roles
• Nichols, A.L. & Maner, J.K. “The Good-Subject Effect” (2008) — Perspectives on Psychological Science, a modern reassessment