In social psychology, demand characteristics refer to any inadvertent hints that lead participants to alter their natural behavior based upon their insights into the objectives of the experiment (Nichols & Maner, 2010).
They are aspects of a study that reveal what the purpose may be, thereby inadvertently causing participants to adjust their behaviors to confirm to those cues. For example, when conducting a survey about job satisfaction, if participants tend to smile or look pleased every time a positive question is asked, it might signal participants to give positive responses (Weathington, Cunningham & Pittenger, 2010).
Demand characteristics can greatly impact the reliability and validity of an experiment’s outcome.
Demand Characteristics Definition
Demand characteristics were first defined in the works of Martin Orne in 1959, and have been widely explored for over 50 years.
A few clear definitions of demand characteristics are provided below, starting with Orne’s definition himself:
- “[Demand characteristics are] the totality of cues which convey an experimental hypothesis to the subject” (Orne, 1962, p. 779).
- “[Demand characteristics] refers to participants being aware of what the researcher is trying to investigate, or anticipates finding, and what this implies for how participants are expected to behave.” (McCambridge, Bruin & Witton, 2012)
- “[Demand characteristics are] cues that make participants aware of what the experimenter expects to find or how participants are expected to behave” (Nichols & Maner, 2010)
Of all potential sources of bias in experimental research, demand characteristics have emerged as one of the most potent and pervasive.
This is because they can subtly influence participant behavior without their conscious knowledge. Take, for instance, experimenter bias, where the experimenter’s known or unknown expectations influence the subject’s behavior (Rosenthal, 2012).
In essence, the experimenter might give subconscious context cues about the desired behavior, influencing experimental outcomes (Creswell & Creswell, 2017).
Examples of Demand Characteristics
Demand characteristics involve cues about the research’s aims and hypothesis, such as rumors of the study, cues in the setting of the laboratory, and explicit or implicit communication with and between participants.
Below is an explanatio of each:
- Rumors of the study refer to any information about the study that participants may have heard before participating (Schwarz & Clore, 2016). These rumors could give participants a preconceived notion about the experiment’s purpose which could subsequently influence their behavior. Let’s take an example of a participant who heard that the study is about aggressiveness. This participant might consciously or subconsciously act less aggressive or more aggressive depending on their interpretation of the experiment’s aim (Nichols & Maner, 2010).
- The setting of the laboratory, or any experimental environment, can give cues to the participant about the desired behavior. For instance, if the laboratory environment is dominantly formal with an official-looking experimenter in a lab coat, participants might feel more anxious and try to behave ‘appropriately’, potentially skewing the results (Orne, 2010). Similarly, if the experiment is held in an informal setting, like a cafe or a park, participants may act more casually, affecting the way they respond to experimental tasks.
- Explicit or implicit communication, which refers to both verbal and non-verbal communication cues from the experimenter, can significantly influence a participant’s behavior. The tone of voice, body language, or facial expressions can offer hints about expected behavior (Rosenthal, 2012). For example, if an experimenter smiles or nods approvingly when a participant provides a certain type of response, the participant might continue providing similar responses believing it to be the ‘correct’ or desired response.
In essence, these elements can serve as cues triggering demand characteristics, altering participants’ natural behaviors in a way that could potentially impact the outcome of a study. Thus, it is crucial for researchers to control these aspects during the design and execution of an experiment.
How do Demand Characteristics Affect Research Participants?
The problem with demand characteristics is that it can affect the behaviors of research participants.
Weber and Cook (1972) identified four primary roles that participants can take on when they become aware of the study’s hypothesis through demand characteristics, as described below:
- The Good Subject: The first role is the “good subject.” A good subject tries to behave in a way that supports the experimenter’s hypothesis. Upon deciphering what the experiment is about, they aim to give responses that would advance the research (Weber & Cook, 1972). For example, in a study about the benefits of physical exercise on mental health, a participant who has inferred the study’s aim may overstate the positive effects of exercising on their mood to support the researcher’s hypothesis (Nichols & Maner, 2010).
- The Negative Subject: The second role is the “negative subject.” A negative subject behaves in a way that opposes or negates an experiment’s expected results. This contrarian approach can be seen as an attempt to sabotage the experiment purposefully – sometimes, this might be a reaction to feeling manipulated or in an effort to exert control (Weber & Cook, 1972).
- The Faithful Subject: The third role is the “faithful subject.” A faithful subject aims to follow the given instructions strictly without being influenced by any inferred expectations from the experiment. They aim to be honest and genuine in their responses, avoiding any manipulation of their behavior to fit specific outcomes (Weber & Cook, 1972).
- The Apprehensive Subject: The fourth role is the “apprehensive subject,” also known as “evaluation apprehension.” Apprehensive subjects are overly concerned about their performance and how they may be judged based on their responses. This apprehension causes them to alter their behavior towards what they perceive to be more socially desirable or acceptable (Weber & Cook, 1972).
These roles underline how the understanding of demand characteristics by participants can occasion different responses, emphasizing the need for stringent control of such aspects to ensure the validity of the research findings.
How to Control for Demand Characteristics
Controlling for demand characteristics is crucial in ensuring the collection of unbiased, reliable, and valid data in psychological research.
Here are a few common techniques employed by researchers:
- Use a double-blind procedure. This technique ensures that neither the participant nor the experimenter knows the purpose of the experiment or the expected outcome (Creswell & Creswell, 2017). As a consequence, it reduces the likelihood that expectations or demand characteristics influence the results.
- Use the deception technique. This involves providing participants with a misleading aim of the study or misleading the participants about the nature of the tasks (Athanassoulis & Wilson, 2009). However, such a method should always be conducted ethically, ensuring that no harm is caused to the participants.
- Participant anonymity. When participants are assured that their responses will remain anonymous, they are less likely to modify their behaviors or responses to fit perceived expectations (Saunders, Kitzinger & Kitzinger, 2015).
- Participant training. Engaging in thorough participant training prior to the experiment can also be helpful to counteract demand characteristics. This training would include detailed instructions on the task and encouragement for honest responses (Orne, 2010).
- Post-experiment inquiry. In post-experiment inquiry, participants are asked about their perception of the study’s aims. This can help identify whether participants guessed the purpose of the study or if they altered their responses based on demand characteristics (Orne, 2010).
Comparison to Other Types of Research Bias in Psychology
Demand Characteristics vs Hawthorne Effect
Demand characteristics are often confused with the Hawthorne effect. The main difference is that the Hawthorne effect refers to situations where people react to being observed, but they may not necessarily know the hypothesis of the researcher. By contrast, demand characteristics refers to the experimenter setup cues that reveal the researcher’s hypothesis.
The Hawthorne Effect refers to the change in behavior resulting from the awareness of being observed (Adair, 2014).
Named after a study by Western Electric at their Hawthorne Works factory in Chicago, the phenomenon concerns individuals modifying or improving their behavior in response to their awareness of being observed. Workers in the factory increased their productivity simply because they were aware they were under study
This effect, though typically connected with productivity and work environments, could be applicable to any situation where subjects are aware they are being watched (Levitt & List, 2011).
Both demand characteristics effects and the Hawthorne Effect affect the integrity and reliability of research findings. Demand characteristics can potentially lead participants to behave in line with perceived experimenter expectations and the Hawthorne effect can skew participant behavior based on their awareness of being observed.
However, their differentiation lies in the fact that demand characteristics emerge from experimental setup cues, while the Hawthorne effect is spurred by participants’ consciousness of being observed.
Demand Characteristics vs Social Desirability Bias
Similarly, people often confuse demand characteristics with social desirability bias. Both demand characteristics and social desirability bias influence the behavior of participants in research studies, yet they operate on different principles and cause different effects.
Demand characteristics, as defined earlier, involve cues in an experiment that may subtly suggest to participants the kind of behavior that is expected or desirable, influencing their responses (Nichols & Maner, 2010). This can inadvertently alter the results of a study, impacting its validity and reliability.
In the demand characteristics effect, participant’s perceptions of the researcher’s expectations can lead them to modify their behavior in a manner that conforms to these inferred expectations (Orne, 2010).
In contrast, social desirability bias refers to the tendency of participants to present themselves in a socially acceptable manner (Larson, 2019).
Participants may modify how they respond to research questions in order to appear more favorable or acceptable to others, regardless of cues given by the experimenter.
For instance, an individual may over-report good behavior or under-report bad behavior because they believe that it is the socially appropriate response.
The main difference between the two lies in the sources of the influences. Demand characteristics arise from specific cues and expectations identified by the participants within the research context, whereas social desirability bias is driven by participants’ broader social and self-presentational concerns, often stemming from societal norms and values.
Additional Research Bias Phenomena
- Selection Bias: This type of bias occurs when participants are not randomly selected for a study, causing the sample to be not representative of the target population (Swanson, 2019). It can influence the study’s results by over- or underestimating the effects measured due to the skewed sample. Therefore, using random selection methods is important to control selection bias.
- Confirmation Bias: Confirmation bias emerges when researchers interpret or seek information that confirms their existing hypotheses or beliefs (Nickerson, 1998). This can unjustly strengthen the desired hypothesis while ignoring evidence that might oppose it. It can be reduced by striving for objectivity and testing alternative hypotheses.
- Measurement Bias: This occurs when data is consistently overestimated or underestimated due to faulty measurement tools (Creswell & Creswell, 2017). When the tools used in a study are not accurate, valid, or reliable, the results will be skewed. Ensuring reliable and valid measurements is key to reducing this bias.
- Reporting Bias: Reporting bias takes place when researchers selectively report or exaggerate findings that are deemed significant or fit their expectations, while downplaying or ignoring insignificant findings (Adams, 2010). This can skew the reported findings, misrepresent the scientific truth, and misinform readers. Comprehensive reporting of all findings, regardless of their statistical significance, is essential in reducing this bias.
- Sampling Bias: Sampling bias occurs when a sample does not represent the entire population (Dattalo, 2010). If the sample is not representative, conclusions drawn from it may not be generalizable to the larger population. Random, stratified, or systematic sampling techniques can help reduce this bias.
- Observer Bias: In observer bias, the observer’s own experiences, expectations, or knowledge can influence their interpretation of the participants’ behavior (Rosenthal, 2012). This can lead to incorrect conclusions about the observed phenomena. Blind studies, where the observer doesn’t know the study’s hypothesis, can help avoid this bias.
- Publication Bias: Publication bias is the tendency for researchers and academic journals to publish positive or significant results over negative or neutral ones (Fanelli, 2012). This can skew the scientific literature and make certain outcomes seem more reliable than they truly are. Registration of all clinical trials and publishing all results can help alleviate this bias.
- Confirmation Bias: Confirmation bias (again) arises when researchers favor information that supports their preconceptions, disregarding contrary data (Nickerson, 1998). This can skew the results and interpretations of a study leading to false conclusions. Awareness and deliberate attempts to disconfirm hypotheses can help mitigate it.
- Funding Bias: Funding bias happens when the outcomes of a study are swayed by the entities funding the research (Lundh, 2017). This can lead to positive results for a product or service, compromising the integrity of the results. Transparency in funding sources and rigorous peer review can help control this bias.
- Recall Bias: Recall bias occurs when participants do not remember past events accurately, leading to inaccurate conclusions (Khare & Vedel, 2019). This can distort the results and interpretations of a study. The use of more objective measures and verification of information can help reduce this bias.
Demand characteristics, if left uncontrolled, can produce misleading results in psychological research. However, through appropriate experimental design and careful control measures, such as double-blind designs or subterfuge, these effects can be minimized ensuring the production of reliable and valid results.
Adair, J. G. (1984). The Hawthorne effect: a reconsideration of the methodological artifact. Journal of applied psychology, 69(2), 334. doi: https://psycnet.apa.org/doi/10.1037/0021-9010.69.2.334
Adams, R. J. (2010). Improving health outcomes with better patient understanding and education. Risk management and healthcare policy, 61-72.
Swanson, S. A. (2019). A practical guide to selection bias in instrumental variable analyses. Epidemiology, 30(3), 345-349.
Creswell, J. W., & Creswell, J. D. (2017). Research design: Qualitative, quantitative, and mixed methods approaches. Sage publications.
Dattalo, P. (2010). Ethical dilemmas in sampling. Journal of Social Work Values and Ethics, 7(1), 1-12.
Fanelli, D. (2012). Negative results are disappearing from most disciplines and countries. Scientometrics, 90(3), 891-904. Doi: https://doi.org/10.1007/s11192-011-0494-7
Khare, S. R., & Vedel, I. (2019). Recall bias and reduction measures: an example in primary health care service utilization. Family Practice, 36(5), 672-676.
Larson, R. B. (2019). Controlling social desirability bias. International Journal of Market Research, 61(5), 534-547. doi: https://doi.org/10.1177/1470785318805305
Levitt, S. D., & List, J. A. (2011). Was there really a Hawthorne effect at the Hawthorne plant? An analysis of the original illumination experiments. American Economic Journal: Applied Economics, 3(1), 224-238.
Lundh, A., Lexchin, J., Mintzes, B., Schroll, J. B., & Bero, L. (2017). Industry sponsorship and research outcome: systematic review with meta-analysis. Intensive Care Medicine, 43(11), 1603-1612. Doi: https://doi.org/10.1007/s00134-018-5293-7
McCambridge, J., de Bruin, M., and Witton, J. (2012). The Effects of Demand Characteristics on Research Participant Behaviours in Non-Laboratory Settings: A Systematic Review. PLoS ONE 7(6): e39116. https://doi.org/10.1371/journal.pone.0039116
Saunders, B., Kitzinger, J., & Kitzinger, C. (2015). Participant anonymity in the internet age: from theory to practice. Qualitative research in psychology, 12(2), 125-137. doi: https://doi.org/10.1080/14780887.2014.948697
Nichols, A. L., & Maner, J. K. (2010). The good-subject effect: Investigating participant demand characteristics. The Journal of general psychology, 135(2), 151-166. Doi: https://doi.org/10.3200/GENP.135.2.151-166
Athanassoulis, N., & Wilson, J. (2009). When is deception in research ethical?. Clinical Ethics, 4(1), 44-49. doi: https://doi.org/10.1258/ce.2008.008047
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of General Psychology, 2(2), 175–220. Doi: https://doi.org/10.1037/1089-2618.104.22.168
Orne, M. T. (1959). The nature of hypnosis: Artifact and essence. Journal of Abnormal Social Psycholology 58: 277–299.
Orne, M. T. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American Psychologist 17: 776–783.
Orne, M.T. (2010). Demand characteristics and the concept of quasi-controls. In Rosenthal, R. & Rosnow, R. (Eds.). Artifacts in Behavioral Research. Oxford: Oxford University Press.
Rosenthal, R. (2012). Covert Communication in Classrooms, Clinics, Courtrooms, and Cubicles. American Psychologist, 57(11), 839–849. doi: https://psycnet.apa.org/doi/10.1037/0003-066X.57.11.839
Schwarz, N., & Clore, G. L. (2016). Evaluating psychological research requires more than attention to the N: A comment on Simonsohn’s (2015)“small telescopes”. Psychological Science, 27(10), 1407-1409. doi: https://doi.org/10.1177/0956797616653102
Weathington, B. L., Cunningham, C. J., & Pittenger, D. J. (2010). Research methods for the behavioral and social sciences. New York: Wiley & Sons.
Weber, S. J., & Cook, T. D. (1972). Subject effects in laboratory research: An examination of subject roles, demand characteristics, and valid inference. Psychological bulletin, 77(4), 273. doi: https://psycnet.apa.org/doi/10.1037/h0032351
Dr. Chris Drew is the founder of the Helpful Professor. He holds a PhD in education and has published over 20 articles in scholarly journals. He is the former editor of the Journal of Learning Development in Higher Education. [Image Descriptor: Photo of Chris]