You Have the Right to Understand: The Deleterious Effect of Stress on Suspects’ Ability to Comprehend Miranda
- First Online:
- Cite this article as:
- Scherr, K.C. & Madon, S. Law Hum Behav (2011). doi:10.1007/s10979-011-9283-3
- 47 Views
Miranda v. Arizona (384 U.S. 436, 1966) required that suspects be explicitly warned of the right to avoid self-incrimination and the right to legal representation. This research was designed to examine whether stress, induced via an accusation of wrong-doing, undermined or enhanced suspects’ ability to comprehend their Miranda rights. Participants were randomly assigned to either be accused (n = 15) or not accused (n = 15) of having cheated on an experimental task in a two-cell between-subjects experimental design. Results supported the hypothesis that stress undermines suspects’ ability to comprehend their Miranda rights. Participants who were accused of cheating exhibited significantly lower levels of Miranda comprehension than participants who were not accused of cheating. The theoretical processes responsible for these effects and the implications of the findings for police interrogation are discussed.
The Supreme Court set a landmark criminal procedure precedent in the Miranda v. Arizona (1966) ruling. The Miranda ruling, based on the Fifth Amendment right against compelled self-incrimination and the Sixth Amendment right to counsel, was designed as a prophylactic safeguard against police intimidation during custodial questioning. However, debate over Miranda continues even 40 years after the ruling (Leo & Thomas, 1998; White, 2001). Central to this debate is concern that Miranda may not afford suspects the protections that were initially intended by the Supreme Court. For example, whereas many researchers have documented that adults of average intelligence have a satisfactory understanding of Miranda, these findings have been reported with the caveat that assessments of Miranda comprehension have most often been examined in low stress situations (e.g., Grisso, 1998). This fact has led social scientists and legal scholars to hypothesize that stress caused by police accusation (Gudjonsson, 2003; Irving, 1980) may undermine suspects’ ability to fully understand their constitutional rights (Kassin et al., 2007; Oberlander & Goldstein, 2001; Rogers, Gillard, Wooley, & Fiduccia, 2010).
The potentially deleterious effect of stress on Miranda comprehension has profound implications for the judicial system because it raises the possibility that Miranda may not be fully affording suspects the prophylactic safeguards it was designed to provide. In the current research, therefore, we experimentally examined how stress—induced via an accusation of wrong-doing—affected Miranda comprehension. We begin by discussing research relevant to Miranda comprehension after which we present the theoretical and empirical basis underlying our research.
Miranda Comprehension Research
Several key factors have been empirically demonstrated to influence Miranda comprehension. One such factor is reading difficulty. There are considerable differences in the reading difficulty of Miranda warnings across jurisdictions and these differences influence suspects’ comprehension (Rogers, 2008a). For example, recently detained adults showed significantly better comprehension of Miranda warnings when the warnings were written at 6th grade level than when they were written between an 8th and 12th grade level (Rogers, 2008b). The method by which Miranda warnings are administered also influences suspects’ comprehension. Suspects demonstrate better comprehension when Miranda warnings are administered in a written fashion than an oral fashion (Rogers, 2008b). It is also clear that cognitive deficits associated with mental retardation and substance use (Hazelwood, 2009) substantially undermine suspects’ ability to comprehend Miranda (e.g., Clare & Gudjonsson, 1991; Everington & Fulero, 1999; O’Connell, Garmoe, & Goldstein, 2005). Moreover, suspects with mental health conditions have been found to have a reduced ability to comprehend Miranda compared to individuals without mental health conditions (Cooper & Zapf, 2008).
An additional factor that has recently been demonstrated to affect Miranda comprehension is stress. Rogers and colleagues (2010) examined the influence of situational factors on individuals’ ability to comprehend Miranda, including stress evoked by an accusation of having committed a mock crime. In the study, participants either were or were not accused of having stolen a watch. Although participants were informed ahead of time that they might have to pretend to steal the watch, be apprehended, be read their Miranda rights, and have their understanding of these rights assessed, participants who were accused of stealing the watch nonetheless reported feeling more stress and demonstrated worse recall and reasoning on Miranda comprehension instruments than did participants who were not accused of stealing the watch.
The findings of this study are important because they provide initial evidence in support of the idea that stress can have a disadvantageous influence on suspects’ ability to understand their Miranda rights. However, because these findings emerged with respect to an accusation that pertained to a mock crime, additional evidence arising from research that involves an accusation of actual wrong-doing is needed to strengthen the claim that stress undermines suspects’ Miranda comprehension. This is particularly important given that an accusation of actual wrong-doing would likely elicit more stress than an accusation of having committed a mock crime and stress has previously been shown to have a complex relation to cognitive functioning, sometimes undermining it and other times enhancing it.
Stress and Cognitive Functioning
Several lines of research support the idea that stress can undermine cognitive functioning. According to processing efficiency theory (Eysenck, 1982, 1992, 1997), for instance, stress expends valuable cognitive resources that would otherwise be used for efficient cognitive functioning, and by so doing, compromises the efficiency of individuals’ working memory system (e.g., Derakshan & Eysenck, 1998; MacLeod & Donnellan, 1993). Empirical findings strongly support this prediction. A meta-analysis found that high levels of stress negatively impacted both the accuracy of eyewitness identifications and the recall accuracy of crime-related details (Deffenbacher, Bornstein, Penrod, & McGorty, 2004). Likewise, high levels of anxiety undermined eyewitness memory performance (e.g., Bothwell, Brigham, & Pigott, 1987; Deffenbacher, 1994; Peters, 1988) and, as mentioned above, stress, induced via a mock crime, reduced individuals’ ability to comprehend their Miranda rights (Rogers et al., 2010). Stress has also been shown to impair memory (Eichenbaum, Otto, & Cohen, 1992; LeDoux, 1995; Newcomer, Craft, Hershey, Askins, & Bardgett, 1994), compromise the processing of inaccessible information (Bargh & Thein, 1985; Pratto & Oliver, 1991), undermine performance (Lupien et al., 2005; MacKenzie, Smith, Hasher, Leach, & Behl, 2007; Sandstrom, Rhodin, Lundberg, Olsson, & Nyberg, 2005) and lead to a loss of distance cues in perception (Callaway & Thompson, 1953). Overall, therefore, there is myriad evidence demonstrating that stress can impair cognitive functioning, thus generating the hypothesis that stress caused by police accusation compromises suspects’ ability to comprehend Miranda.
However, it is also possible that stress may enhance Miranda comprehension. Theory and research dating back to the 1950s indicates that negative emotions—including stress and anxiety—cause people to narrow their attentional focus to central and relevant information. For example, stress causes people to focus more on cues that are central to the visual field than to cues that are peripheral to it (Chagut & Algom, 2003; Easterbrook, 1959). People experiencing anxiety better remember strongly related constructs than weakly related constructs (Derryberry & Tucker, 1994). And, anxiety, caused by threat-relevant words, leads people to focus more on threat-relevant information than threat-irrelevant information (Watts, McKenna, Sharrock, & Trezise, 1986). Accordingly, it is possible that the stress and anxiety that suspects experience as a result of police accusation could have an advantageous influence on their ability to comprehend their Miranda rights. Suspects could, for example, focus acutely on the content of Miranda warnings, thereby enhancing their understanding of them. To the extent that stress improves cognitive functioning, then the alternative hypothesis—that stress enhances Miranda comprehension—is generated.
The current research examined two competing predictions regarding the effect of stress on Miranda comprehension. Consistent with theory and research indicating that stress reduces cognitive functioning (e.g., Grisso, 1998; Kassin et al., 2007; Oberlander & Goldstein, 2001), we hypothesized that stress undermines Miranda comprehension. Alternatively, based on theory and research indicating that stress can improve cognitive functioning by virtue of causing individuals to narrow their attentional focus (e.g., Chagut & Algom, 2003; Derryberry & Tucker, 1994; Easterbrook, 1959), we hypothesized that stress enhances Miranda comprehension. We tested these competing hypotheses with an experiment in which we manipulated participants’ stress with an accusation of actual wrong-doing. The procedures we employed were adapted from a paradigm developed by Russano, Meissner, Narchet, and Kassin (2005). In the Russano et al. (2005) paradigm, participants are paired with a confederate with whom they complete logic problems. The experimenter instructs the pair to work independently on some problems and together on others. The confederate asks some participants (but not others) for help on a problem the pair was instructed to work on alone. Shortly after, the pair is separated and the participant is accused of cheating. In the current research, we modified this paradigm such that all participants were asked for help on a problem, but only half of the participants were accused of cheating. The manipulation of the accusation was intended to influence participants’ stress levels such that participants who were accused of cheating would feel more stress than those who were not accused of cheating.
Thirty-two college students (16 women, 16 men) participated in exchange for course credit. Participants ranged in age from 18 to 24 (M = 19.32). The sample included 29 Caucasians, 1 African-American, and 2 participants who were multi-ethnic. All participants were native English speakers. Two participants were removed from the sample due to suspicion causing the final sample to consist of 30 participants, 15 in each condition.
Participants were randomly assigned to either be accused (n = 15) or not accused (n = 15) of having cheated on an experimental task in a two-cell between-subjects experimental design.
Miranda comprehension was assessed with Grisso’s (1998) Instruments for Assessing Understanding and Appreciation of Miranda Rights, which includes four instruments. The four instruments are the Comprehension of Miranda Rights (CMR), Comprehension of Miranda Rights-Recognition (CMR-R), Comprehension of Miranda Vocabulary (CMV), and Function of Rights in Interrogation (FRI). Each of these instruments is described in more detail next.
Comprehension of Miranda Rights
The CMR assesses participants’ ability to understand the basic meaning of each of the Miranda warning statements. The CMR requires that participants paraphrase each of the four Miranda warnings (e.g., “You have the right to remain silent”). If participants give initial answers that are unsatisfactory, follow-up questions can be asked to give participants the opportunity to display an adequate understanding of the statement. Responses for each of the four statements can be adequate, intermediate, or inadequate, with scores of 2, 1, or 0, respectively. Thus, total scores on the CMR range from 0 to 8.
Comprehension of Miranda Rights-Recognition
The CMR-R assesses participants’ comprehension of their Miranda rights with minimal need to construct a verbal response. The CMR-R requires that participants identify a differently worded statement as being the same or not the same as an original statement. For example, one of the four Miranda statements is read and is followed by three other statements that either have the same meaning or have a different meaning than the first statement. Thus, for each of the four Miranda statements, there are three subsequent statements that the participant has to respond to by saying the subsequent statement either has a similar or a not similar meaning than the first Miranda warning statement. Participants’ responses (i.e., same or different) are then recorded. For each response, participants can get a score of 1 or 0. Thus, for all four statements and the three accompanying statements under each, scores can range from 0 to 12.
Comprehension of Miranda Vocabulary
The CMV assesses participants’ ability to correctly define key terms within Miranda (e.g., entitled, attorney, etc.). Participants are read six words and are asked to correctly define each one. Responses are scored as being adequate, intermediate, or inadequate with scores of 2, 1, or 0, respectively. Thus, total scores on the CMV range from 0 to 12.
Function of Rights in Interrogation
The FRI assesses various beliefs about the consequences of Miranda during the interrogation process and a subsequent trial. The FRI is composed of three subscales. The Nature of Interrogation subscale assesses participants’ understanding of the antagonistic relationship between interrogators and suspects. The Right to Counsel subscale assesses the supportive relationship between an attorney and a client. The Right to Silence subscale assesses participants’ ability to appreciate the fact that, regardless of the authority, individuals have a right to remain silent. Participants are shown four pictures of hypothetical situations and asked a total of 15 questions regarding each picture—each question pertains to only one of the subscales. Participants’ responses are scored as being adequate, intermediate, or inadequate with scores of 2, 1, or 0, respectively. Scores for each subscale range from 0 to 10 and scores on the total FRI instrument range from 0 to 30.
American College Test
Prior investigations have demonstrated that intelligence can have a significant influence on individuals’ comprehension of Miranda (e.g., Clare & Gudjonsson, 1991; O’Connell, Garmoe, & Goldstein, 2005). To account for this potential influence among our participants, we assessed American College Test (ACT) scores, which have been shown to correlate significantly with standard measures of intelligence (e.g., California test, Otis-Lennon, Lorge-Thorndike, Henmon-Nelson; Koenig, Frey, & Detterman, 2008). Participants indicated their ACT scores in response to an open-ended item that asked them to report their scores as accurately as possible.
Participants provided demographic information pertaining to their age, gender, and race.
We assessed the effectiveness of the accusation manipulation with pilot participants (N = 40). Following the same procedures of the main experiment, pilot participants either were (n = 18) or were not (n = 22) accused of cheating. In addition, shortly after the accusation manipulation, pilot participants reported how stressed, worried, anxious, nervous, concerned, tense, and scared they felt on rating scales with anchors 1 (Not at all) to 7 (Extremely). Pilot participants’ responses to these seven questions were averaged to create one variable per participant with higher values indicating greater stress, α = .97.
Participants were run individually. Upon arrival, a male experimenter greeted the participant and asked her/him to be seated at a counter with a confederate posing as another participant in the study. The pair was then given a consent document to read and sign. The experimenter then explained to the pair that the study was designed to examine the influence of individual differences on people’s ability to make decisions alone and with a partner. The participant then reported their demographic information and ACT scores individually on a computer while the confederate appeared to do so as well in a separate room. After providing this information, the pair was brought into a different room and given 3 min to get acquainted with one another. This was done to increase the likelihood that the participant would provide an answer when asked for help by the confederate. After the pair had a chance to become acquainted, the experimenter told them that the next phase of the study involved the pair completing some logic problems. The pair was given three logic problem packets, two individual packets and one team packet obtained from Russano et al. (2005). The pair was told that they needed to work alone on the individual packets, but that they needed to discuss their answers and strategies when completing the team packet. The experimenter then left the room and closed the door.
While the pair was completing the logic problems, the confederate asked the participant for help on the final individual logic problem after which the pair completed the final team logic problem. After the pair had completed all of the logic problems, they informed the experimenter that they were done. The experimenter then gave the pair a filler survey to complete while he ostensibly corrected the logic problem packets. After the pair had completed the survey, the experimental manipulation occurred.
In the accusation condition, the experimenter returned to the pair’s room, looking irritated, and told the pair that he had to check on something. After a minute, the experimenter informed the pair that there might be a problem and that he would need to talk to each of them separately. The experimenter then escorted the confederate out of the room and told the participant he would be back in a moment to talk to her/him. After 5 min had passed, the experimenter returned to the participant and explained that he had talked with the other participant and there was indeed a problem. The experimenter went on to explain that he was suspicious that the two had shared answers on one of the individual logic problems because they both had the same wrong answer. The experimenter said that he was not sure how he should handle this situation, so he called his advisor and he could tell that his advisor was annoyed. The experimenter further explained that he was not sure what his advisor would do or how he would handle the situation and that the advisor may even consider this a case of cheating. The experimenter said that the advisor wanted to talk to the participant about this situation himself, but that he was in a meeting and would come to the lab in about 10 min. In the meantime, the participant was supposed to finish up the study. In the no accusation condition, the experimenter simply told the confederate to go back to their original room and that he would be with her/him shortly. Following the experimental manipulation, all participants were informed that the last part of the study was designed to assess verbal and creative individual differences. Participants were then individually administered Instruments for Assessing Understanding and Appreciation of Miranda Rights (Grisso, 1998). Their responses were audio recorded for scoring purposes.
After the experimenter finished assessing participants’ comprehension of Miranda, participants were probed for suspicion. Participants were then fully debriefed, with extra care being taken to assure accused participants that they were not in any trouble. After all questions regarding the study were answered, participants were thanked and dismissed.
Scoring of Miranda Comprehension
Using standardized scoring procedures (Grisso, 1998), three coders, blind to condition, independently coded participants’ audio recorded responses to each of the Instruments for Assessing Understanding and Appreciation of Miranda Rights except for the CMR-R which is scored in a purely objective manner.
Interrater Agreement for Miranda Comprehension Instruments
Interrater agreement among the three coders with respect to the Miranda comprehension instruments was examined with an intraclass correlation (Suen & Ary, 1989). The results indicated that the three coders had a good level of agreement, ICC = .71 (Cicchetti, 1994). Therefore, the coded responses were averaged to yield one Miranda comprehension score for each of the instruments and subscales used in the main analyses.
A one-way analysis of variance (ANOVA) was used to examine if there were differences in participants’ ACT scores across the two accusation conditions. The independent variable for this analysis was the accusation manipulation. The dependent variable was participants’ ACT scores. Results indicated that participants’ ACT scores did not differ across the accusation conditions (MAccused = 24.50; MNot Accused = 24.87), F (1, 29) = .08, p = .78. Furthermore, an independent samples t-test indicated that there were no significant differences between men’s (M = 24.46) and women’s (M = 24.85) ACT scores, t (28) = .31, p = .76. We also conducted several zero-order correlations between participants’ ACT scores and their scores on the Miranda comprehension instruments. These correlations indicated that participants’ ACT scores evidenced marginally significant correlations of moderate magnitudes with three of the four comprehension instruments, i.e., the CMR (r = .31, p = .09), the CMR-R (r = .27, p = .16), and the CMV (r = .31, p = .10).1 Accordingly, we included participants’ ACT scores as a covariate in all of the main analyses.
To test for gender differences in participants’ Miranda comprehension scores, we performed a multivariate analysis of covariance (MANCOVA). Gender was included as the subject variable and participants’ ACT scores were included as a covariate. The dependent variables were participants’ scores on the four instruments and the three subscales of the FRI used to assess Miranda comprehension. The results indicated that there were no significant differences between men’s and women’s responses, F (7, 21) = 1.05, p = .426. Therefore, we did not include participants’ gender as a variable in any of the subsequent analyses.
An independent samples t-test was used to examine the effectiveness of the accusation manipulation. We examined whether pilot participants’ self-reported levels of stress differed across the accusation and no accusation conditions. The results indicated that pilot participants who were accused of cheating reported significantly higher levels of stress (M = 3.75) than pilot participants who were not accused of cheating (M = 2.48), t (38) = 3.04, p = .004, 95% CI: .43; 2.12, d = .99. This result provides support for the effectiveness of the accusation manipulation.
A MANCOVA, followed by separate analyses of covariance (ANCOVAs), were conducted to test the competing hypotheses regarding the effect of stress on Miranda comprehension. The independent variable for these analyses was the accusation manipulation. Participants’ ACT scores were included as a covariate. For the MANCOVA, the dependent variables were participants’ coded responses to the four individual instruments and the three subscales of the FRI instrument of the Instruments for Assessing Understanding and Appreciation of Miranda Rights (Grisso, 1998). Results of the MANCOVA analysis indicated that participants in the accusation condition had significantly lower Miranda comprehension scores than participants in the no accusation condition, F (7, 21) = 3.27, p = .016.
Means, standard deviations in parentheses, F values, and effect sizes of Instruments for Assessing Understanding and Appreciation of Miranda Rights for both conditions
Accused (n = 15)
Not accused (n = 15)
The Miranda ruling was designed to provide a safeguard for suspects against police intimidation. However, social scientists and legal scholars have long hypothesized that stress, caused by police accusation, may undermine suspects’ ability to fully understand their constitutional rights as described in Miranda warnings. Our research provided empirical support for this important hypothesis. Consistent with a large body of research showing that stress adversely affects cognitive functioning, our results indicated that stress, induced through an accusation of wrong-doing, significantly undermined participants’ comprehension of Miranda. Specifically, participants who were accused of wrong-doing demonstrated significantly lower Miranda comprehension scores than participants who were not accused of wrong-doing.
Magnitude of Effects
By using an experimental design, the current research was able to infer a causal relationship between the accusation manipulation and Miranda comprehension. Moreover, because the pilot study demonstrated that accused participants experienced more stress than participants who were not accused, there is reason to believe that the accusation manipulation influenced participants’ Miranda comprehension scores, at least in part, because it affected their stress levels. However, because this research was conducted as a lab experiment, it suffers from the potential lack of generalizability. Indeed, it would have been neither ethical nor feasible to expose our participants to the same degree of stress that accompanies police accusation. Nonetheless, it is important to understand what this means with respect to our findings. Because the degree of stress experienced by our participants was very likely less than that experienced by suspects accused of serious crimes by police, the observed results are likely conservative. This possibility is particularly noteworthy given that for six of the seven analyses we performed, the accusation manipulation had a large effect on our participants’ ability to comprehend Miranda, with effect sizes of d ≥ .80 (Cohen, 1988). Therefore, it seems likely that the effects that were observed in these data are conservative estimates of effect that stress has on Miranda comprehension during police accusation.
Furthermore, the effect sizes observed in the current research were larger than those observed in previous research that induced stress via a mock crime (Rogers et al., 2010). This is important to note because it provides evidence that an accusation of actual wrong-doing likely elicits more stress and more greatly undermines Miranda comprehension than does an accusation of having committed a mock crime. Moreover, if actual accusations of wrong-doing do elicit a higher degree of stress, as our results suggest, then it would seem reasonable to conclude that the more stress suspects experience, the less able they will be to fully comprehend their Miranda rights.
Another way to appreciate the effect that the accusation had on our participants’ ability to comprehend Miranda is to compare their comprehension scores to the comprehension scores of other samples reported in the literature. Such a comparison reveals that participants who were accused of wrong-doing in our research demonstrated an average comprehension score that was relatively equivalent to the average comprehension scores of juveniles (Grisso, 1998), adults who were diagnosed as psychotic, and adults who were patients at a psychiatric ward (Cooper & Zapf, 2008). These comparisons are particularly noteworthy given that we relied exclusively on college students as participants and college students are typically assumed to have a higher level of intelligence than the general population. Thus, the detrimental effect that the accusation had on our participants’ Miranda comprehension scores was not only large in a statistical sense, but also meaningful in a practical sense.
Of course, this does not necessarily mean that the comprehension scores of these participants were unacceptably low from a legal standpoint. It is entirely possible that the degree to which the accusation undermined these participants’ comprehension scores was not so large as to consider their comprehension of Miranda to be unacceptable during trial. But, as we discussed earlier, the effects that were observed in these data are likely conservative estimates of the effect that stress may have on suspects’ ability to comprehend Miranda during actual police accusations. In the naturalistic setting, then, the stress that suspects will likely experience in response to police accusation could reduce their Miranda comprehension to such a degree that their understanding of Miranda would be considered unacceptable, even from a legal standpoint.
Theoretical Reasons for the Observed Effects
Existing theory suggests that the tendency for stress to have undermined our participants’ comprehension of Miranda may reflect a compromised working memory system. According to the processing efficiency theory (Eysenck, 1982, 1983, 1997), individuals under stress make more errors, require more processing time, and rely on cognitive shortcuts when engaged in problem solving and decision making. Cognitive efficiency is especially likely to be compromised when individuals are engaged in a difficult task (Eysenck, 1997), such as when suspects attempt to comprehend the legal vernacular of Miranda warnings. Thus, the reduced comprehension exhibited by our participants may have occurred because they relied on less effortful and more well-rehearsed cognitive processing. For example, because lawyers are often characterized in our society as being too expensive for most citizens (American Bar Association, 2002), suspects may rely on this heuristic and not realize that Miranda states that suspects have access to free counsel. Indeed, right to free counsel is one of the worst comprehended components of Miranda (Rogers, 2008b).
Implications for the Legal System
The effect that the accusation manipulation had on our participants’ comprehension of Miranda has important implications for the integrity of the criminal justice system as well as for the civil liberties of suspects. Most notably, it suggests a need for standardization in the administration of Miranda warnings. Indeed, until very recently Miranda warnings were thought to be relatively uniform. However, recent investigations of different jurisdictions across the United States have shown that warnings differ in length, reading difficulty, and method of administration (Greenfield, Dougherty, Jackson, Podboy, & Zimmermann, 2001; Rogers, Harrison, Shuman, Sewell, & Hazelwood, 2007). Thus, because warnings are not administered in a standardized fashion, it seems plausible that stress, caused by police accusation, could interact with various characteristics of Miranda warnings. For example, stress may undermine Miranda comprehension more strongly when the warnings are longer versus shorter, presented at a more difficult versus a less difficult reading level, and presented in an oral versus a written format. Empirical evidence showing that stress undermines suspects’ comprehension of the warnings under these conditions would support reforms to standardize Miranda warnings in terms of length, reading difficulty, and method of administration.
Our findings also have implications for Miranda waivers. If it is the case that stress undermines individuals’ ability to comprehend Miranda in naturalistic settings, then it is possible that the effects we observed might also influence Miranda waivers. For example, people have a tendency to engage in informational social influence processes whereby they look to others as a way to determine correct, appropriate, or socially desirable behavior (Cialdini & Griskevicius, 2010). Applying this tendency to a Miranda administration situation raises the possibility that suspects will be especially susceptible to police tactics used to attain Miranda waivers. That is, because stress is compromising suspects’ ability to appreciate the significance of Miranda, suspects may decide whether or not to waive their Miranda rights based on cues given by the police. If police minimize the importance of Miranda warnings (Leo, 1996; Simon, 1991), then suspects may assume that waiving their rights is in their long-term interests even though it is not. This could be a main reason why four out of every five suspects waive their Miranda rights (Leo, 1996; Schulhofer, 1996).
Limitations and Future Directions
There are two limitations of the current study that warrant discussion. First, the current research manipulated stress via an accusation of wrong-doing. Although pilot data demonstrated that the accusation manipulation significantly influenced participants’ stress, this does not rule out the possibility that the accusation manipulation had additional effects on participants’ Miranda comprehension scores for reasons that were unrelated to stress. In other words, it is possible that some characteristic of the accusation that was not associated with stress was partly responsible for the differences that we observed in participants’ Miranda comprehension scores. Although our data cannot rule out this possibility, current theoretical perspectives specifically hypothesize that stress caused by police accusation undermines suspects’ comprehension of Miranda rights (Grisso, 1998; Kassin et al., 2007; Oberlander & Goldstein, 2001), thereby supporting our interpretation that the accusation manipulation at least partly affected participants’ Miranda comprehension scores because of its effect on their stress levels.
A second potential limitation of our research pertains to the perceived relevancy of the Miranda measures. In our study, participants were led to believe that the Miranda comprehension instruments were a measure of individual differences. Thus, it is not likely that participants who were accused of wrong-doing perceived the Miranda comprehension instruments to be relevant to the accusation. This is an important issue because it raises the possibility that participants might have demonstrated better comprehension of Miranda had they perceived the assessment instruments as relevant to their current circumstances. Although the current research cannot directly address this issue, it is worthwhile to point out that police often attempt to reduce suspects’ perceptions of the relevancy of Miranda warnings. For example, legal scholars have observed that police will employ various tactics that make it appear as though the warnings are not particularly relevant to suspects’ long-term outcomes by minimizing and trivializing the importance of the warnings and by administering the warnings in a nonchalant fashion (Leo, 1996; Simon, 1991). Accordingly, the scenario used in this research closely resembles a situation that would be observed outside the lab. Nonetheless, the issue of perceived relevancy needs to be addressed with future research to draw more definitive conclusions about the effect that perceived relevancy has on suspects’ ability to comprehend Miranda.
The landmark decision of the Supreme Court in the case of Mirandav. Arizona (1966) afforded suspects prophylactic safeguards against police intimidation. The current investigation addressed the influence that stress has on suspects’ ability to comprehend and grasp the significance of these warnings. Consistent with the hypothesis that stress worsens suspects’ ability to comprehend Miranda, participants in our research who experienced stress in response to having been accused of wrong-doing, had significantly worse levels of Miranda comprehension than participants who were not under stress as a result of not having been accused of wrong-doing. This finding supports theoretical speculation that, although adults generally have a good level of understanding of Miranda normal settings, the psychological state of individuals upon arrest is likely to be very different and have a disadvantageous influence on their ability to effectively comprehend Miranda (Grisso, 1998; Oberlander & Goldstein, 2001; Rogers et al., 2010). In this way, the current findings raise the possibility that the protections Miranda was designed to afford suspects may not be realized to the full extent initially intended and that research needs to identify ways to improve suspects’ Miranda comprehension to increase the extent to which suspects are afforded protections against police intimidation.
We also conducted the seven individual main analyses without participants’ ACT scores using seven ANOVAs. Results remained the same in terms of trends and significance with two exceptions. The CMR-R and the RC subscale of the FRI instrument that were significant when ACT scores were included as a covariate (ps = .04) became marginally significant when ACT scores were excluded from the analyses, ps = .07.