Bargaining for Freedom: A Person-by-Situation Approach to Studying Plea-Bargain Decision-Making
Lauren Clatch, University of Minnesota, Department of Psychology
Discounting scholars have studied decisions about monetary rewards, or gains, and found that the probability and delay of the receipt of those gains influence their subjective value. Plea-bargain decisions inherently contain the features of probability and delay in the decision context: Trial’s outcome is uncertain while the plea offer’s outcome is certain, and trial is delayed while the plea offer is relatively immediate. This dissertation’s four studies apply discounting paradigms to plea-bargain decision-making and find that probability of trial conviction and delay until trial influence decision making. Additionally, Studies 3 and 4 highlight how other situational features, Factual Innocence and Attorney Advice, influence plea decisions. Overall, the person-by-situation approach from personality and social psychology adopted by this dissertation offers discounting scholars a richer understanding of the personal and situational factors that can influence decision-making.
In a plea-bargain decision, a person charged with a crime has the choice whether to accept an offer made by the prosecutor to accept guilt, forego a criminal trial, and go straight to the sentencing phase in which the offered sentence is the presumptive sentence. Most people accept that offer. But is that because they are guilty and the evidence is stacked up against them? Or are there other features of the plea-bargain choice context that influence that decision?
I use discounting as a theoretical starting point to understand plea bargain choices because discounting researchers study binary choices between one option that is uncertain and delayed and a second option that is certain and immediate. In the plea-bargaining setting, trial is uncertain and delayed, and the plea bargain is relatively certain and immediate. I was drawn to the potential utility of discounting paradigms to explain plea-bargain decision making because the general psychological patterns described by discounting research might help explain why so many people accept plea bargains. Specifically, discounting research has shown that, in general, people prefer certain and immediate outcomes (for review see Frederick, Loewenstein, & O’Donoghue, 2002), which if applied to a plea-bargaining setting could show why people prefer the plea offer over trial.
Certain features of monetary-decision contexts are especially important to predict behavior: probability of the outcome, delay/time of occurrence, and commodity valence (gains vs. losses, termed the “sign effect”). Especially of interest to social psychologists, commodity type seems to matter, too, as
revealed by research in the last decade. In the context of losses, Harris (2012) found descriptively different patterns of behavior when the commodity was lost money compared to non-monetary losses. In particular, with monetary commodities, the median participant chose to put off losses as long as possible and chose to experience gains without any delay. A non-monetary gain of receiving sincere praise from someone you respect showed an overwhelming choice for immediate praise, but the non-monetary loss of a bee sting showed bi-modality at the poles with many people choosing to get the losses over with and many others choosing to put off the loss as long as possible.
The traditional way of assessing discounting involves a complex series of questions (called the amount-adjusting titration procedure) to triangulate the subjective value that people assign to an offered outcome (deduced from an indifference point). In particular, experimentalists ask participants a series of binary, forced-choice questions, varying a feature of the choice context within subjects (e.g., delay or probability) to see how it affects decisions. In each
within-subjects condition, the amount of money (or here, the sentence associated with the plea or trial) is varied until an indifference point could be deduced—which is a point where a person is equally likely to choose the two outcomes.
Using discounting’s methodological paradigm, this dissertation’s series of four studies provides an opportunity to examine more deeply a recent finding that discounting occurs in the plea-bargaining context in similar ways as in monetary-decision contexts (Clatch & Borgida, 2021).
The key dependent variable in this research, Trial Aversion, was measured using the amount-adjusting titration procedure. The higher the indifference point, the more trial aversion a participant is exhibiting. Simply put, the higher the sentence a person is willing to accept through a plea deal, the more trial aversion they are evincing.
All four online surveys contained a vignette describing an unfortunate hypothetical situation in which the participant is asked to imagine driving down a residential road while texting using voice activation when they hit and injure a child. Participants then are instructed that they are being (hypothetically) charged with a criminal offense because it is illegal to use your hands to text while driving (i.e., they are actually innocent); then they answer a series of binary, forced-choice questions in which they have to decide whether to accept a plea bargain or to go to trial. Last, participants completed a number
of individual difference measures (e.g., an impulsivity scale), open-ended questions (e.g., why do most people accept pleas?), and demographic questions (e.g., gender).
All four studies had mixed-factor designs with some variables manipulated between-subjects and others manipulated within-subjects. All four studies manipulated Probability of Trial Conviction (e.g., 5%, 50%, 95%) and Delay until Trial (e.g., 1 day, 1 week, 1 month). Studies 1, 2, and 3
manipulated Waiting-for-Trial Location (while in Jail vs. out on Bail). Study 2 included a methodological manipulation to determine whether there are differences between single-item measurements of the key dependent variable and the traditional titration procedures (that use a series of binary questions).
Studies 3 and 4 manipulated participants’ Factual Innocence by randomly assigning them to read a vignette where they texted while using voice activation (Innocent) or texted while using their hands (Guilty). Study 4 manipulated Attorney Advice (Plea Advice vs. Trial Advice) such that participants received advice about whether they should go to trial or accept a guilty plea.
All four studies used online participant pools. Study 2 obtained a nationally-representative sample based on the 2019 U.S. Census data through Lucid Marketplace.1 Studies 3 and 4 obtained nationally-representative samples based on gender and age (using the same 2019 Census data) and oversampled Black Americans2 to provide more statistical power for race sub group comparisons, especially the one most relevant in criminal justice research. People of color, especially black men, have long been overrepresented in the U.S. criminal justice system (Spohn, 2011).
Simple linear mixed-effects regressions and mixed-effects multiple regressions were conducted to test the effects of experimental and individual difference variables on plea decisions. Study 1 demonstrated that participants’ impulsivity influenced their plea bargain decisions by changing their perceptions of probability of trial conviction—that is, participants with lower impulsivity were especially sensitive to the chances of losing at trial and accordingly accepted more pleas. This finding was consistent across all four studies. All three studies that manipulated Probability, Delay, and Pretrial Detention found the three-way interaction first presented by Clatch and Borgida (2021) such that likelihood of losing at trial combined with pretrial detention and a delay until trial induced more plea acceptance. All three studies that measured self-blame and perceptions of innocence demonstrated that participants that blamed themselves for the car accident or injuring the child or perceived themselves to be less innocent for those acts (regardless of their factual innocence) were more likely to accept a plea.
Study 2 demonstrated that the traditional method of measuring Subjective Value in discounting studies (i.e., titration procedures involving a series of binary questions) presents a more reliable measurement of participant preferences than single-item measurement. Studies 3 and 4 demonstrated a nuanced influence of factual innocence/guilt. Although innocents accepted fewer deals on average than those who were guilty, the similarity of the groups’ plea decisions was striking and challenges current assumptions that innocents accept guilty pleas only “at the margins.” For example, nearly one third of innocent participants (29%) accepted the plea bargain no matter what deal they were offered (i.e., even a criminal sentence that was just three days less than their trial sentence), and being guilty only increased that percentage 11 points, with 41% of guilty participants accepting the plea every time. At the other end, 12% of guilty participants exercised their trial right every time a plea bargain was offered, but this rate only increased 4% in innocent participants with 16% of them exercising their trial right every time.
Studies 3 and 4 also, because of their oversample of Black Americans, found that Black participants were more likely to go to trial because they were less sensitive to their probability of losing at trial than White participants.
Study 4 demonstrated that Attorney Advice to accept a plea or go to trial did not (as a main effect) change participants’ plea decisions, but an interaction with Factual Innocence was found. In particular, innocent participants seemed to be less malleable than their counterparts such that guilty participants were very responsive to attorney advice.
This series of studies informs both discounting theory and normative legal theory on plea bargaining. In particular, discounting theory has relied primarily on studies examining monetary decision-making, and the findings in the non-monetary context of plea-bargain decisions wherein a person’s freedom is at stake suggest both commonalities with and differences between the two decision-making processes. Uncertain outcomes are avoided relative to certain outcomes regardless of whether it is money or freedom being decided about. But delayed outcomes are not uniformly avoided relative to immediate outcomes. In the monetary decision context, people prefer immediacy, but when deciding about their freedom people only prefer immediacy when they are detained pretrial, suggesting that in non-monetary decision contexts, situational factors may qualify what seem to be robust findings in monetary contexts.
Traditional legal theories about plea bargaining suggests that probability of trial conviction and the length of the sentence are the only features that are relevant to decision-making (Bibas, 2004; Easterbrook, 2013). However, this series of studies confirms that not only do other situational features like pretrial detention and delay until trial influence decisions, but also person-specific variables like decision-maker race, self-blame, and perceptions of innocence influence plea decisions. Accordingly, this project suggests that the normative models of plea-bargain decision-making are much too narrow and should be expanded to accurately represent the complexity of the decision making process.
This body of work challenges the (hopeful and) intuitive assumption that people’s guilt status is the most important factor in their plea-bargain decision making process. Instead, contextual variables such as how likely it is that they will lose at trial (as interpreted by the prosecutor and defense attorney), how long they would have to wait until trial (determined by judges’ and attorneys’ schedules), and whether they are detained pretrial all influence the decisions regardless of guilty status.
This project’s method of measuring Trial Aversion also not only reveals how likely it is for a person to take a plea but also (because length of a prison sentence in days is a continuous variable) how harsh of a plea they would be willing to accept. So applying the discounting paradigm to this setting and leveraging the conceptual power of the indifference point, it becomes clear that people are willing to trade away their Constitutional right to a criminal trial even when they are innocent for a marginally shorter sentence and some certainty about the immediate outcome. These factors conspire with people’s self-blame and self-perceptions of guilt to induce pleas in one third of the factually innocent participants—a plea deal that was just three days less than their sentence would purportedly be if they lost at trial.
“Plea bargaining is a defining, if not the defining, feature of the present federal criminal justice system” (Brown & Bunnell, 2006). Moreover, William Blackstone coined the well-known maxim: “[I]t is better that ten guilty persons escape, than one innocent suffer” (1826, p. 358). Although the real-world ratio of escaped guilty persons and wrongly convicted innocents is a known unknown, there is evidence that innocents are wrongly convicted through their own decision to accept a plea bargain: 38 of the Innocence Project’s 351 people exonerated by DNA evidence accepted a plea bargain (Innocence Project, 2017).
This dissertation project demonstrates that a variety of psychological processes implicated in the plea-bargaining decision-making process— including probability discounting and self-perceptions of guilt and blame— affect both innocent and guilty defendants. Future structural or procedural changes to the plea-bargaining system may be able to encourage more innocent defendants to go to trial rather than accept a false guilty plea. Pretrial detention appears to be a prime candidate: reducing the number of people in pretrial detention will make the (inherently delayed) trial option more appealing.