OR IG I NAL ART I C L E Consent searches and underestimation of compliance: Robustness to type of search, consequences of search, and demographic sample Roseanna Sommers1 | Vanessa K. Bohns2 1University of Michigan Law School, Ann Arbor, Michigan, USA 2Cornell University ILR School, Ithaca, New York, USA Correspondence Roseanna Sommers, University of Michigan, 701 South State Street, Ann Arbor, MI 48109, USA. Email: rosesomm@umich.edu Funding information NSF, Grant/Award Number: 1823661 Abstract Most police searches today are authorized by citizens’ consent, rather than probable cause or reasonable suspicion. The main constitutional limitation on so-called “consent searches” is the voluntariness test: whether a reasonable person would have felt free to refuse the officer’s request to conduct the search. We investigate whether this legal inquiry is subject to a systematic bias whereby uninvolved decision-makers overstate the voluntariness of consent and underestimate the psychological pressure individuals feel to comply. We find evidence for a robust bias extending to requests, tasks, and populations that have not been examined previously. Across three pre-registered experiments, we approached participants (“Experiencers”) with intrusive search requests and measured their behavioral compliance and self-reported feelings of psychological freedom. Another group of participants (“Forecasters”) reported whether they would comply if hypotheti- cally placed in the same situation. Study 1 investigated participants’ willingness to allow experimenters access to their unlocked personal smartphones in order to read through the search histories on their web browsers—a private sphere where many individuals feel they have something to hide. Results revealed that whereas 27% of Forecasters reported they would permit such a search, 92% of Experiencers complied when asked. Study 2 rep- licated this underestimation-of-compliance effect when individuals were asked to permit a search of their purses, backpacks, and other bags—traditional searches not eligible for [Correction added on 31 January 2024, after first online publication: Author affiliations in the author byline were updated in this version.] DOI: 10.1111/jels.12375 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2023 The Authors. Journal of Empirical Legal Studies published by Cornell Law School and Wiley Periodi- cals LLC. 4 J Empir Leg Stud. 2024;21:4–34.wileyonlinelibrary.com/journal/jels mailto:rosesomm@umich.edu http://creativecommons.org/licenses/by-nc-nd/4.0/ http://wileyonlinelibrary.com/journal/jels http://crossmark.crossref.org/dialog/?doi=10.1111%2Fjels.12375&domain=pdf&date_stamp=2023-12-26 the heightened legal protection extended to digital devices. Study 3 replicated the gap between Forecasters’ projections and Experiencers’ behavior in a more representative sample, and found it persists even when participants’ predictions are incentivized monetarily. K E Y W O R D S compliance, consent, police searches, psychology INTRODUCTION The Constitution protects people from unreasonable searches by government agents, a guarantee that is supported by the Fourth Amendment’s requirement that “no Warrants shall issue, but upon probable cause” (Schneckloth v. Bustamonte, 1973; U.S. Const. amend. IV). The rationale for the warrant requirement rests on privacy: as the Supreme Court explained in Wolf v. Colorado (1949, p. 27), “[t]he security of one’s privacy against arbitrary intru- sions by the police—which is at the core of the Fourth Amendment—is basic to a free society.” But police officers often conduct searches without warrants and without reasonable suspicion: searches that are based on consent (LaFave, 2007). So-called “consent searches,” which are a recognized “exception[] to the requirements of both a warrant and probable cause” (Schneckloth v. Bustamonte, 1973, p. 219), are ubiq- uitous. Exact numbers are hard to come by, but commentators estimate that they affect “tens of thousands if not hundreds of thousands of people a year” (Eith & Durose, 2011; Gau, 2013; McGlinchy, 2018; Nadler & Trout, 2012; Strauss, 2001, p. 214). Upwards of 90% of searches conducted by police are justified based on con- sent, rather than a warrant or probable cause (Burke, 2015; Simmons, 2005; Sutton, 1986).1 These searches are disproportionately performed on motorists of color (Baumgartner et al., 2020; Gau, 2013; McGlinchy, 2018; Pierson et al., 2020). In the vast majority of searches, no contraband is discovered (Bar-Gill & Friedman, 2012; Dittmeier, 1967; McGlinchy, 2018; State v. Carty, 2002). Legal commentators have observed that whereas traditional police searches are “publicly authorized”—greenlit by the government (e.g., a court or adminis- trative agency issuing a warrant; an agent of the government such as a police officer deciding probable cause exists)—searches based on consent are “privately authorized”—greenlit by an individual private citizen (Rotenberg, 1991, p. 175). Publicly authorized searches rely on public justifications, and tend to be subject to limitations that correspond to those justifications: they are “confined by par- ticularity of both place to be searched and item to be seized” (Rotenberg, 1991, 1But see Warren and Tomaskovic-Devey (2009), noting a decline over time in the proportion of searches conducted pursuant to consent. CONSENT SEARCHES AND COMPLIANCE 5 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense p. 175). Consent-based searches, by contrast, devolve power to the people to decide for themselves whether to permit a search, and within what scope. Consent-search jurisprudence thus leaves it up to private individuals to protect themselves. In theory, individuals approached with consent requests will make decisions in line with their personal preferences, based on their own weighing of privacy considerations against other factors, such as a desire to cooperate with police and aid police in reducing crime. “There is nothing constitutionally suspect in a person’s voluntarily allowing a search,” the Supreme Court has firmly stated (Schneckloth v. Bustamonte, 1973, p. 243). Thus, few constitutional guardrails on consent searches exist. The main limitation is that consent must be given freely and voluntarily: it cannot be the product of “duress or coercion, express or implied” (Schneckloth v. Bustamonte, 1973, p. 227). The voluntariness of consent is determined by “the totality of the surround- ing circumstances” (Schneckloth v. Bustamonte, 1973, p. 227). Factors that courts have recognized as bearing on the voluntariness of consent include whether officers brandished their weapons, made intimidating movements, issued explicit threats or commands, spoke in a loud or menacing voice, or blocked the individual’s egress (Dittmeier, 1967; Nadler, 2002; Strauss, 2001; Sutherland, 2006). Police encounters that are “polite, friendly, and informal” are generally deemed not coercive (Nadler, 2002, p. 199). While voluntariness does not require officers to inform citizens of their right to refuse consent to a search, officers are not permitted to falsely state or imply that the individual has no choice but to submit to the search (Schneckloth v. Bustamonte, 1973). The key test of voluntariness, the Supreme Court has explained, is how free a reasonable person would feel in the situation: “the appropriate inquiry is whether a reason- able person would feel free to decline the officers’ requests or otherwise termi- nate the encounter” (Florida v. Bostock, 1991, p. 436).2 This means that the legal standard for assessing the constitutionality of consent searches addresses not the justifications that might support the need to perform such a search (e.g., evidence suggesting that the individual is hiding something illegal), but the voluntariness with which the individual grants consent to the police. Thus, the legal inquiry in consent-search cases requires outside observers to perform a highly difficult psychological task: Decision-makers must determine how a rea- sonable person would feel in a given encounter with the police, and must make judgments about the kinds of influence that tend to render people unable to refuse consent. 2In particular, courts ask how a reasonable person would respond to police pressure in order to determine if the suspect’s will was overborne by police coercion, an inquiry that requires a social psychological assessment of the likely effects of police pressure on citizens’ feelings, beliefs, and behaviors. This legal inquiry is subjective, taking into account the individual circumstances of the suspect (e.g., age, intelligence, prior experience with the police, etc.), while also analyzing how people tend to respond psychologically to police pressure (Schneckloth v. Bustamonte, 1973). 6 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense While prior research in psychology has demonstrated that suspiciousness judg- ments are likely to be influenced by biases in threat detection (e.g., racial bias) (Correll et al., 2014; Donders et al., 2008; Wilson et al., 2017) and hindsight bias (i.e., the belief that a given outcome was foreseeable all along) (Kagehiro et al., 1991; Roese & Vohs, 2012), comparatively little work has examined the psy- chological processes through which judges assess the voluntariness with which a requestee grants consent (Kagehiro, 1988, 1990).3 The current research begins to fill this gap in the literature by drawing on research on the psychology of compliance and a promising new experimental paradigm for studying assessments of voluntari- ness in consent-search contexts. THE PSYCHOLOGY OF JUDGING VOLUNTARINESS Are judges able to recognize the difficulty of saying “no”? Prior research from psychology suggests that third-party decision-makers make striking mistakes when attempting to predict other people’s attitudes and behaviors. When people attempt to engage in “perspective-taking,” they typically draw on their own experience as a starting point and adjust from there (Bohns & Flynn, 2015). Their adjustments are often insufficient, however, which can lead to systematic social prediction errors (Epley et al., 2004). Successful perspective-taking often requires uninvolved decision-makers to make not just one, but two adjustments (Van Boven et al., 2005, 2013). First, they must adjust from “self in current emo- tional state” to “self in different emotional state.” Then, they must adjust again from “self” to “other,” which can lead to its own set of errors (Nordgren, Banas, & MacDonald, 2011; Nordgren, McDonnell, & Loewenstein, 2011). Prior work on “hot-cold empathy gaps” has demonstrated the errors that can result when attempting to predict one’s own preferences and behaviors in a different emotional state (Van Boven et al., 2013). These dynamics help explain the large differences that have been documented between what people imagine they would do and what they actually do in a variety of social situations. For example, a person who merely imagines being sexually harassed, or bearing wit- ness to homophobic or racist remarks, will tend to underestimate the extent to which fear and discomfort will prevent them from speaking up against such offenses. Thus, outside observers imagine feeling more emboldened to stand up for themselves or others in such contexts than people prove to be when they find themselves in the relevant situations (Crosby & Wilson, 2015; Kawakami et al., 2009, 2019; Woodzicka & LaFrance, 2001). 3But see Kagehiro (1988), reporting the results of a 2 � 2 � 2 vignette study with 12 participants per condition, in which some participants imagined “you and your roommate” interacting with a police officer (actor condition) while other participants imagined “two friends” interacting with an officer (observer condition). The study showed that “observers overestimated consentors’ perceived choice in revoking consent once given (M = 7.0 vs. 5.2, on a 0–10 scale)” (Kagehiro, 1990, p. 189). CONSENT SEARCHES AND COMPLIANCE 7 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense These dynamics, and resulting errors, are exacerbated when a second judg- ment is introduced: when rather than merely predicting one’s own preferences and behaviors in a different emotional situation than one is currently in, one is asked to predict another person’s preferences and behaviors in a different emo- tional context. Research has found that individuals considerably underestimate the emotional distress another person is likely to experience, thus mispredicting their resulting behavior, in the face of embarrassment (Van Boven et al., 2005) and bullying (Nordgren, Banas, & MacDonald, 2011). Similarly, studies on the psychology of compliance have found that people who merely imagine being approached with a request have been found to underestimate just how difficult it is to say “no.” For example, Flynn and Lake (2008) documented that face-to-face requests for small favors, such as requests to borrow one’s cell phone, complete a brief questionnaire, or donate to a charitable cause, yielded compliance rates of over 50%. Even when the request was unethical, and targets of the requests expressed discomfort, com- pliance rates remained quite high (Bohns et al., 2014). In one study, 64% of participants agreed to a request to vandalize a library book by scribbling the word “pickle” on one of the pages. All told, Bohns (2016) reviewed 14 studies in which more than 14,000 participants were asked a variety of requests of this sort and found that compliance rates were consistently high—over 50%— regardless of the type of request. Importantly, before any of these individuals made these requests, they first predicted how often they would be turned down. Results show that whereas requesters generally expected low compli- ance rates—an average of 28% across all 14 studies—the actual average com- pliance rate was 54%, almost double requesters’ predictions (Bohns, 2016). Applied to consent searches, the foregoing research suggests that compliance with search requests is likely to be high, and that third-party decision-makers (e.g., judges) tasked with judging the voluntariness of consent are likely to underappreciate the extent of the psychological pressure to comply. That is, forecasters sitting in a “cold” state are likely to underappreciate the emotions (e.g., discomfort, fear) others experience when considering the prospect of say- ing “no” to a police officer. CONSENT SEARCHES AND THE PSYCHOLOGY OF JUDGING VOLUNTARINESS Fourth Amendment scholars have frequently criticized the Supreme Court’s consent-search jurisprudence, arguing that the justices have blithely adopted a picture of voluntariness at odds with psychological research and common sense (Barrio, 1997; Maclin, 2008; Simmons, 2005; Williams, 2007). Commentators have frequently cited Stanley Milgram’s laboratory studies (Milgram, 1974) to draw conclusions about the voluntariness of consent to 8 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense searches by police (Simmons, 2005, pp. 802–807), arguing that “obedience to authority is deeply ingrained [and] people will obey authority even when it is not in their best interest to do so” (Strauss, 2001, p. 236). However, these arguments have historically been subject to challenge due to the number of disanalogies between the context of Milgram’s studies and that of consent searches. The most important is that Milgram’s experimenter pressured participants to continue administering (what they believed to be) painful electric shocks, telling them, “The experiment requires that you continue” and eventu- ally, “You have no other choice, you must continue.” As Simmons (2005, p. 806) explains, “There is no question that if a police officer told a suspect, ‘The law requires that you allow me to search,’ or, ‘You have no other choice, you must allow me to search your bag,’ the consent would be deemed involuntary.” Thus, the application of Milgram’s results to the policing context is tenuous. A second important difference between Milgram’s laboratory and the real-world phenomenon of consent searches is that Milgram’s participants were asked to harm another person, whereas with consent searches, the consenters are ordinar- ily the ones who face the risk of potential harm (Sommers & Bohns, 2019).4 More recent research on the topic of compliance has sought to address these critiques, and has continued to confirm the central insights of this research: namely, that it is extremely difficult for people to say “no” to face-to-face requests, and thus they do so infrequently. In a series of laboratory studies, Sommers and Bohns (2019) approached participants with a request to unlock their password-protected smartphones and hand them over to an experimenter to search through. A separate group of participants answered whether, hypotheti- cally, they would agree to the same request. The researchers found that whereas most Forecasters believed a reasonable person would refuse such a request, the majority of Experiencers complied. Furthermore, Experiencers reported feeling less free to refuse the request than Forecasters anticipated. In a second study, the researchers found that informing people of their right to refuse did little to alter the rates at which Experiencers complied or how free they reported feeling. The authors concluded that “a popular reform proposal—requiring police to advise citizens of their right to refuse consent—may have little effect,” once one takes into account the psychology of compliance (Sommers & Bohns, 2019, p. 1962). Taken together, these studies aid in assessing the psychological pressure indi- viduals experience when an officer requests consent to a search, for a number of reasons. First, these laboratory studies are less susceptible than Milgram’s experiments to the critiques outlined above. The scripts that were used in these studies were straightforward requests (e.g., “Before we begin the study, can you please unlock your phone and hand it to me?”; “I’m playing a prank on a friend, would you write the word ‘pickle’ in this library book so they don’t recognize 4Of course, Milgram’s studies may be more relevant when it comes to third-party consent searches, which often involve a decision-maker choosing whether to impose a potential privacy invasion on another person. CONSENT SEARCHES AND COMPLIANCE 9 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense my writing?”; “Could I borrow your phone to make a call?”). Requesters were instructed not to pressure participants into complying if they hesitated or said “no.” Thus, there is little concern that requesters in these studies put more pres- sure on targets to comply than a law enforcement agent would be legally permit- ted to use. In addition, analogous to consent-search contexts, the individuals who would potentially be harmed by complying with these requests are the tar- gets themselves, who take on the potential risk of, for example, engaging in an act of vandalism for which they could be sanctioned, or handing over their cell phone to a stranger and risking that it could be stolen. While these studies offer a promising new paradigm for studying the percep- tion of voluntariness in consent searches, several questions remain about the gen- eralizability and external validity of these initial findings. In particular, the search context in these studies was dissimilar to a real search context in three key ways, which we address in the current research. First, research subjects may feel they have nothing to hide, leading them to consent with less apprehension, as com- pared to a person stopped by police and asked to submit to a Fourth Amendment search. Second, the law recognizes the special status of digital privacy; indeed, asking to search a person’s smartphone may be different from asking to search a more traditional personal item. Finally, prior research on the psychology of com- pliance has been conducted on undergraduate samples that are not demographi- cally representative of the population of individuals who tend to be subjected to consent searches. While academic research tends to rely on convenience samples that consist of mostly female undergraduate students of Caucasian or Asian descent, consent searches are most often performed on Black men (Baumgartner et al., 2020; Gau, 2013; McGlinchy, 2018; Pierson et al., 2020). The present research attempts to address each of these limitations in turn, testing the generalizability of the initial findings and thus their usefulness for legal appli- cation. As courts have already begun citing the results of the initial experimental studies on consent searches (People v. Tacardon, 2022; State v. Diego, 2021; State v. Hauge, 2022; State v. Turnquest, 2019; United States v. Meadows, 2021), test- ing the robustness of these laboratory findings is a pressing concern. THE PRESENT RESEARCH The current studies use the paradigm established by Sommers and Bohns (2019) in which researchers ask one group of participants (“Experiencers”) to consent to a search of something on their person, either asking them to unlock and hand over their password-protected smartphones or to permit a search of their personal bags, while a separate group of participants (“Forecasters”) predicts how they would respond to this request, presented hypothetically. In three pre-registered studies, we test whether the initial results that emerged from this paradigm generalize to additional requests and populations that have not been examined previously. In 10 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Study 1, participants are asked to consent to a search of a domain where many participants feel a sense of having something to hide: the search histories on their web browsers. In Study 2, participants are asked to consent to a search of their purse or backpack—a more traditional search request that invites less legal scru- tiny than searches of digital devices. Finally, Study 3 seeks to replicate the study in a sample that is more diverse in race, age, and education—and closer to the popu- lation that is disproportionately asked to consent to police searches. STUDY 1: SEARCH WITH SOMETHING TO HIDE Laboratory research on the psychology of consent faces external validity limita- tions when drawing inferences about police searches. One key difference between the research context and the policing context is that research partici- pants risk losing their privacy, whereas those approached by police risk much more than an invasion of privacy. Those who harbor contraband risk losing their liberty and suffering the various collateral consequences that come with criminal justice system entanglement, including those concerning employment, housing, immigration, child custody, criminal justice debt, and so on (Ahrens, 2000; Pager, 2003; Pager et al., 2009; Sobol, 2015). Ethically, we could not mimic the stakes of a real police search in a research study. Nonetheless, we could address the possibility that participants might see no cost to consenting to a search, by specifying that the search would implicate a pri- vate sphere where many participants feel they have “something to hide”: their web search histories. Prior research shows that people’s Google histories “act as a ‘digi- tal truth serum’ for deeper and darker thoughts” (Shermer, 2018): many people search the web for things that are unflattering, shameful, or even illegal (Buhi et al., 2009; Stephens-Davidowitz, 2017). Thus, a request to check participants’ web search history is a request to look through an exceedingly private sphere. Fur- thermore, prior work on concrete versus abstract construal levels suggests that an experimenter’s request that vaguely references the need to “search some things” is likely to prompt a less vivid representation in participants’ imaginations than a concrete request to search a particular thing (Fujita, 2008). For this reason, we can expect that the specific request to look through participants’ web search histories is likely to be perceived as especially invasive and objectionable. Method Pre-registration This study was pre-registered as AsPredicted.com (#28698; https://aspredicted. org/2aa8j.pdf). CONSENT SEARCHES AND COMPLIANCE 11 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense http://aspredicted.com https://aspredicted.org/2aa8j.pdf https://aspredicted.org/2aa8j.pdf Participants All 205 students enrolled in a behavioral science course at a university in the Northeast were invited to take part in the study for extra credit. One hundred fifty-three students participated in the study. Participants were 41% male, 59% female; 5% Black, 44% White, 10% His- panic/Latinx, 26% Asian/Asian–American, and 16% choosing another or multiple designations. Ages ranged from 17 to 29 years, with a median age of 19 years. Procedure All participants came into the lab, were greeted by an experimenter, and led to separate rooms where they were asked to have a seat. The experimenter was always one of three research assistants (two female, one male) who had been trained to follow a script. These experimenters wore their regular attire. In all studies, the typical research consent process was modified such that participants were not asked to sign an informed consent document agreeing to participate in the research study until after the experimenter’s search request (described below) was made. This modification to the standard informed con- sent procedure, in which a consent form is provided at the beginning of the study, was approved by the Institutional Review Board at the university where the study took place. The purpose of the modification was to ensure that the research consent process did not itself affect participants’ responses to the search request. Experiencing condition After taking a seat, participants in the Experiencing condition were immediately asked by the experimenter: “Before we begin the study, can you please unlock your phone and hand it to me? I’ll just need to take your phone outside of the room for a moment in order to check your web search history.” If the participant complied, the experimenter took the phone out of the room, waited for 5 seconds, and brought the phone back. If the participant at any point said “no” or otherwise declined to hand over the phone, the experi- menter immediately moved on to the questionnaire phase of the study. Following this interaction, participants filled out a questionnaire reporting whether they had agreed to hand over their phone in response to the experi- menter’s request moments before (Yes/No). Participants also rated on 7-point Likert scales how easy, comfortable, and awkward it had been for them to say “no” to this request, and how free and pressured they felt. 12 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense To measure whether participants attended to and remembered that the experimenter would be checking their web search histories, participants were asked to report what the experimenter had requested to search. They were instructed to write “do not recall” if they did not remember. Next, participants were asked a series of questions about how they use their phone: how comfortable they are with the idea of an experimenter checking their web search history (1 = not at all; 7 = extremely); whether they clear their search history regularly (Yes/No); and which of several activities they do on their phone without additional security measures (e.g., an additional passcode protecting specific apps). They also completed a demographic questionnaire. Finally, for each Experiencer, the experimenter reported whether the partici- pant had hesitated, asked questions, expressed unease, or apologized during their encounter. The experimenter also noted whether the participant had, in fact, handed over their phone (Yes/No). The experimenter additionally made notes about any comments or noteworthy behaviors by participants. Forecasting condition After being greeted by the experimenter and taking a seat, participants in the Forecasting condition were immediately given a questionnaire that asked what a reasonable person would do if the experimenter asked them to unlock and hand over their phone so that their web search history could be accessed (request presented verbatim) (Yes/No). Forecasters also answered whether they personally would surrender their phone in such a situation (Yes/No). Next, they answered the same 7-point Likert scales reporting how easy, com- fortable, and awkward it would be for them personally to say “no” to this request, and how free and pressured they would feel. Finally, they answered the same questions as Experiencers: recall of the purpose for which the phone would be searched; comfort with their web search histories being checked; how they use their phones; and demographic questions. Debriefing Before exiting the study, participants in both conditions were compensated, thanked for their participation, and debriefed. During the debriefing, we con- firmed that participants did not recognize the experimenter from another context and had not heard anything about the study beforehand. Results Phone use In order to verify that this population’s personal smartphones contained sensi- tive personal information, which would make a search of their devices a signifi- cant privacy intrusion, we first analyzed participants’ responses to what was CONSENT SEARCHES AND COMPLIANCE 13 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense contained on their phones. Most participants use their phones for personal mes- saging (99.34%), email (97.71%), storing contacts (98.64%), photos (99.32%), shopping (90.72%), playing games (92.66%), and storing sensitive information such as social security numbers and passwords (88.00%)—without requiring sep- arate passwords for these apps. 42.31% report using their phone for banking. Only 24.84% regularly clear their web search history on their phones. Comprehension check When asked to recall the reason the experimenter gave for wanting to search their phones, 91.50% of participants (94.67% of Experiencers and 88.46% of Forecasters) accurately reported that the experimenter intended to access their web search histories. A minority (5.23%) reported that they did not recall and 3.27% wrote something nonspecific, such as “I would need you to unlock your phone and hand it to me,” or inaccurate, such as “[W]e are not given the request.” Compliance behavior As hypothesized, few Forecasters (20.51%) predicted that a reasonable person would agree to hand over their phone, whereas most Experiencers (92.00%) actu- ally surrendered their phones. This difference was significant, χ2(1, N = 153) = 76.27, p < 0.001, φ = 0.72 (Figure 1). Forecasters also reported whether they personally would agree to this request. As predicted, few Forecasters (26.92%) predicted they personally would comply. This was significantly lower than the rate at which Experiencers actu- ally complied, χ2(1, N = 153) = 64.19, p < 0.001, φ = 0.66 (Figure 1). Within subjects, there was no significant difference between Forecaster’s judgments of what a reasonable person would do versus what they personally would do, McNemar’s χ2(1) = 1.45, p = 0.23. Feelings of freedom Consistent with our pre-registration, we combined participants’ ratings of how pressured, free, comfortable, awkward, and easy it would be (for Forecasters) or was (for Experiencers) to refuse the experimenter’s request into a single measure of feelings of freedom (α = 0.83), and conducted a one-tailed t-test assessing the directional prediction that Forecasters would indicate greater feelings of free- dom than Experiencers. As predicted, Forecasters imagined feeling more free 14 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense (M = 3.99, SD = 1.16) than Experiencers actually reported feeling (M = 3.04, SD = 1.24), tone-tailed(151) = 4.85, p < 0.001, d = 0.78. We observed no significant individual experimenter effects on (i) compliance behavior among Experiencers; (ii) forecasted compliance among Forecasters; or (iii) feelings of freedom. None of these measures differed by participants’ self- reported race/ethnicity, gender, or age, with one exception: overall, female par- ticipants reported feeling less free than male participants. We did not anticipate this result or pre-register this analysis, and it did not replicate in Studies 2 and 3. No significant interaction between gender and condition was observed; within each gender group, Forecasters gave significantly higher feelings-of-freedom rat- ings than did Experiencers. Comfort with one’s web search history being accessed Participants in both conditions were asked, “How comfortable are you with the idea of the experimenter checking your web search history?” (1 = Not at all; 7 = Extremely). Forecasters’ average rating was 3.77 (SD = 1.64). Experiencers’ average rating was 3.57 (SD = 1.88), a nonsignificant difference, t(151) = 0.69, p = 0.49, d = 0.11. 20.51 % 26.92 % 92.00 % 0.00 0.25 0.50 0.75 1.00 Would a reasonable person comply? Would you comply? Did you comply? P ro po rt io n co m pl yi ng Forecasters Experiencers Study 1: Forecasted vs. Actual Compliance Behavior F I G U R E 1 Most Experiencers (92.00%) complied with the request to allow researchers to look through their web search histories on their unlocked personal smartphones. Far fewer Forecasters (26.92%) predicted they would comply or that a reasonable person would comply (20.51%). Error bars represent bootstrapped 95% confidence intervals. CONSENT SEARCHES AND COMPLIANCE 15 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Discussion Study 1 found that the vast majority of Experiencers complied, while most Forecasters asserted that they would not comply, and that it would not be reason- able to comply. The post-task survey revealed that most Experiencers were aware of, and remembered, that the researchers were asking to look through their web sea- rch histories. Thus, most people did something that most regarded as unreasonable. A limitation of prior research was the difference—psychological as well as legal—between people who “have nothing to hide” and those who are in posses- sion of something they do not want discovered. The results of Study 1 suggest that even when people are asked to consent to a search of a domain where they may feel they have something to hide, most Experiencers still comply—and Forecasters still underestimate compliance. While participants in the current study may have been concerned with hiding information that might lead to embarrassment more so than with tangible legal consequences, embarrassment is nonetheless considered to be a powerful motivator, on par with concerns about doing something potentially unethical or illegal (Sabini et al., 2001). STUDY 2: SEARCHING NONDIGITAL BELONGINGS Are smartphones special? A person’s smartphone contains “a comprehensive cache of their communications, a complete inventory of their friends and acquaintances, [and] a perfect record of their whereabouts” (Sommers & Bohns, 2019, p. 1980). For this reason, the Supreme Court has singled out smartphones as deserving greater protection than other objects, such as wallets or purses (Marshall et al., 2019; Riley v. California, 2014). Legal commentators have observed that “[c]ourts and legislatures are increasingly recognizing that ‘digital is different’” (Weisburd, 2019, p. 717). Despite the novel questions raised by electronic information, traditional pri- vacy remains both significant and politically relevant. People’s cars, suitcases, and backpacks may not receive the same heightened constitutional protections that their smartphones do, but a search of one’s physical belongings can still be intru- sive and consequential. Thus, it is important to understand whether the psycholog- ical phenomena explored here generalize to these nondigital search domains. In Study 2, we requested to search participants’ “bags.” We chose “bag” because it can refer to a purse, backpack, briefcase, or shopping bag. Pre-registration This study was pre-registered at AsPredicted.com (#35593; https://aspredicted. org/um4du.pdf). 16 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense http://aspredicted.com https://aspredicted.org/um4du.pdf https://aspredicted.org/um4du.pdf Participants All 170 students enrolled in a behavioral science course at a university in the Northeast were invited to take part in the study for extra credit. Ninety-four stu- dents took part. Participants were 36% male, 63% female; 6% Black, 55% White, 9% His- panic/Latinx, 29% Asian/Asian-American, and 1% choosing another or multi- ple designations. Ages ranged from 18 to 27 years, with a median age of 19 years. Procedure We followed the same procedure as in Study 1, with the following modifications. As before, the experimenter was always one of three research assistants (2 female, 1 male) dressed in regular attire. This time, the request presented to participants was: “Before we begin the study, can you please hand me your bag? I’ll just need to take your bag outside of the room for a moment in order to search it.” As before, participants in both conditions filled out a questionnaire asking whether they had agreed (Experiencers) or would personally agree (Forecasters) to hand over their bag (Yes/No). They rated how easy it was and how comfort- able, awkward, free, and pressured they felt (Experiencers) or imagined someone would feel (Forecasters) in response to this request. This time, participants in both conditions indicated whether “a reasonable person” would hand over their bag (Yes/No). Exploratory measures In lieu of reporting how they use their phones, participants reported how they use their bags. They were asked, “Do you have anything personal, sensitive, or expensive in your bag that you would rather not have the experimenter touch” (Yes/No) and, if they answered affirmatively, to specify. They also answered demographic questions. As in Study 1, before exiting the study participants were compen- sated, thanked for their participation, and debriefed. During the deb- riefing, we confirmed that participants did not recognize the experimenter from another context and had not heard anything about the study beforehand. CONSENT SEARCHES AND COMPLIANCE 17 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Results Bag use 34.04% of participants reported that they had something personal, sensitive, or expensive in their bags that they “would rather not have the experimenter touch.” Common items were laptops, headphones, wallets, and feminine hygiene products. Compliance behavior As predicted, fewer Forecasters (63.83%) predicted that a reasonable person would agree to hand over their backpack or bag than Experiencers, whose com- pliance rate was 100% (41 out of 41). This difference was significant, χ2(1, N = 94) = 18.38, p < 0.001, φ = 0.47 (Figure 2). Forecasters also indicated whether they personally would agree to this request. Again, fewer Forecasters (74.47%) predicted they personally would agree to a search of their bags. This was significantly lower than the 100% compliance rate observed among Experiencers, χ2(1, N = 94) = 11.56, p < 0.001, φ = 0.38 (Figure 2). As an exploratory measure, Experiencers also reported what a reasonable person would do: 87.23% said a reasonable person would comply. Compared to 63.83 % 74.47 % 100.00 % 0.00 0.25 0.50 0.75 1.00 Would a reasonable person comply? Would you comply? Did you comply? P ro po rt io n co m pl yi ng Forecasters Experiencers Study 2: Forecasted vs. Actual Compliance Behavior F I G U R E 2 All Experiencers (100.00%) complied with the request to search their bag, while significantly fewer Forecasters (75.47%) predicted they would comply or that a reasonable person would comply (63.83%). Error bars represent bootstrapped 95% confidence intervals for outcomes in which the variance exceeded zero. 18 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Forecasters’ beliefs about reasonable compliance (63.83%), this is significantly higher, χ2(1, N = 94) = 5.76, p = 0.016, φ = 0.27. Within subjects, Forecasters were no more likely to report that they person- ally would hand over their phones than that a reasonable person would hand over their phone, McNemar’s χ2(1) = 1.45, p = 0.23. Experiencers were signifi- cantly more likely to comply than to say a reasonable person would comply, McNemar’s χ2(1) = 30.19, p < 0.001. Feelings of freedom We again combined participants’ ratings of how pressured, free, comfortable, awkward, and easy it would be (for Forecasters) or was (for Experiencers) to refuse the experimenter’s request into a single measure of feelings of freedom (α = 0.72) and conducted a one-tailed t-test assessing the directional prediction that Forecasters would indicate greater feelings of freedom than Experiencers. Contrary to our pre-registered prediction, Forecasters did not imagine feeling more free (M = 3.24, SD = 0.89) than Experiencers actually reported feeling (M = 3.29, SD = 0.91), tone-tailed(92) = �0.27, p = 0.61, d = 0.06. As before, (i) compliance behavior among Experiencers; (ii) forecasted com- pliance among Forecasters; and (iii) feelings of freedom did not differ based on which of the three experimenters participants interacted with. Moreover, none of these measures differed by participants’ self-reported race/ethnicity, gender, or age, with one exception: younger Experiencers reported feeling significantly less free to refuse than did older Experiencers. We did not anticipate this result or pre-register this analysis, and it did not replicate in Studies 1 and 3. No signif- icant interaction between age and condition was observed. Comfort generally with one’s bag being searched Participants in both conditions were asked how comfortable they were with the idea of the experimenter checking their bags (1 = Not at all; 7 = Extremely). Forecasters’ average rating was 3.77 (SD = 1.39), while Experiencers’ average rating was 4.00 (SD = 1.71), t(92) = 1.97, p = 0.052, d = 0.15. Discussion Study 2 replicated the key finding with a different item as the target of the sea- rch, showing that the “digital is different” principle emerging in courts and legis- latures (Weisburd, 2019, p. 717) is no reason to expect that the psychology of compliance operates differently for physical personal objects. Indeed, bags do CONSENT SEARCHES AND COMPLIANCE 19 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense not receive the heightened legal solicitude that smartphones do, yet participants underestimate the pressure to comply with search requests in this arena as well. Surprisingly, we observed no difference between conditions in feelings of freedom. It is possible that due to the familiarity of this request—our partici- pants were likely to have much more experience being asked to allow a search of their bags than their phones—participants in the “Forecaster” condition were able to draw from recent real-world experiences to answer these questions, lead- ing to null results between the conditions on these items. In line with this possi- bility, it is worth noting that even Forecasters’ predictions of compliance were much higher (i.e., were more accurate) for the bag search than the phone search in Study 1 (64% vs. 27%). Despite this, Forecasters still failed to anticipate the extraordinarily high actual compliance rate of 100% that was observed among Experiencers in this study. This result indicates that although Forecasters answered the psychological freedom questions no differently from Experiencers, they expected some Experiencers to refuse, when none in fact did. STUDY 3: REPLICATION AND EXTENSION WITH A DEMOGRAPHICALLY DIVERSE SAMPLE A disproportionate number of people who are stopped and asked to submit to consent searches are racial minorities (Baumgartner et al., 2020; Gau, 2013; McGlinchy, 2018; Pierson et al., 2020). As such, the racialized nature of policing in the United States is important to acknowledge. Because previous research on the voluntariness of consent has been conducted on samples that are predomi- nantly white and Asian undergraduates (Sommers & Bohns, 2019), replicating the results with a more representative sample is critical (Ceci et al., 2010; Henrich et al., 2010; Marshall et al., 2019). Study 3 included some additional exploratory measures designed to investi- gate potential psychological mechanisms underlying the underestimation- of-compliance phenomenon. We asked participants to rate the degree to which they contemplated (or would contemplate, hypothetically) the risks and benefits of handing over their phones, and the extent to which they responded without really thinking. We also asked Experiencers to report what they thought a rea- sonable person would do. This question allowed us to investigate whether, after having gone through the task, participants judged compliance to be reasonable. In addition, we asked both Experiencers and Forecasters to guess what per- centage of Experiencers would end up complying with the request, and we paid them a monetary bonus if they guessed accurately (within one percentage point of the correct answer). The purpose of this task was to incentivize accuracy (i.e., focus participants’ attention on what outcome they thought was most likely). We wanted them to try their best to predict, not express their opinion or wishes about the search request. 20 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Method Pre-registration This study was pre-registered as AsPredicted.com (#18998; https://aspredicted. org/8ce4r.pdf). Participants Participants were drawn from a community sample in downtown Chicago. We aimed to recruit 200 participants; due to lab scheduling, we collected data from 203. Per our pre-registration, we excluded from further analyses any participant who did not have a working personal smartphone in English and any partici- pant who recognized the experimenter, yielding a final sample of n = 187. Participants were 55% male, 44% female, 1% choosing another designation; 69% Black, 20% White, 6% Hispanic/Latinx, 4% Asian/Asian-American, and 1% choosing another or multiple designations. Black men made up the largest demo- graphic subgroup (n = 79, 38.73%), followed by Black women (n = 58, 28.43%), white women (n = 22, 10.78%) and white men (n = 19, 9.31%). Participants’ ages ranged from 19 to 88 years, and the median age was 36 years. 28% had obtained post-secondary education, 19% had graduated from college or vocational school; 38% had obtained some college education, 14% had obtained high school or equiv- alent, and 2% had obtained less than high school or equivalent. Participants were paid a guaranteed $10. They could additionally earn a $1 bonus depending on how they performed on the prediction task. Procedure The procedure was largely the same as in Studies 1 and 2, with a few key changes. As this was a different lab in a different city, the experimenters were always one of seven research assistants (five female, two male) dressed in their regular attire. As before, experimenters recorded notes on their interactions with Experiencers. This time, the request presented to Experiencers and, hypothetically, to Forecasters was: “Before we begin the study, can you please unlock your phone and hand it to me? I’ll just need to take your phone outside of the room for a moment to check for some things.” This same script was used in prior research by Sommers and Bohns (2019, p. 1962). As before, participants in both conditions filled out a questionnaire CONSENT SEARCHES AND COMPLIANCE 21 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense http://aspredicted.com https://aspredicted.org/8ce4r.pdf https://aspredicted.org/8ce4r.pdf asking whether they had agreed (Experiencers) or would personally agree (Forecasters) to hand over their phone in response to the experimenter’s request (Yes/No). They also rated how easy, comfortable, awkward, free, and pressured they had felt, or would feel, in response to this request. Exploratory measures Next, as exploratory measures, participants in both conditions reported whether a “reasonable person” would hand over their phone in this scenario, as well as how easy, comfortable, awkward, pressured, and free to refuse a reasonable per- son would have felt. Next, participants reported sociodemographic information. Then, they reported how mindlessly they behaved (Experiencers) or would behave (Forecasters) using two 7-point scales measuring the extent to which they handed over their phones without thinking and whether they considered the benefits and risks of handing over their phones. We also asked how much they trusted the experimenter whom they met that day. Incentivized prediction task Next, participants were asked to make a prediction about how many partici- pants would comply with the search request: “Imagine that the researcher you met today approached 100 participants like yourself and asked them to hand over their phones. What percent of participants do you think would agree to hand over their phone?” Participants were offered a bonus payment if their guess fell within one percentage point of the correct answer. For both groups, this question came at the end of the survey; thus, Experiencers answered this question minutes after having been personally subjected to this same script. Debriefing Finally, participants in both conditions were paid, thanked for their participa- tion, and debriefed. They were informed that their interaction with the experi- menter had been videotaped and were given an opportunity to request that their tapes and/or data be deleted. As before, we confirmed that participants did not recognize the experimenter from another context and had not heard anything about the study beforehand. Results Phone use Most participants reported using their phones for personal messaging (91.44%), email (93.05%), storing contacts (87.70%), photos (87.17%), banking (68.98%), playing games (58.82%), and shopping (57.75%). Nearly half (41.71%) reported using their phones for storing sensitive information such as social security 22 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense numbers and passwords. Thus, it appears that for a majority of participants, granting unsupervised access to their unlocked personal phones would amount to a significant risk to their privacy. Predicted versus actual compliance rates Our key hypothesis was that Forecasters would imagine refusing to hand over their phones at greater rates than observed among Experiencers. Only 61.05% of Forecasters predicted they would hand over their phones, and only 46.32% thought a reasonable person would do so. Among Experiencers, 87 out of 92 participants complied with the experimenter’s request, yielding a compliance rate of 94.57%. This difference was significantly higher than Forecasters’ projec- tions, χ2(1, N = 187) = 28.25, p < 0.001, φ = 0.40 (Figure 3). Experiencers also reported what they thought a “reasonable person” would do in response to the experimenter’s request: 76.09% thought a reasonable person would comply. This was significantly higher than Forecasters’ assessment of what a reasonable person would do, χ2(1, N = 187) = 16.18, p < 0.001, φ = 0.31. Looking within subjects, we observe that a greater percentage of Experiencers complied than thought would be reasonable, χ2(1, N = 92) = 13.47, p < 0.001 (McNemar’s test). Of the 87 Experiencers who themselves chose to hand over their phones, 18 thought a reasonable person would refuse, indicating that 20.69% of compliers thought their compliance had been unrea- sonable. Of the 5 Experiencers who refused, 4 thought a reasonable person would similarly refuse and 1 thought a reasonable person would comply. 46.32 % 61.05 % 94.57 % 0.00 0.25 0.50 0.75 1.00 Would a reasonable person comply? Would you comply? Did you comply? P ro po rt io n co m pl yi ng Forecasters Experiencers Study 3: Forecasted vs. Actual Compliance Behavior F I G U R E 3 Most Experiencers (94.57%) complied with the search request, while far fewer Forecasters (61.05%) predicted they would comply or that a reasonable person would comply (46.32%). Error bars represent bootstrapped 95% confidence intervals. CONSENT SEARCHES AND COMPLIANCE 23 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Among Forecasters, a greater percentage imagined that they personally would comply than reported would be reasonable to comply, χ2(1, N = 95) = 7.04, p = 0.008 (McNemar’s test). Of the 58 Forecasters who imagined they personally would comply, 19 thought a reasonable person would refuse, indi- cating that 32.76% thought they would comply unreasonably. (Of the 37 Fore- casters who imagined they would refuse, 5 thought a reasonable person would comply.) Feelings of freedom We combined participants’ ratings of how pressured, free, comfortable, awkward, and easy it would be (for Forecasters) or was (for Experiencers) to refuse the experi- menter’s request into a single measure of feelings of freedom (α = 0.64). Consistent with our prediction, Forecasters imagined feeling more free (M = 4.92, SD = 1.26) than Experiencers actually reported feeling (M = 4.58, SD = 1.42), tone-tailed(185) = 1.75, p = 0.041, d = 0.26. In judging how free a reasonable person would feel, no difference was observed between Experiencers (M = 4.38, SD = 1.12) and Fore- casters (M = 4.62, SD = 1.35), t(185) = 1.31, p = 0.19, d = 0.19. Within subjects, Experiencers did not report feeling more or less free than they imagined a reasonable person would tpaired(91) = 1.51, p = 0.13. Fore- casters thought they personally would feel more free than they imagined a rea- sonable person would feel, tpaired(94) = 2.16, p = 0.03. As before, we observed no significant individual experimenter effects on (i) compliance behavior among Experiencers; (ii) forecasted compliance among Forecasters; or (iii) feelings of freedom. Nor did any of these mea- sures differ based on participants’ self-reported race/ethnicity, gender, or age. Mindlessness and trust We averaged ratings for the questions “To what extent did you respond to the experimenter’s request without thinking” and “To what extent did you con- sider the potential risks and benefits of handing over your phone before decid- ing whether to do so?” (reverse-scored) to create a composite measure of mindlessness. Experiencers (M = 4.28, SD = 1.42) reported more mindless- ness than Forecasters predicted (M = 3.77, SD = 1.47), t(184) = 2.44, p = 0.015, d = 36. Experiencers (M = 5.45, SD = 1.56) also expressed greater trust in the experimenter than did Forecasters (M = 4.60, SD = 1.91), t(184) = 3.31, p = 0.001, d = 0.49. 24 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense Incentivized prediction As predicted, Experiencers’ estimates (67.30%, SD = 28.42) are significantly higher than Forecasters’ estimates (43.79%, SD = 32.37), t(185) = 5.27, p < 0.001, d = 0.77. Both groups, however, underestimated compliance com- pared to the true observed compliance rate, which (as reported above) was 94.57% (Figure 4). Forecasters’ average estimate of 43.79% is thus a significant underprediction, t(94) = 15.29, p < 0.001, d = 1.57. Experiencers’ estimate of 67.30% is also a significant underprediction, t(91) = 9.20, p < 0.001, d = 0.96. Discussion Study 3 demonstrates that the main finding—the overwhelming majority of Experiencers comply, and Forecasters significantly underestimate compliance— replicates among a more diverse sample of participants that more closely mir- rors the population that is disproportionately subjected to consent searches in the United States. Participants were older, more likely to be African-American, and less likely to have attended college; still, they largely reported using their smartphones for personal messaging, banking, and other private matters. Whereas 95% of Experiencers complied with the experimenter’s request, 61% of Forecasters thought they personally would do the same. The 61% figure is higher than observed in Study 1, but Forecasters still significantly 67.30 % 43.79 % True ComplianceTrue Compliance 0.00 0.25 0.50 0.75 1.00 Of 100 participants, how many would agree to hand over their phones? A ve ra ge e st im at e Forecasters Experiencers Estimates of Compliance F I G U R E 4 Experiencers and Forecasters estimated the percentage of participants who would comply with the search request. Experiencers’ average guess (67.30%) significantly exceeded Forecasters’ (43.79%), but both groups significantly underestimated compliance when compared to true compliance—the percentage of Experiencers in the experiment who ultimately complied (94.57%). Error bars represent bootstrapped 95% confidence intervals. CONSENT SEARCHES AND COMPLIANCE 25 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense underestimated compliance. They also overstated how free they would feel to refuse, as compared to how Experiencers actually felt. We can compare these results to those observed in Sommers and Bohns’s (2019) research (Study 1), which used the same script with participants drawn from a campus population. Experiencers complied at roughly the same rates (94% in the present study vs. 97% in the prior research), suggesting that the results are not attributable to students’ being exceptionally vulnerable to social influence. Forecasted compliance, however, differed between the two studies: whereas only 28% of the student Forecasters in the earlier study imagined they would comply, 62% of the present sample’s Forecasters imagined complying. Finally, when it came to feelings of freedom, the present sample gave higher overall ratings than the student sample, a trend that was observed among both Forecasters and Experiencers.5 It is difficult to pinpoint the cause(s) of these dif- ferences between studies; they could be due to the variation in geography (Midwest vs. Northeast); study milieu (an urban setting vs. a semirural univer- sity campus); time period (the original studies were conducted in 2018, while the present studies were conducted between 2019 and 2021); the divergent ways in which people from different populations tend to use their smartphones; or myr- iad other differences between the two studies. Still, within each population and context, the primary prediction is borne out: those who merely imagine a con- sent interaction underestimate the likelihood of compliance and the perceived pressure to comply, as compared to those who actually experience it. Study 3 expands upon prior work by beginning to examine potential psycho- logical mechanisms that explain why Forecasters mispredict compliance. It seems that Forecasters commit an error whereby they imagine that they would contemplate the risks and benefits before deciding, and imagine that they would not answer mindlessly. But when Experiencers are actually put in the sit- uation, they do (by their own self-report) answer mindlessly and without think- ing through the risks and benefits. By asking Experiencers to report whether a reasonable person would com- ply, Study 3 reveals that 21% of compliers thought that their own compliance was unreasonable. For this group, the decision to consent may be one that, upon reflection, they regretted. Most compliers, however, seemed to believe that what they had done was reasonable. This stands in stark contrast to Forecasters, a majority of whom asserted that compliance would have been unreasonable. The divergence could reflect Experiencers’ rational updating in response to new information (e.g., “I complied and nothing bad ultimately happened, so it was reasonable”). It is also consistent with attitude change following cognitive dissonance (Elliot & Devine, 1994; Festinger, 1962). The same can be said for participants’ ratings of how much they trusted the experimenter: Experiencers’ 5The prior research reported M = 3.28 (SD = 0.95) among Experiencers and M = 4.14 (SD = 1.44) among Forecasters, n = 203 (Sommers & Bohns, 2019, p. 1986). 26 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense significantly higher trust ratings could reflect rational updating in light of the experimenter’s unobjectionable handling of Experiencers’ phones, or it could reflect cognitive dissonance (e.g., having behaved as if they trust the researcher, they shifted their attitudes such that they came to regard the researcher as more trustworthy). It is also possible that Experiencers’ heightened trust stemmed from their slightly more involved interaction with the experimenters. The incentivized prediction result suggests that having experienced the con- sent request oneself—just minutes earlier—makes Experiencers more accurate than Forecasters at estimating the overall level of compliance. However, per- sonal experience does not fully “debias” people. Thus, we can expect that even if every courtroom were staffed with judges who had themselves been approached by police and asked to submit to a consent search, these decision-makers would still underestimate the prevalence of compliance in the general population. With that said, this group might forecast more accurately than a group that had never personally experienced such requests. GENERAL DISCUSSION These findings extend previous research by demonstrating that the underestimation-of-compliance effect is robust to the type of search, to the con- sequences of the search, and to a sample that is more demographically represen- tative of the population that tends to be subjected to consent searches. In addition, the novel incentivized prediction task demonstrates that even when decision-makers are focused on accuracy, they still significantly underestimate the likelihood of compliance. This latter finding suggests that Forecasters are not strategically posturing when they assert that a reasonable person would refuse to allow the search; they are genuinely misperceiving what the average person is likely to do. Underestimation of compliance is thus an error of social perception—a failure to understand how people respond to being asked to con- sent to a search. These findings carry several implications for legal reform. First, the findings underscore that the legal standard for determining the voluntariness of consent is subject to a systematic bias: uninvolved fact-finders are likely to overstate the ease of refusal and to underestimate the psychological pressure to comply. Second, the results call into question whether diversifying the bench is likely to address this social perception bias. The ideological makeup of the judiciary has been criticized as heavily pro-police (e.g., Lvovsky, 2016; Neidig, 2021; Neily, 2019). While addressing those criticisms may enhance the legitimacy and fairness of the judicial system, we suggest that it will not address the particular problem identified here, which runs deeper than ideology. The results of the incentivized prediction task in Study 3 suggest that even a decision-maker who is motivated solely by a desire to project accurately how another person feels is CONSENT SEARCHES AND COMPLIANCE 27 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense still likely to understate the psychological pressure to comply. Thus, even a per- fectly neutral decision-maker is likely to overstate the voluntariness of consent, due to the ordinary workings of their social psychology. Finally, as Study 3 illus- trates, racially diverse decision-makers are not immune from the bias. Importantly, Experiencers who had just been through the consent-search task gave more accurate estimates of compliance than did Forecasters. This result suggests that prior personal experience may mitigate the bias by giving participants an appreciation for how commonplace compliance is. Nonetheless, even these participants still significantly underestimated compliance. Whereas Forecasters’ best guess was that 1 in 2 would refuse, Experiencers’ best guess was that 1 in 3 would refuse. The correct answer was that just over 1 in 20 refused. The big picture The primary lesson that emerges from this research is not to miss the forest for the trees. While there may be psychologically interesting moderators to be dis- covered based on the particular script used, the particular task presented, and the particular population studied, the overarching insight from this line of work is that people are overwhelmingly compliant. People are so compliant that Experiencers who have just been through the task themselves significantly underestimate the level of compliance. The main effect of compliance research—that people comply a lot—is surprising (as demonstrated by Fore- caster responses), large in magnitude, and replicable across a variety of conditions. In light of this overarching insight, we can return to policing’s embrace of searches based on voluntary consent, as opposed to evidence (e.g., probable cause or reasonable suspicion) bearing on the likelihood that the suspect is in posses- sion of contraband. This shift, we can observe, represents a massive redistribu- tion of power toward the police—the magnitude of which is difficult for people to fathom. Indeed, police officers themselves have expressed surprise at how easy it is for them to obtain consent from individuals. In the late 1980s, when the New Jersey state police first developed a practice of stopping every driver who committed a moving violation along a certain stretch of highway, newspa- per reports relayed that officers were “surprised at how often and readily motor- ists gave consent” to searches for drugs (Rotenberg, 1991, p. 190). After interviewing detectives in seven cities in 1986, a research team with the National Center for State Courts observed, “Listening to law enforcement officers …, one would think consent was the easiest thing in the world to come by” (Sutton, 1986, p. 41). As described earlier, consent searches represent a distinct justification for government searches. Whereas searches based on probable cause or reasonable 28 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense suspicion tend to be justified based on evidence presented to a public official suggesting that the individual is hiding something illegal, consent-based searches are justified based on the voluntary choice of the individual who accedes to the search. As the present studies make clear, a legal regime in which the legality of searches rests on consent is one in which most people will comply with police requests when asked. With this psychological reality in view, we can appreciate just how significant the rise of consent searches6 has proven to be, practically speaking. A legal regime in which private individuals are left to protect them- selves, by refusing requests that are presented to them, is a regime in which vastly more searches will be carried out. Thus, a turn toward individual choice is, in effect, a turn toward police power. Existing consent-search jurisprudence pays exquisite attention to the specific facts of each case, as judges seek to determine whether the situation coerced the individual to comply (Schneckloth v. Bustamonte, 1973). But, we can see from traffic stop data that actual variation in compliance behavior is quite low: upwards of 95% of motorists consent to police searches when asked (State v. Carty, 2002). Analyzing the specifics of each situation is required by the “totality of the circumstances” standard, but we can surmise that there is much less variation in compliance behavior than might be expected based on individ- ual variation in internal preferences regarding searches. Indeed, research from social psychology dramatizes just how little variation there is likely to be in peo- ple’s responses to such requests; the “situationist” literature from the 1970s and 1980s foretells the relative uniformity in individuals’ responses to powerful forms of social influence. As Ross (1977) observed, “many of the best known and most provocative studies in our field depend, for their impact, upon the reader’s erroneous expectation that individual differences and personal disposi- tions will overcome relatively mundane situational variables,” such as the pres- sure to conform to a group (Asch, 1955) or comply with a request (Milgram, 1963). Limitations The current studies have attempted to address several key ways in which Sommers and Bohns’s (2019) studies differed from a real police search context, finding that the original results are robust to particularly objectionable searches, searches of physical rather than merely digital objects, and to a more representa- tive participant sample. Nonetheless, a laboratory context unavoidably differs from a real search context in numerous respects. Most saliently, a laboratory is 6Sources dating back to the 1950s and 1960s expressed concern about “the increasing use of the consent search” (Dittmeier, 1967, p. 300) and called for “courts [to] place their stamp of disapproval upon this increasing practice of federal officers searching a home without a warrant on the theory of consent” (United States v. Arrington, 1954, p. 637). CONSENT SEARCHES AND COMPLIANCE 29 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense inevitably a safer and more familiar environment for most requestees than the milieu surrounding a police stop. Still, our studies placed both Forecasters and Experiencers in the same relatively safe environment, and found sizeable differ- ences between the two perspectives. Furthermore, the present research furnishes a more analogous context than the Milgram studies, as well as the original Sommers and Bohns studies, did previously. Tellingly, both prior studies have been cited by numerous courts considering real searches conducted by police, suggesting that despite the limitations of laboratory research, courts have found the empirical research to be illuminating (People v. Tacardon, 2022; State v. Carty, 2002; State v. Diego, 2021; State v. Hauge, 2022; State v. Turnquest, 2019; United States v. Meadows, 2021). Of further note, because of our specific interest in consent searches as they have been legally defined and implemented in the United States, we have focused on American samples. Whether the underlying tendency to underesti- mate the voluntariness of consent generalizes beyond this specific context remains to be investigated. Previous research has examined whether the underestimation-of-compliance effect persists across cultural differences: Bohns et al. (2011) compared participants recruited in China to those recruited in the United States and found that although the effect replicated, it was attenuated in the Chinese sample—a result that aligns with prior work on social influence in collectivistic cultures. Accordingly, it is possible we would find a similar pat- tern of results if the current studies were run in a more collectivistic cultural setting. CONCLUSION In summary, the robustness of the findings—across tasks, scripts, and populations—underscores that we should not miss the forest for the trees. Peo- ple are highly compliant. The magnitude of their compliance is difficult to appreciate. While numerous research questions remain about boundary condi- tions, individual differences, and mechanisms, we can see that the decision to embrace consent as the constitutional basis for police searches represents a mas- sive power giveaway to the police, enabling them to search more people than ever before, on a scale that would surprise judges, policymakers, members of the public, and even the police themselves. ACKNOWLEDGMENTS The authors wish to thank Shay Elbaum, Anjali Baliga, Emily Dawson, Taylor Galdi, Charlotte Gemperle, Eshaan Jain, Courtney Noll, Pierson Ohr, Nethra Raman, Emily Smid, Mae Veidlinger, Mike White, Vicki Xie, and Warda Yousuf for invaluable research assistance. We are grateful to the staff and research assistants at the PIMCO Decision Research Laboratories at the 30 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense University of Chicago Booth’s Center for Decision Research, especially Bryan Baird and Kaushal Addanki. We received helpful comments from Brian Bornstein; Adi Liebovitch; Gabriel Mendlow; Janice Nadler; Eve Primus; and participants at the 2022 Conference on Empirical Legal Studies, the American Bar Foundation Speaker Series, and the seminar on Social, Behavioral, and Experimental Economics held by the School of Information at the University of Michigan. This research was funded by NSF Grant #1823661. DATA AVAILABILITY STATEMENT Data necessary to replicate the results of this article are available on the Open Science Framework (https://osf.io/4t7ym/) or upon request from the corresponding author. REFERENCES Ahrens, D. (2000). Not in front of the children: Prohibition on child custody as civil branding for criminal activity. NYU Law Review, 75, 737. Asch, S. E. (1955). Opinions and social pressure. Scientific American, 193(5), 31–35. Bar-Gill, O., & Friedman, B. (2012). Taking warrants seriously. Northwestern University Law Review, 106, 1609. Barrio, A. J. (1997). Rethinking Schneckloth v. Bustamonte: Incorporating obedience theory into the Supreme Court’s conception of voluntary consent. University of Illinois Law Review, 215, 215–251. Baumgartner, F. R., Epp, D. A., Shoub, K., & Love, B. (2020). Targeting young men of color for search and arrest during traffic stops: Evidence from North Carolina, 2002–2013. In The poli- tics of protest (pp. 188–212). Routledge. Bohns, V. K. (2016). (Mis)understanding our influence over others: A review of the underestimation-of-compliance effect. Current Directions in Psychological Science, 25(2), 119– 123. https://doi.org/10.1177/0963721415628011 Bohns, V. K., & Flynn, F. J. (2015). Empathy gaps between helpers and help-seekers: Implications for cooperation. In Emerging trends in the social and behavioral sciences (pp. 1–15). John Wiley & Sons, Ltd.. https://doi.org/10.1002/9781118900772.etrds0113 Bohns, V. K., Handgraaf, M. J. J., Sun, J., Aaldering, H., Mao, C., & Logg, J. (2011). Are social prediction errors universal? Predicting compliance with a direct request across cultures. Journal of Experimental Social Psychology, 47(3), 676–680. https://doi.org/10.1016/j.jesp.2011.01.001 Bohns, V. K., Roghanizad, M. M., & Xu, A. Z. (2014). Underestimating our influence over others’ unethical behavior and decisions. Personality and Social Psychology Bulletin, 40(3), 348–362. Buhi, E. R., Daley, E. M., Fuhrmann, H. J., & Smith, S. A. (2009). An observational study of how young people search for online sexual health information. Journal of American College Health, 58(2), 101–111. Burke, A. S. (2015). Consent searches and Fourth Amendment reasonableness. Florida Law Review, 67, 509. Ceci, S. J., Kahan, D. M., & Braman, D. (2010). The WEIRD are even weirder than you think: Diversifying contexts is as important as diversifying samples. Behavioral and Brain Sciences, 33(2–3), 87–88. Correll, J., Hudson, S. M., Guillermo, S., & Ma, D. S. (2014). The police officer’s dilemma: A decade of research on racial bias in the decision to shoot. Social and Personality Psychology Compass, 8(5), 201–213. CONSENT SEARCHES AND COMPLIANCE 31 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense https://osf.io/4t7ym/ https://doi.org/10.1177/0963721415628011 https://doi.org/10.1002/9781118900772.etrds0113 https://doi.org/10.1016/j.jesp.2011.01.001 Crosby, J. R., & Wilson, J. (2015). Let’s not, and say we would: Imagined and actual responses to witnessing homophobia. Journal of Homosexuality, 62(7), 957–970. Dittmeier, T. E. (1967). Consent search: Waiver of Fourth Amendment rights. Saint Louis University Law Journal, 12, 297. Donders, N. C., Correll, J., & Wittenbrink, B. (2008). Danger stereotypes predict racially biased attentional allocation. Journal of Experimental Social Psychology, 44(5), 1328–1333. Eith, C. A., & Durose, M. R. (2011). Contacts between police and the public, 2008. US Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. Elliot, A. J., & Devine, P. G. (1994). On the motivational nature of cognitive dissonance: Dissonance as psychological discomfort. Journal of Personality and Social Psychology, 67(3), 382–394. https://doi.org/10.1037/0022-3514.67.3.382 Epley, N., Keysar, B., Van Boven, L., & Gilovich, T. (2004). Perspective taking as egocentric anchoring and adjustment. Journal of Personality and Social Psychology, 87(3), 327–339. Festinger, L. (1962). A theory of cognitive dissonance. Stanford University Press. Florida v. Bostock. (1991). 501 U.S. 429. Flynn, F. J., & Lake, V. K. (2008). If you need help, just ask: Underestimating compliance with direct requests for help. Journal of Personality and Social Psychology, 95(1), 128–143. Fujita, K. (2008). Seeing the forest beyond the trees: A construal-level approach to self-control. Social and Personality Psychology Compass, 2(3), 1475–1496. Gau, J. M. (2013). Consent searches as a threat to procedural justice and police legitimacy: An analysis of consent requests during traffic stops. Criminal Justice Policy Review, 24(6), 759–777. Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2–3), 61–83. Kagehiro, D. K. (1988). Perceived voluntariness of consent to warrantless police searches. Journal of Applied Social Psychology, 18(1), 38–49. Kagehiro, D. K. (1990). Psycholegal research on the Fourth Amendment. Psychological Science, 1(3), 187–193. https://doi.org/10.1111/j.1467-9280.1990.tb00196.x Kagehiro, D. K., Taylor, R. B., Laufer, W. S., & Harland, A. T. (1991). Hindsight bias and third- party consentors to warrantless police searches. Law and Human Behavior, 15(3), 305–314. https://doi.org/10.1007/BF01061715 Kawakami, K., Dunn, E., Karmali, F., & Dovidio, J. F. (2009). Mispredicting affective and behav- ioral responses to racism. Science, 323(5911), 276–278. Kawakami, K., Karmali, F., & Vaccarino, E. (2019). Confronting intergroup bias: Predicted and actual responses to racism and sexism. In Confronting prejudice and discrimination (pp. 3–28). Elsevier. LaFave, W. R. (2007). Criminal procedure (Vol. 2, 4th ed.). Thomson West. Lvovsky, A. (2016). The judicial presumption of police expertise. Harvard Law Review, 130(8), 1995–2081. Maclin, T. (2008). The good and bad news about consent searches in the Supreme Court. McGeorge Law Review, 39, 56. Marshall, E. W., Groscup, J. L., Brank, E. M., Perez, A., & Hoetger, L. A. (2019). Police surveil- lance of cell phone location data: Supreme Court versus public opinion. Behavioral Sciences & the Law, 37(6), 751–775. McGlinchy, J. C. (2018). “Was that a yes or a no?” reviewing voluntariness in consent searches. Vir- ginia Law Review, 104, 301. Milgram, S. (1963). Behavioral study of obedience. The Journal of Abnormal and Social Psychology, 67(4), 371–378. Milgram, S. (1974). Obedience to authority: An experimental view (1st ed.). Harper & Row. Nadler, J. (2002). No need to shout: Bus sweeps and the psychology of coercion. The Supreme Court Review, 2002, 153–222. 32 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense https://doi.org/10.1037/0022-3514.67.3.382 https://doi.org/10.1111/j.1467-9280.1990.tb00196.x https://doi.org/10.1007/BF01061715 Nadler, J., & Trout, J. D. (2012). The language of consent in police encounters. In The Oxford hand- book of language and law. Oxford University Press. https://doi.org/10.1093/oxfordhb/ 9780199572120.013.0024 Neidig, H. (2021). Biden under pressure to pick new breed of federal prosecutors. The Hill https:// thehill.com/homenews/administration/562364-biden-under-pressure-to-pick-new-breed-of- federal-prosecutors/ Neily, C. (2019). Are a disproportionate number of federal judges former government advocates? Cato Institute. https://www.cato.org/study/are-disproportionate-number-federal-judges-former- government-advocates Nordgren, L. F., Banas, K., & MacDonald, G. (2011). Empathy gaps for social pain: Why people underestimate the pain of social suffering. Journal of Personality and Social Psychology, 100(1), 120–128. Nordgren, L. F., McDonnell, M.-H. M., & Loewenstein, G. (2011). What constitutes torture? Psy- chological impediments to an objective evaluation of enhanced interrogation tactics. Psycho- logical Science, 22(5), 689–694. Pager, D. (2003). The mark of a criminal record. American Journal of Sociology, 108(5), 937–975. Pager, D., Western, B., & Sugie, N. (2009). Sequencing disadvantage: Barriers to employment facing young black and white men with criminal records. The Annals of the American Academy of Political and Social Science, 623(1), 195–213. People v. Tacardon. (2022). 14 Cal. 5th 235, 521 P.3d 563. Pierson, E., Simoiu, C., Overgoor, J., Corbett-Davies, S., Jenson, D., Shoemaker, A., Ramachandran, V., Barghouty, P., Phillips, C., & Shroff, R. (2020). A large-scale analysis of racial disparities in police stops across the United States. Nature Human Behaviour, 4(7), 736–745. Riley v. California. (2014). 573 U.S. 373. Roese, N. J., & Vohs, K. D. (2012). Hindsight bias. Perspectives on Psychological Science, 7(5), 411–426. Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortions in the attribution pro- cess. In Advances in experimental social psychology (Vol. 10, pp. 173–220). Elsevier. Rotenberg, D. L. (1991). An essay on consent(less) police searches. Washington University Law Quarterly, 69, 175. Sabini, J., Siepmann, M., & Stein, J. (2001). The really fundamental attribution error in social psy- chological research. Psychological Inquiry, 12(1), 1–15. https://doi.org/10.1207/ S15327965PLI1201_01 Schneckloth v. Bustamonte. (1973). 412 U.S. 218. Shermer, M. (2018). Web searches reveal (in aggregate) what we’re really thinking. Scientific Ameri- can https://www.scientificamerican.com/article/web-searches-reveal-in-aggregate-what-were- really-thinking/ Simmons, R. (2005). Not voluntary but still reasonable: A new paradigm for understanding the con- sent searches doctrine. Indiana Law Journal, 80, 773. Sobol, N. L. (2015). Charging the poor: Criminal justice debt and modern-day debtors’ prisons. Maryland Law Review, 75, 486. Sommers, R., & Bohns, V. K. (2019). The voluntariness of voluntary consent: Consent searches and the psychology of compliance. Yale Law Journal, 128, 1962–2033. State v. Carty. (2002). 174 N.J. 351, 806 A.2d 798. State v. Diego. (2021). 169 N.E. 113. State v. Hauge. (2022). 973 N.W.2d 453. State v. Turnquest. (2019). 305 Ga. 758, 827 S.E.2d 865. Stephens-Davidowitz, S. (2017). Everybody lies: How Google search reveals our darkest secrets. The Guardian. https://www.theguardian.com/technology/2017/jul/09/everybody-lies-how-google- reveals-darkest-secrets-seth-stephens-davidowitz CONSENT SEARCHES AND COMPLIANCE 33 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense https://doi.org/10.1093/oxfordhb/9780199572120.013.0024 https://doi.org/10.1093/oxfordhb/9780199572120.013.0024 https://thehill.com/homenews/administration/562364-biden-under-pressure-to-pick-new-breed-of-federal-prosecutors/ https://thehill.com/homenews/administration/562364-biden-under-pressure-to-pick-new-breed-of-federal-prosecutors/ https://thehill.com/homenews/administration/562364-biden-under-pressure-to-pick-new-breed-of-federal-prosecutors/ https://www.cato.org/study/are-disproportionate-number-federal-judges-former-government-advocates https://www.cato.org/study/are-disproportionate-number-federal-judges-former-government-advocates https://doi.org/10.1207/S15327965PLI1201_01 https://doi.org/10.1207/S15327965PLI1201_01 https://www.scientificamerican.com/article/web-searches-reveal-in-aggregate-what-were-really-thinking/ https://www.scientificamerican.com/article/web-searches-reveal-in-aggregate-what-were-really-thinking/ https://www.theguardian.com/technology/2017/jul/09/everybody-lies-how-google-reveals-darkest-secrets-seth-stephens-davidowitz https://www.theguardian.com/technology/2017/jul/09/everybody-lies-how-google-reveals-darkest-secrets-seth-stephens-davidowitz Strauss, M. (2001). Reconstructing consent. Journal of Criminal Law & Criminology, 92, 211. Sutherland, B. A. (2006). Whether consent to search was given voluntarily: A statistical analysis of fac- tors that predict the suppression rulings of the federal district courts. NYU Law Review, 81, 2192. Sutton, P. (1986). The Fourth Amendment in action: An empirical view of the search warrant pro- cess. Criminal Law Bulletin, 22, 405–421. U.S. Const. amend. IV. (n.d.). United States v. Arrington. (1954). 215 F.2d 630 (7th Cir.). United States v. Meadows. (2021). No. 6:21-CR-27-REW-HAI, 2021 WL 4782259 (E.D. Ky.). Van Boven, L., Loewenstein, G., & Dunning, D. (2005). The illusion of courage in social predic- tions: Underestimating the impact of fear of embarrassment on other people. Organizational Behavior and Human Decision Processes, 96(2), 130–141. Van Boven, L., Loewenstein, G., Dunning, D., & Nordgren, L. F. (2013). Changing places: A dual judgment model of empathy gaps in emotional perspective taking. In Advances in experimental social psychology (Vol. 48, pp. 117–171). Elsevier. Warren, P. Y., & Tomaskovic-Devey, D. (2009). Racial profiling and searches: Did the politics of racial profiling change police behavior? Criminology & Public Policy, 8(2), 343–369. Weisburd, K. (2019). Sentenced to surveillance: Fourth Amendment limits on electronic monitoring. North Carolina Law Review, 98(4), 717–778. Williams, D. R. (2007). Misplaced angst: Another look at consent-search jurisprudence. Indiana Law Journal, 82, 69. Wilson, J. P., Hugenberg, K., & Rule, N. O. (2017). Racial bias in judgments of physical size and formidability: From size to threat. Journal of Personality and Social Psychology, 113(1), 59–80. Wolf v. Colorado. (1949). 338 U.S. 25. Woodzicka, J. A., & LaFrance, M. (2001). Real versus imagined gender harassment. Journal of Social Issues, 57(1), 15–30. How to cite this article: Sommers, R., & Bohns, V. K. (2024). Consent searches and underestimation of compliance: Robustness to type of search, consequences of search, and demographic sample. Journal of Empirical Legal Studies, 21(1), 4–34. https://doi.org/10.1111/jels.12375 34 CONSENT SEARCHES AND COMPLIANCE 17401461, 2024, 1, D ow nloaded from https://onlinelibrary.w iley.com /doi/10.1111/jels.12375 by C ornell U niversity, W iley O nline L ibrary on [09/12/2025]. See the T erm s and C onditions (https://onlinelibrary.w iley.com /term s-and-conditions) on W iley O nline L ibrary for rules of use; O A articles are governed by the applicable C reative C om m ons L icense https://doi.org/10.1111/jels.12375 Consent searches and underestimation of compliance:Robustness to type of search, consequences of search,and demographic sample