Refusal to Participate

Abstract

Refusal to participate, which MeSH classifies under dissent and disputes, is the declining of involvement in an activity, request, treatment, or research study. Cognitive psychology treats a refusal not as a simple absence of behaviour but as a motivated choice with identifiable causes and measurable consequences. The theory of psychological reactance explains refusal as the response to a perceived threat to a valued freedom, an intertwined state of anger and counterarguing that can make a persuasive appeal backfire. Survey methodology shows that what matters for a result is not the refusal rate itself but whether the propensity to refuse is correlated with what the study measures, the condition under which non-participation biases an estimate. This article surveys reactance, nonresponse bias, the volunteer subject, and refusal as the exercise of autonomy.

Keywords: refusal to participate, psychological reactance, nonresponse bias, volunteer subject, informed consent

Declining to take part is one of the most ordinary acts a person can perform, and precisely because it is so ordinary it is easy to treat as the mere absence of something rather than as a behaviour in its own right. Cognitive and social psychology take the opposite view: a refusal is a motivated decision, shaped by how the request is framed, by whether it is experienced as a threat to freedom, and by the person's reasons for engaging in the first place (Brehm, 1966). Studying refusal matters doubly, because it is both a phenomenon to be explained and a methodological problem for every science that relies on people agreeing to be studied (Groves, 2006).

What Refusal to Participate Is

Refusal to participate denotes the declining of involvement in an activity, procedure, request, or study that is expected of or offered to a person. MeSH files it under dissent and disputes, a placement that captures one true feature of refusal, that it is a form of disagreement, while understating another: refusal is dissent expressed by withdrawal rather than by argument, a way of opposing a course of action by simply not joining it. The indexing choice is a tool for organizing the literature rather than a theory of the behaviour, and the psychological accounts that matter cut across it.

What unites the cases is that a person who refuses has weighed a request and declined it, so the interesting questions are about the weighing. Three literatures supply most of the answers. The theory of psychological reactance treats refusal as a motivated resistance to control, triggered when a request threatens a freedom the person values (Brehm, 1966). Survey methodology treats refusal as nonresponse and asks when it distorts what a study concludes (Groves & Peytcheva, 2008). The ethics and psychology of consent treat refusal as the exercise of autonomy, a right to be respected rather than a resistance to be overcome (Appelbaum, 2007). The reactance tradition remains the most developed theoretical account, and a sustained programme of research has kept refining it (Steindl et al., 2015).

Psychological Reactance

The dominant psychological explanation of refusal is the theory of psychological reactance. Brehm proposed that people value their freedom to think and act as they choose, so that when a specific freedom is threatened or eliminated, a motivational state is aroused that drives them to restore it (Brehm, 1966). The clearest signature of reactance is the boomerang effect: a heavy-handed attempt to push someone toward a position produces movement away from it, because compliance would mean surrendering the threatened freedom, so the appeal lowers the very compliance it sought. Reviewing the theory four decades on, Miron and Brehm concluded that this core claim had held up well and organized a large body of findings across persuasion, health, and consumer behaviour (Miron & Brehm, 2006).

For most of its history reactance was treated as an unmeasurable intervening state, which limited the theory's testability. Dillard and Shen resolved this by proposing the intertwined model, in which reactance is best represented as a combination of two measurable responses, the anger aroused by the threat and the counterarguing the person generates against the message (Dillard & Shen, 2005). This operational definition let researchers quantify reactance and test its role directly, and a meta-analytic review confirmed the intertwined structure across many studies, establishing that threats to freedom arouse a joint affective and cognitive response that in turn predicts resistance to the appeal (Rains, 2013).

The practical lesson is that the forcefulness of a request is itself a cause of refusal, independent of the merits of what is being requested. A message that uses controlling, dogmatic, or explicitly pressuring language raises perceived threat to freedom and therefore the reactance it provokes, so an appeal can increase the very refusal it was meant to prevent. A programme of research synthesizing these developments has mapped how message features, individual differences, and restoration behaviours fit together, giving a detailed picture of when persuasion backfires and how it can be designed to arouse less reactance (Steindl et al., 2015).

Nonresponse and Its Bias

When refusal is studied as a methodological problem rather than a psychological one, it appears as survey nonresponse, and the central insight of the modern literature is counterintuitive. Groves argued that a high refusal rate is not, in itself, evidence that a survey's estimates are biased; what produces bias is a correlation between a person's propensity to respond and the very characteristic the survey is trying to measure (Groves, 2006). If refusers and respondents do not differ on the measured variable, even a large amount of refusal leaves the estimate unbiased; if they differ sharply, even a modest amount of refusal distorts it.

This decoupling of rate from bias was confirmed quantitatively. A meta-analysis of surveys with validation data found that the nonresponse rate was a poor predictor of nonresponse bias across studies, so that lowering the refusal rate does not reliably lower the bias and may not be the best use of resources (Groves & Peytcheva, 2008). The finding reframed decades of survey practice that had treated the response rate as the master indicator of quality.

Figure 1

Because rate and bias are only loosely coupled, the most intensive efforts to reduce refusal can be of limited value. A systematic review of methods for increasing response rates to postal questionnaires catalogued many tactics that reliably raise participation, from incentives to shorter instruments, but raising the rate addresses the bias only to the extent that late or persuaded responders resemble the hard refusers the survey still misses (Edwards et al., 2002). The modern recommendation is to measure and model who refuses, not merely to count how many do.

The Volunteer Subject

A refusal that biases a sample has a mirror image in the people who agree, and the classic treatment of this is Rosenthal and Rosnow's analysis of the volunteer subject. Assembling the evidence across many studies, they showed that people who volunteer for research differ systematically from those who decline: volunteers tend to be better educated, higher in need for social approval, more sociable, and of higher social class, among other reliable differences (Rosenthal & Rosnow, 1975). Because a self-selected sample over-represents these characteristics, any finding drawn from willing volunteers may not generalize to the population that includes the refusers, a threat to external validity that self-selection builds into a study before the first measurement is taken.

This self-selection bias is an enduring and general problem, not a historical curiosity. A concise modern treatment distinguishes participation bias, self-selection bias, and response bias as related ways that who chooses to take part, and who chooses to answer, can skew a result, and it remains a routine caution in fields from dermatology to epidemiology (Elston, 2021). The practical significance is that a refusal is not statistically neutral: the people a study fails to recruit are precisely those whose absence can most distort its conclusions.

Autonomy and the Right to Refuse

Refusal is not only a cause of bias or a product of reactance; it is also, in research and medicine, the exercise of a protected right. The doctrine of informed consent holds that a competent person may decline to take part in research or to accept a treatment, and that this decision must be respected. Appelbaum set out the clinical standard for assessing the decisional capacity that makes such a refusal authoritative: the ability to understand the relevant information, appreciate its application to one's own situation, reason about the options, and express a choice (Appelbaum, 2007). On this account a competent refusal is not a problem to be solved but a decision to be honoured, and the ethical failure would be to override it.

Why people decline is illuminated by self-determination theory, which distinguishes autonomous motivation, where a person acts from genuine interest or conviction, from controlled motivation, where they act under external or internal pressure (Ryan & Deci, 2000). The theory predicts that controlling, pressuring contexts undermine the sense of autonomy and so provoke the disengagement and refusal that reactance theory describes from the motivational side; a refusal can therefore be read as a signal about the social environment rather than a defect in the person. The two frameworks meet in the design of persuasion: because anger at a threatened freedom drives the reactant response, the emotional state the appeal evokes matters as much as its content, and messages that arouse or coincide with anger amplify the refusal they produce (Richards & Larsen, 2017).

Worked Example

Consider the central quantitative claim of the nonresponse literature, as rendered in the interactive nonresponse-bias model above. A survey estimates the mean of some variable from the people who respond, but the quantity of interest is the mean of the whole sample including those who refuse. A standard approximation expresses the bias of the respondent mean as the product of the nonresponse rate and the difference between respondents and nonrespondents on the variable: bias equals the refusal proportion times the quantity respondent mean minus nonrespondent mean.

Take a survey of annual doctor visits with a refusal rate of 40 percent, so the refusal proportion is 0.40. Suppose respondents average 4.0 visits a year and the refusers, who are healthier and less engaged with the health system, would have averaged 2.0. The bias in the respondent mean is 0.40 times the difference of 4.0 minus 2.0, that is 0.40 times 2.0, or 0.8 visits, so the survey would report about 4.0 when the true sample mean is 4.0 minus 0.8 times nothing: the full-sample mean is 0.60 times 4.0 plus 0.40 times 2.0, which equals 2.4 plus 0.8, or 3.2 visits. The respondent estimate of 4.0 overstates the truth of 3.2 by exactly the 0.8-visit bias the formula predicts.

Now hold the steep 40 percent refusal rate fixed but suppose the refusers do not differ from respondents, each averaging 4.0 visits. The bias becomes 0.40 times 4.0 minus 4.0, that is 0.40 times zero, or zero, and the respondent mean of 4.0 is exactly correct despite the high refusal rate. The two scenarios share an identical, large refusal rate and differ only in whether refusal is correlated with the measured variable, which is the whole of Groves's point: the rate sets how much a difference is amplified, but with no difference there is nothing to amplify (Groves, 2006).

Discussion

The three literatures surveyed here treat refusal from different angles that prove to be complementary. Reactance theory explains the psychology of the refusing act, locating its cause in a threatened freedom and its signature in the boomerang effect. Survey methodology explains the consequences of refusal for knowledge, showing that non-participation harms an estimate only when it is selective. The ethics of consent explains the standing of refusal, grounding it in autonomy and the capacity to decide. What they share is the premise that refusal carries information: about the person's motivation, about the sample's representativeness, and about the legitimacy of the request.

The design principle that follows is consistent across the perspectives: refusal is better understood and accommodated than suppressed. A persuader who lowers the threat to freedom arouses less reactance than one who presses harder; a methodologist who models who refuses learns more than one who merely drives the rate down; a clinician who honours a competent refusal acts more ethically than one who overrides it. Across a half-century the reactance tradition in particular has matured from an unmeasurable intervening state into a quantified and well-reviewed account of when and why people push back (Rosenberg & Siegel, 2018).

Current Directions

The most developed contemporary synthesis is in persuasive health communication, where reactance is now a routine design consideration. A review of the literature maps how message features such as controlling language and explicit threats to freedom raise reactance, and how techniques including choice-enhancing postscripts, empathy, and self-affirmation can lower it, turning the theory into a practical toolkit for messages that inform without provoking refusal (Reynolds-Tylus, 2019). This applied front keeps the fifty-year-old theory current by testing it against the problem of persuading people who are free, and entitled, to say no (Rosenberg & Siegel, 2018).

On the methodological side, the study of who refuses continues to sharpen. A systematic review of survey nonresponse and its association with health confirms that refusers and respondents differ in ways that bear directly on health estimates, so that nonresponse is not random and its bias must be modelled rather than assumed away (Rybak et al., 2023). The volunteer-subject problem has also been revisited in new settings: a recent analysis in applied linguistics describes a volunteer participation paradox, in which the very recruitment practices meant to protect participants intensify self-selection, so that samples drift further from the populations researchers hope to describe (Hiratsuka & Suzuki, 2025). Both strands reflect a shared turn away from treating refusal as noise and toward modelling it as a systematic, informative part of the data.

Key Researchers

Paul S. Appelbaum

(contemporary). Defined the clinical and legal standards by which a person's competence to give or withhold informed consent is assessed, grounding a refusal to participate in research or treatment as the exercise of a decisional right rather than an obstacle. ORCID

Jack W. Brehm

(1928-2009). Formulated the theory of psychological reactance, the motivational state aroused when a free behaviour is threatened, which remains the dominant explanation of why people refuse requests and persuasion they perceive as coercive. Wikidata

Edward L. Deci

(1942-2026). Co-founded self-determination theory, which distinguishes autonomous from controlled motivation and reframes refusal as the predictable response to controlling contexts that thwart the basic need for autonomy. Wikipedia

James Price Dillard

(contemporary). Developed the intertwined cognitive-affective model of reactance, defining it as a blend of anger and counterarguing and giving the refusal response to persuasion an operational definition that could be tested empirically. ORCID

Robert M. Groves

(contemporary). Reframed survey nonresponse as a problem of bias rather than rate, showing that a high refusal rate biases an estimate only when the propensity to refuse is correlated with what the survey measures. Wikipedia

Stephen A. Rains

(contemporary). Provided the meta-analytic consolidation of the intertwined reactance model, confirming across studies that threats to freedom arouse a joint anger-and-cognition response that predicts resistance to a persuasive appeal. ORCID

Robert Rosenthal

(1933-2024). Documented the volunteer subject, showing systematically that people who agree to take part in research differ from those who refuse, so that self-selected participation biases samples and threatens external validity. Wikipedia

Lijiang Shen

(contemporary). Co-developed and validated the measurement of reactance as anger and counterarguing and extended it to the boomerang effect, where a persuasive health appeal increases the refusal it sought to prevent. ORCID

Glossary

Autonomy.

The capacity and right to govern one's own choices, the value whose protection underlies informed consent and whose thwarting provokes refusal.

Boomerang Effect.

The signature of reactance in which a forceful appeal produces movement away from the advocated position, increasing the very refusal it was meant to prevent.

Compliance.

Agreement to a request; the outcome that reactance reduces when a message is experienced as a threat to freedom rather than an invitation.

Controlled Motivation.

Action undertaken under external or internal pressure rather than genuine interest; the motivational context that self-determination theory links to disengagement and refusal.

Counterarguing.

The generation of thoughts against a persuasive message; the cognitive half of the intertwined model of reactance.

Informed Consent.

The agreement of a competent, informed person to take part in research or treatment, which equally entails the protected right to refuse.

Intertwined Model.

Dillard and Shen's representation of reactance as a combination of anger and counterarguing, the operational definition that made the state measurable.

Nonresponse Bias.

The distortion of a survey estimate that arises when the propensity to respond is correlated with the variable being measured; the product of the refusal rate and the refuser-respondent difference.

Nonresponse Rate.

The proportion of a sample that does not respond; a poor predictor of nonresponse bias on its own.

Psychological Reactance.

The motivational state aroused by a threat to a valued freedom, which drives a person to restore the freedom, often by refusing or doing the opposite of what is urged.

Reactance Proneness.

A stable individual difference in the tendency to experience reactance, making some people more disposed than others to refuse in response to pressure.

Refusal to Participate.

The declining of involvement in an activity, request, treatment, or study; a motivated choice studied as reactance, as nonresponse, and as the exercise of autonomy.

Response Rate.

The proportion of a sample that participates; long treated as a master indicator of survey quality, though it tracks bias only loosely.

Self-Determination Theory.

The theory of motivation distinguishing autonomous from controlled regulation, which explains refusal as a response to contexts that thwart the need for autonomy.

Self-Selection Bias.

The distortion that arises when the people who choose to take part differ systematically from those who refuse, compromising a sample's representativeness.

Threat to Freedom.

A perceived constraint on a behaviour a person believes they are free to perform; the proximal trigger of psychological reactance.

Volunteer Subject.

A research participant who self-selects into a study and who differs systematically from those who decline, so that findings from volunteers may not generalize.

Frequently Asked Questions

What does refusal to participate mean in psychology?

It is the declining of involvement in an activity, request, treatment, or research study, treated not as a simple absence of behaviour but as a motivated choice with identifiable causes, studied through the lenses of reactance, survey nonresponse, and autonomy (Brehm, 1966).

What is psychological reactance?

It is the motivational state aroused when a person perceives a threat to a freedom they value, which drives them to restore that freedom, often by refusing or by moving toward the opposite of what is urged (Brehm, 1966).

Why can pressuring someone backfire?

Because a controlling or heavy-handed appeal is itself a threat to freedom, it arouses anger and counterarguing that push the person away from the advocated position, the boomerang effect that reactance theory predicts (Dillard & Shen, 2005).

Does a high survey refusal rate mean the results are biased?

Not necessarily; a high refusal rate biases an estimate only when the propensity to refuse is correlated with the variable being measured, so the rate alone is a poor predictor of bias (Groves & Peytcheva, 2008).

What is the volunteer subject problem?

People who agree to take part in research differ systematically from those who refuse, being on average better educated and higher in need for approval, so a self-selected sample can fail to represent the wider population (Rosenthal & Rosnow, 1975).

Is refusing to take part the same as not caring?

No; self-determination theory treats refusal as an informative response to a controlling context that thwarts autonomy, so a refusal often signals something about the environment rather than indifference in the person (Ryan & Deci, 2000).

What is the difference between a refusal rate and refusal bias?

The rate is simply how many people decline, while the bias is how much their absence distorts a result; the two are only loosely coupled, because bias also depends on how much refusers differ from participants (Groves, 2006).

Do people have a right to refuse to participate in research or treatment?

Yes; the doctrine of informed consent holds that a competent person may decline, and the clinical standard assesses their capacity to understand, appreciate, reason, and express a choice so that a competent refusal is respected (Appelbaum, 2007).

References