Students’ Reasoning About Whether to Report When Others Cheat: Conflict, Confusion, and Consequences

Nearly all students believe academic cheating is wrong, yet few students say they would report witnessed acts of cheating. To explain this apparent tension, the present research examined college students’ reasoning about whether to report plagiarism or other forms of cheating. Study 1 examined students’ conflicts when deciding whether to report cheating. Most students gave reasons against reporting a peer (e.g., social and physical consequences, a lack of responsibility to report) as well as reasons in favor of reporting (e.g., concerns about welfare, justice, and fairness). Study 2 provided experimental confirmation that the contextual factors referenced by Study 1 participants in fact influenced decisions about whether to report cheating. Overall, the findings indicate that students often decide against reporting peers’ acts of cheating, though not due to a lack of concern about integrity. Rather, students may refrain from reporting because of conflicting concerns, lack of information about school policy, and perceived better alternatives to reporting.

Why do students refrain from reporting cheating, even though they say cheating is wrong? One possible explanation is that students simply do not care about academic integrity, and their rejections of cheating are insincere. Alternatively, students may decide against reporting because they prioritize competing concerns (e.g., others' welfare), are unfamiliar with school policies, or perceive better alternatives to reporting. In Study 1, we investigated potential barriers to reporting by interviewing undergraduate students about why they would be willing-or unwilling-to report a peer. Then, to test whether the concerns identified in Study 1 would influence students' decisions, Study 2 manipulated these factors in surveys that prompted students to decide whether to report a peer.

Academic Cheating and Plagiarism
We define academic cheating-referred to in this paper as cheating-as an academic action that violates institutional rules and would yield academic advantages to one or more students if carried out successfully (Barnhardt, 2016;Cizek, 2003;Murdock et al., 2016). Examples of cheating include bringing prohibited crib notes into an exam or giving homework answers to a classmate (Waltzer et al., 2017). Plagiarism, a subtype of cheating, involves presenting others' words or ideas as one's own in an academic assignment (Moss et al., 2018;Park, 2003). Plagiarism can be more ambiguous than other forms of cheating, and students who plagiarize often do not realize they are doing so (Gullifer & Tyson, 2014;Pennycook, 1996;Power, 2009;Waltzer et al., 2017). Thus, reasons against reporting plagiarism (e.g., insufficient evidence, uncertainty about the plagiarizer's intentions) could be distinct from reasons against reporting cheating in general (e.g., social repercussions, consequences to the cheater). Unlike prior work (Harding et al., 2007;Jendrek, 1992;McCabe et al., 2012;Rennie & Crosby, 2002;Yachison et al., 2018), the present research examined reasoning about plagiarism separately from reasoning about other forms of cheating.
Students frequently encounter a variety of cheating in school. Decades of research have documented that most students cheat at least once in college (Davis et al., 1992;Genereux & McLeod, 1995;McCabe et al., 2012;Murdock et al., 2016;Park, 2003;Stone et al., 2009;Whitley, 1998;Yardley et al., 2009). Students also seem to be aware of their peers' cheating. For example, Brown (2002) found that 94% of senior nursing students had witnessed cheating. Moreover, many students are asked by their peers to participate in facilitating dishonesty (Scrimpshire et al., 2017).

How do Students Decide Whether to Report Cheating?
Despite their frequent encounters with cheating, few students-in some studies as few as 1%-say they are willing to report cheating to a teacher (Baird, 1980;Burton & Near, 1995;Jendrek, 1992;Lim & See, 2001;Nuss, 1984;Stone et al., 2009). This is true even as many students say they should report cheating (Jendrek, 1992;Rennie & Crosby, 2002;Yachison et al., 2018). In an experimental setting, Yachison et al. (2018) found that when students were not asked any questions about a peer's act of cheating on an exam, only 9% of students proactively reported cheating, but when asked direct questions about the cheating incident, 63% of students reported the cheating behavior to the experimenter. These choices depended on contextual factors: Students were less likely to report if they were collaborating with the cheater and more likely to report if they were being graded relative to the cheater (Yachison et al., 2018; see also Scrimpshire et al., 2017).
Students' unwillingness to report cheating may seem surprising given evidence that people often protest, report, or otherwise intervene on perceived wrongdoing, even when the interveners are not directly affected by the violations (Darley & Latané, 1968;Dozier & Miceli, 1985;Fehr & Fischbacher, 2003, 2004Henrich et al., 2006;Jackson et al., 2014;Latané & Darley, 1970;Latané & Nida, 1981;Near & Miceli, 1985;Philpot et al., 2020). Why do so few students report acts of cheating to their schools, sometimes in violation of institutional encouragement and honor codes? Here, we consider four possible factors: (1) competing considerations, (2) informational assumptions about school policies¸ (3) perception of better alternatives, and (4) lack of care about academic integrity.

(1) Competing Considerations
One explanation for why students refrain from reporting is that they are faced with a dilemma: On one hand, students disapprove of cheating (Davis et al., 1992); on the other hand, they perceive reasons against reporting a peer (Rennie & Crosby, 2002). In responding to such conflicts, people need to balance competing concerns, ultimately prioritizing one over the other (Keeney & Raiffa, 1976;Kohlberg, 1971;Nucci, 2004;Turiel, 2015;Turiel & Dahl, 2019;Waytz et al., 2013). In deciding whether to report cheating, students may balance concerns about reputational backlash or harmful consequences to the reported cheater against their concerns about the wrongness of cheating (Nitsch et al., 2005;Pupovac et al., 2019). Hence, students may sometimes refrain from reporting cheating because other concerns seem more important, not because they disregard the value of academic integrity. Existing research on students' reasoning about reporting is limited, and most previous studies have either tested explanatory factors pre-determined by the researchers (Jenkel & Haen, 2012;Pupovac et al., 2019) or obtained qualitative responses from small focus groups (Rennie & Crosby, 2002). The present research used open-ended interview prompts and manipulated scenarios in surveys to examine the considerations influencing students' decisions about whether to report.

(2) Informational Assumptions
A second possible contributor is students' beliefs about policies and facts regarding academic integrity. Informational assumptions about the world guide many social evaluations and decisions (Ajzen et al., 2011;Neff & Helwig, 2002;Turiel et al., 1991Turiel et al., , 1987Wainryb, 1991). Prior research suggests that students in schools with an honor code understand their academic policy better (Schwartz et al., 2013), and students in honor code schools are more willing than students in non-honor code schools to report cheating (Arnold et al., 2007;McCabe et al., 2001). Many of the reasons that students hold against reporting (e.g., unclear guidelines, the futility of reporting, and requirement of proof; Pupovac et al., 2019;Rennie & Crosby, 2002) rest on students' beliefs about the process and consequences of reporting cheating. If students lack information about the process of reporting, or if they believe that reporting would ruin the cheater's life (Nitsch et al., 2005), they may refrain from reporting despite a general concern with integrity. The present research examined students' statements about rules and consequences of reporting to assess the relation between informational assumptions and decisions to report.

(3) Better Alternatives
A third explanation is that students prefer alternative ways of responding to cheating. For instance, a student may have accurate knowledge about academic integrity policies and want to stop the cheating yet choose to confront the cheater instead of reporting to the instructor. In this way, students may resolve the dilemma of whether to report by finding an alternative solution besides reporting or doing nothing. If so, studies focused on reporting as the sole response to cheating may overlook high rates of other interventions. The present research analyzed participant statements for suggestions of alternative actions.

(4) Lack of Care
A fourth possible explanation for low rates of reporting is that students refrain from reporting because they find cheating acceptable and do not care that other students cheat. Indeed, some scholars have argued that students think cheating is "not a big deal" since "everyone cheats," and that moral judgments about right and wrong play a minimal role in decisions about cheating (Houston, 1976, p. 301; see also Brown, 2002;McCabe, 1997;McCabe et al., 2012;Sykes & Matza, 1957). As one prominent researcher put it, "Morality does not seem to be a major influence on student decisions to cheat or not to cheat" (McCabe, 1997, p. 444). If students view cheating as normal or acceptable, they should express little or no reason to report cheating and experience little conflict. In contrast, recent evidence suggests that most students judge cheating as wrong in most situations (Baird, 1980;Davis et al., 1992;Jensen et al., 2002;Semerci, 2006;Waltzer et al., 2019). Consistent with their judgments, most students avoid cheating in most cases (Levine et al., 2010;Rundle et al., 2019;Waltzer et al., 2019). Their judgments that cheating is wrong are rooted in concerns beyond just avoiding sanctions, such as fairness, academic standards, honesty, and others' welfare (Miller et al., 2011;Waltzer & Dahl, in press). Insofar as students judge cheating as wrong, we would expect them to experience conflict around their decisions not to report. The present research measured students' evaluations and reasoning about reporting-including whether they do not care about cheating-and assessed whether the provided reasons would influence others' decisions to report.

The Present Research
To address the apparent tension between students' disapproval of cheating and their resistance to reporting it, the present research investigated college students' reasoning about reporting a hypothetical peer's act of misconduct. To extend prior work, we used betweensubjects (Study 1) and within-subjects (Study 2) comparisons to examine responses to plagiarism separately from other forms of cheating.
In Study 1, we conducted structured interviews with undergraduate students about their decisions to report a hypothetical peer's act of plagiarism or cheating. Students first made a decision about whether they would report, and then explained why. In doing so, we were able to measure a large number of students' considerations without constraining their response options, allowing us to identify barriers that students naturally consider in deciding to report. We hypothesized that many students would decide against reporting, but these decisions would rarely be due to a lack of concern about cheating. Instead, we expected students' reasoning about these decisions would be marked by difficulties, as evidenced by conflicting statements, lack of information about school policy, and spontaneous mentions of alternatives to reporting.
Study 2 provided an experimental test of the reasons that participants expressed in Study 1 by varying these reasons across hypothetical vignettes in an online survey. This within-subject design offered increased statistical power and an experimental test of the situational features that appeared to guide students' decisions in Study 1. We expected decisions about reporting to vary across scenarios, demonstrating a link between students' endorsed reasons and hypothetical decisions. Specifically, we expected that factors students raised against reporting in Study 1 would reduce decisions to report in Study 2, and that factors raised in favor of reporting in Study 1 would increase decisions to report in Study 2.
In both studies, participants were undergraduate students attending a large public university in the Western United States. Students were recruited through a university subject pool and given course credit for participation, in accordance with the university's Institutional Review Board (IRB) policy (UCSC IRB #HS2549). The research procedures for both studies were approved by the university's IRB.
The university in which this research was conducted is a non-honor code secular research university. Although the university does not have an honor code policy, the academic policy does urge students to report witnessed acts of cheating. It states that: "Students need to avoid academic misconduct themselves and are enjoined to report any cases of academic misconduct that are known to them. Students making such reports may maintain anonymity." (UCSC Division of Undergraduate Education, 2019).

Method
Further methodological details are included in the Supplementary Online Materials (SOM). The SOM, anonymized data, analysis scripts, and high-resolution versions of the visuals are all publicly available on the Open Science Framework (https:// osf. io/ fngex/).

Participants
Participants were 141 undergraduate students (93 women, 43 men, 3 non-binary, and 2 people not reporting a gender; M age = 19.76 years; SD age = 1.76). Most participants (80%) were born in the United States, and nearly half (46%) were first-generation college students.

Procedure
Students participated in an in-lab, audio-recorded structured interview. To reduce potential effects of social desirability, the interviewer explained that the study was for psychological research and that it was confidential and completely separate from university administration. Participants were assured that there were no right or wrong answers, and that the researchers just wanted to understand their opinions. Using similar methods, participants in previous studies have been comfortable and fully willing to share their thoughts on the topic of academic cheating with us (Waltzer et al., 2018;Waltzer & Dahl, in press). Audio recordings were masked, transcribed, and then deleted; participants' responses were stored without any identifiable information, in accordance with university IRB policy.
In a between-subjects design, participants were asked about whether they would report a hypothetical peer's act of either plagiarism (plagiarism condition, n = 71) or other cheating (other cheating condition, n = 70). The interviewer asked the following question: "If you knew that someone was engaging in [plagiarism/cheating], would you report it to anyone?" Participants were then asked to justify their decision ("Why?").

Data Coding and Analysis
Participants' responses, which were transcribed from audio recordings, were separated into individual statements for coding. Two coders independently classified statements based on coding schemes for decisions, references to school policy, alternative actions, and type of concern (see Table 1). These categorization schemes were developed through a mix of bottom-up and top-down approaches. Members of the research team reviewed a subset of the data to deductively generate categories that captured common types of responses (bottomup approach). Meanwhile, theoretically-relevant categories (e.g., lack of care) were also added (top-down approach). Because statements about not caring were too rare to be analyzed on their own (only mentioned by 5 participants), they were included in the evaluation of act category (type of concern, Table 1). Reliability was assessed by computing Cohen's kappa scores (κ) for both coders' categorization of a random subset of the data (20% of all responses) (McHugh, 2012).
Applying the categorization schemes to all open-ended responses allowed for the data to be treated as fully quantitative: Each statement was assigned a number for each category (1 = present, 0 = absent). Data were then analyzed using generalized linear models (McCullagh & Nelder, 1989). Models included fixed effects of condition (plagiarism vs. other cheating) and decision to report. Models used logistic link functions and binomial error distributions, and hypotheses were tested using likelihood ratio tests.

Results
Below, we report key findings on students' decisions about whether to report, difficulties in their decision-making, and the concerns students raised. Additional analyses are reported in the SOM (Supplementary Online Materials, https:// osf. io/ fngex/).

Overall Decisions About Whether to Report
Overall, 48% of participants said they would not report, 40% said their decision would depend on some factor, and 11% said they would report. Participants' decisions about whether to report did not differ by condition (plagiarism vs. other cheating), Ds(1) < 0.35, ps > .559.

Conflict and Informational Assumptions in Decisions About Whether to Report
Forty-nine percent of participants made statements both for and against reporting. Many participants expressed ignorance/rejection of school policies (e.g., "I have no idea how I would go about reporting it," 43%) or made ambiguous statements (e.g., "The teacher's probably going to catch them," 62%) that hinted at confusion about or misinterpretation of school policies. Only 28% referenced actual policy (e.g., "Because it [plagiarism] is against the rules"). Without being prompted, 30% of participants spontaneously offered an alternative course of action (e.g., "I would still tell them they probably shouldn't do that and try to guilt them into stopping"). None of the above response categories varied by condition, Ds(1) < 2.03, ps > .155. Students' references to school policy (ignorance/rejection, reference to actual policy, or ambiguous) varied as a function of their decision about whether to report, Ds(2) > 9.81, ps < .008. Post-hoc Wald tests revealed that participants who decided to report made more references to actual policy and had lower rates of ignorance/rejection and ambiguous statements, compared to those who said it depends or decided to not report (Table 2), ps < .007.

Reasoning About Whether to Report
Most participants (65%) raised multiple types of concerns when reasoning about whether to report (M overall = 2.35 concerns). Participants who said it depends provided more reasons than those who would report or would not report, ps < .001 (Table 3). Figure 1 summarizes reasons for and against reporting. The most common reasons against reporting were concerns about responsibility (e.g., "it's not my job to point people out," 35%), consequences (unspecified consequences: 27%; grades and punishment: 24%; social and physical consequences: 21%), and act evaluations (e.g., "I feel like cheating isn't that bad," 21%). The most common reasons for reporting were act evaluations (28%) and welfare, justice, and fairness concerns (e.g., "I feel like it's wrong and dishonest and unfair to a lot of people," 16%).

Discussion
Study 1 indicated that students were highly conflicted about whether to report plagiarism and other forms of cheating. Most participants gave reasons both for and against reporting, and almost a third of participants spontaneously offered alternative courses of action besides reporting. Moreover, many students indicated beliefs that did not align with their school's policies about the process or consequences of reporting. The fact that students who opted against reporting expressed more confusion or ignorance compared to those who would report implies that inaccurate or incomplete information can guide students away from reporting, highlighting possible areas for educational intervention. There were no significant differences between participants who were interviewed about plagiarism and those who were interviewed about other forms of cheating (see General Discussion). To further test for potential differences, Study 2 compared decisions about plagiarism and other forms of cheating across multiple scenarios in a within-subjects design.
The concerns expressed by participants (e.g., about responsibility, severity of the act, and consequences) suggested several factors that may guide students' decisions about whether to report. Still, the design of Study 1 prevented us from concluding that these factors did in fact guide students' decisions. To examine the link between reasoning and prospective actions more closely, Study 2 prompted students to make decisions about reporting across situations that differed in the presence of the concerns expressed in Study 1. For instance, some participants reasoned they would be more inclined to report if the act of cheating impacted their grades (e.g., "I'd be pissed off if I get a worse grade, below average, because someone else is cheating"); in Study 2, this concern was manipulated in the scenario, "Imagine you know that someone is cheating and in doing so they negatively affect the grades of you and your classmates" (for a similar approach, see Dahl et al., 2018). Using this design, Study 2 allowed for an experimental test of whether the factors referenced by participants in Study 1 mattered for judgments about reporting cheating. Study 2 adopted a within-subjects design to increase statistical power.
We expected open-ended responses in Study 2 to follow similar patterns as those in Study 1, reflecting conflicting concerns about whether to report. To further measure conflict, participants in Study 2 were asked whether they should report in addition to whether they would report. We hypothesized most students would say that they should report but also that they would not do so. We expected cross-situational variability in decisions to report. Specifically, we hypothesized that situations based on reasons in favor of reporting would yield more decisions to report, and vice versa for situations based on reasons against reporting.

Type of concern
Reason for reporting Reason against reporting

Participants
We recruited 103 undergraduate students (76 women, 24 men, 1 other response, and 2 people not reporting a gender; M age = 20.21; SD age = 1.52) to participate in an online survey.

Materials and Procedure
Participants responded to a 30-minute online survey administered through Qualtrics. The survey was designed to validate and build on the findings from Study 1. In the first section of the survey, to validate findings from Study 1, participants were asked the same general question as in Study 1 (i.e., "If you knew that someone was engaging in [plagiarism/cheating], would you report it to anyone?"), selecting either would report or would not report. They also rated their likelihood of reporting (sliding scale from 0 = "Extremely unlikely" to 10 = "Extremely likely"), chose whether they should or should not report, and then provided an open-ended explanation for their responses. All participants answered these questions about both plagiarism and cheating (within-subjects design). This survey also built on Study 1 by using the findings from participants' open-ended responses to develop scenarios that were relevant to students' concerns. In the second section of the survey, participants responded to several modified situations that were created using the features that participants referenced in Study 1 (e.g., "Imagine you know that someone is cheating, and in doing so they receive a better grade than those who worked honestly," see Appendix). All scenarios were presented to each participant in randomized order. For each situation, all participants provided three responses: they made a dichotomous choice (report or do not report), rated their likelihood of reporting (sliding scale from 0 = "Extremely unlikely" to 10 = "Extremely likely"), and rated how good or bad it would be to report (from 0 = "Really bad" to 10 = "Really good").

Data Analysis
Analyses of condition effects (plagiarism vs. other cheating) on dichotomous dependent variables were conducted using McNemar's method and exact binomial tests for testing hypotheses. For all other analyses, data were analyzed using generalized linear mixed models with random intercepts for participants and fixed effects of decision about whether to report (Hox, 2010). Models for dichotomous dependent variables used logistic link functions and binomial error distributions. For analyses using continuous dependent variables, models used identity link functions, normal error distributions, and full maximum likelihood estimation. Hypotheses were tested using likelihood ratio tests.

Results
Key findings related to the central research questions are reported here. Additional analyses are reported in the SOM.

General Questions About Reporting
Overall Decisions About Whether to Report In most of the general cases (55%), students said that they should report but would not report, evidencing conflict. Students said they should report and would report in 25% of cases. They said they should not and would not report in 18% of cases. Their general decisions to report did not differ between plagiarism and other cheating, ps > .249. The mean likelihood of reporting across these cases (M likelihood = 4.06, SD = 2.48) was close to the midpoint (5.00), suggesting that participants tended to experience some conflict about whether to report.

Conflict, Informational Assumptions, and Reasoning About Whether to Report As in
Study 1, participants made statements both for and against reporting in several cases (28%). In line with the idea that students are concerned about cheating, this conflict was more likely among participants who said they would not report (35%) than those who said they would (7%), D(1) = 23.13, p < .001. Unlike in Study 1, few participants spontaneously mentioned alternative actions (3% of cases) 1 . Similar to Study 1, participants often expressed ignorance/ rejection of school policy (33%) and made ambiguous statements about policies (37%). Participants were more likely to express ignorance or rejection of policies or make ambiguous statements when they decided not to report, Ds(1) > 22.71, ps < .001 (Table 4). There were no significant effects of condition (plagiarism vs. other cheating) on participants' mentions of both courses of action, alternative actions, or school policies, ps > .115.
On average, participants reasoned about more concerns when they said they would not report (1.70) than when they said they would (1.22), D(1) = 8.81, p = .003. As in Study 1, participants often referenced their lack of responsibility (32% of cases) and various consequences (grades and punishment: 10%; social and physical: 16%; unspecified: 7%) as reasons against reporting, and typically referenced welfare, justice, and fairness (28%), act evaluations (13%), and rules (13%) as reasons for reporting. Would report (n = 55) 2% 11% 40% 1 In a follow-up online survey with a separate sample of undergraduates (N = 40), we directly prompted for alternative actions. As expected, most participants (73%) responded yes to the question "Would you consider taking any alternative actions instead of reporting a student for cheating or plagiarizing?" In their open-ended explanations, most (90%) said they would talk to the person (e.g., seeking to understand why they cheated, warning them of consequences, or offering assistance to avoid cheating).

Questions About Reporting Across Situations
Variability Across Situations Participants' decisions varied substantially across scenarios.
Only 11% of participants said they would not report in any case, and only 1% said they would report in every case. At the same time, the average likelihood of reporting varied substantially between participants. Illustrating both within-and between-participant variability, Fig. 2 visualizes case-by-base responses for each participant.
When comparing scenarios that had both a plagiarizing and a cheating version (see Appendix), the type of act (plagiarism vs. other cheating) did not significantly predict decisions, D(1) = 0.08, p = .783.

Effect of Manipulated Features on Decisions to Report As expected, scenarios that
were based on the features mentioned in Study 1 as motivations in favor of reporting did yield more decisions to report (48%, D[1] = 61.12, p < .001) and higher likelihood of reporting (M likelihood = 5.12, D[1] = 57.46, p < .001) compared to the general case (28%, M likelihood = 4.06). Reflecting the importance of contextual features, some cases yielded broad agreement (e.g., 79% of students would report when their own work was being plagiarized). The scenarios in which students were most willing to report were those in which the reporter had some personal stake in the situation (e.g., when the act of cheating would impact them directly). Additionally, as expected, scenarios based on motivations against reporting yielded less decisions to report (only 22% would report, D[1] = 4.23, p = .040) and lower likelihood of reporting (M likelihood = 3.17, D[1] = 34.32, p < .001) compared to the general case. Again, some scenarios yielded broad agreement (e.g., only 11% of students would report cheating when they had to fill out paperwork and attend academic hearings). The scenarios in which students expressed the least willingness to report were those that involved extra work for the reporter, uncertainty surrounding the circumstances of cheating, or minor offenses (e.g., "plagiarizing just a few words or sentences").

Relation Between Evaluation of Reporting and Decision to Report
There was a strong, positive correlation between participants' ratings of how good or bad it was to report and their rating of how likely they were to report in that situation, Pearson's r = .72, D(1) = 2734.81, p < .001. As suggested by Fig. 3, a substantial proportion of those who gave positive evaluations of reporting (above 5.00) nevertheless chose not to report (36%, vs. 88% of those who gave negative evaluations of reporting). However, this does not mean that these participants' evaluations were unrelated to their intended decisions. Even within the cases for which participants gave positive evaluations of reporting, evaluations and intentions to report were positively correlated, r = .46, D(1) = 619.22, p < .001: Participants who gave less positive evaluations of reporting were also less likely to say they would report, compared to participants who gave more positive evaluations of reporting.

Discussion
As in Study 1, most participants in Study 2 generally decided against reporting, but nevertheless expressed conflict. Most believed they should report, and many made statements that supported both decisions. Decisions against reporting were rarely due to a lack of care: Only one participant said they did not care enough about cheating to report it.
Study 2 also showed that students' judgments about reporting of cheating are sensitive to the specific context of cheating acts. As hypothesized, the reasons students raised in favor of reporting (in Study 1) increased decisions to report (in Study 2), whereas Fig. 3 Bubble plot depicting relation between evaluative rating and likelihood of reporting (Study 2). Larger bubbles indicate more responses. Students were more likely to report acts when they rated reporting in those situations more positively reasons against reporting decreased rates of intended reporting. Further evidencing the link between reasoning and decisions, students' evaluations of reporting and decisions to report were strongly correlated.

General Discussion
Though most students care about academic integrity, they often decide against reporting their peers' acts of cheating. The present research suggested that students do care about integrity, contrasting with prior accounts that imply otherwise (Ariely, 2012;Bandura, 2016;McCabe, 1997;Sykes & Matza, 1957; for a related argument, see Stuhmcke et al., 2016). The findings provided evidence for three factors that help explain why students refrain from reporting, despite their concerns for integrity: competing considerations, informational assumptions, and better alternatives. We discuss each of these points below.

Students Balance Competing Considerations
Both studies indicated that decisions about reporting are complex and involve competing considerations: Students believed the cheating was wrong, but simultaneously considered personal costs (e.g., reputational consequences) or moral concerns (e.g., ruining someone's academic career). Most participants raised multiple concerns when reasoning about whether to report (e.g., "I know that person is doing something wrong, but I don't want to be the reason someone gets in trouble"). This research further illustrates how many moral decisions involve weighing competing concerns (Nucci, 2004;Turiel & Dahl, 2019). When these concerns call for diverging courses of action, a compromise must be made. Reasoning about whether to report cheating, like reasoning about other acts of intervention against wrongdoing, involves conflict (e.g., Latané & Darley, 1970;Waytz et al., 2013).

Students Make Varying Informational Assumptions
Another factor in students' decisions was informational: Many students expressed ignorance about school policies, reporting procedures, consequences of reporting, the cheating student's actions, or the student's motivations for cheating. Consistent with prior research, students' beliefs and knowledge about institutional policies were related to their decisions (Ajzen et al., 2011;Jordan, 2001). Indeed, in real academic situations, students often lack complete information about what constitutes a violation or what the process of reporting entails. For example, many students have not read their school's academic misconduct policy (Gullifer & Tyson, 2014).

Students Perceive Better Alternatives to Reporting
Study 1 showed that students often prefer alternative courses of action over reporting a peer's cheating to the teacher. Without being prompted, students spontaneously proposed alternative courses of action besides reporting or doing nothing (e.g., privately telling the student they are doing something wrong, helping the student complete the work without cheating). These alternative actions further evidence that students are often concerned with stopping cheating even when they decide not to report it. Fewer participants mentioned alternative actions in Study 2. A likely explanation is that Study 2 used online surveys, which tend to elicit shorter responses to open-ended questions than in-person interviews .

Students Do Care About Integrity
The present research provided abundant evidence that students are concerned with academic integrity. In both studies, many participants provided reasons in favor of reporting. In Study 2, there were several scenarios in which the majority chose to report. Most participants in Study 2 also believed they generally should report cheating (80%). Even in the cases that students decided not to report, they expressed conflict and acknowledged reasons in favor of reporting (e.g., "Reporting a student who is cheating is probably the honest thing to do, however it's not my responsibility to worry about what other students are doing"). When students gave reasons against reporting, statements about not caring were exceedingly rare. Study 2 confirmed the influential role students' reasons had in guiding their decisions (Ajzen, 1985(Ajzen, , 1988Ajzen & Fishbein, 2005;Dahl et al., 2018;Knafo et al., 2011;Schwartz, 2012;Turiel, 2003).

Future Directions for Research
This research revealed minimal differences between decisions about plagiarism and other forms of cheating. Although prior work has highlighted that it may be uniquely difficult for students to determine what constitutes plagiarism (Gullifer & Tyson, 2014;Power, 2009;Stuhmcke et al., 2016), the present study removed this difficulty by describing the events presented to participants explicitly as plagiarism. In real academic situations, distinctions between plagiarism and other forms of cheating are characterized by details of the particular acts; when these nuances are stripped away, students may not distinguish plagiarism from other forms of cheating. It is likely that students' responses would have been different if the study had introduced uncertainty about whether the perceived event constituted plagiarism (Waltzer & Dahl, in press). We expect, for instance, that students would be less likely to report an ambiguous case of plagiarism than a clear-cut case of a student hiring someone else to take a test. To address this concern, future research could use hypothetical vignettes depicting concrete examples of student actions (e.g., text from a student's submission) to better capture the uncertainties of identifying plagiarism in real academic situations (Childers & Bruton, 2016;Roig, 1997Roig, , 2001Waltzer & Dahl, in press).
A second area for future research concerns the role of institutional factors (e.g., honor codes) and individual differences (e.g., personality traits). The present work provides a foundation for such future directions by identifying situational features and reasons for and against reporting that could interact with these factors. Building on prior research predicting institutional and individual differences in cheating as well as whistle-blowing (Kisamore et al., 2007;McCabe et al., 2012;Miller et al., 2007;Scrimpshire et al., 2017;Stone et al., 2009;Stone et al., 2012), future research could seek to predict whether some individuals are more likely to report specific acts of cheating than others.
Relations between judgments and actions have received a great deal of theoretical and empirical attention (Blasi, 1980;Thoma, 1994;Stephens, 2018;Turiel, 2003;Waltzer et al., 2019). Previous research has found that people's judgments and values largely guide their behaviors (Ajzen, 1985(Ajzen, , 1988(Ajzen, , 1991Ajzen & Fishbein, 2005;Knafo et al., 2011;Schwartz, 2012). However, in Study 2, most students (55%) seemingly contradicted themselves by saying they should report but would not report. Students' open-ended justifications hint at a distinction between what is objectively right and what is personally required: Many said they felt, from an objective standpoint, that the act "ethically should be reported," but at the same time they personally felt it was "not [their] place" to report and feared the "stigma around reporting." Nevertheless, there was a strong positive correlation between evaluations and decisions about reporting (r = .72). Future research can more closely examine the nature of students' judgments about what they "should" do to determine whether they see reporting as going beyond what is morally required (i.e., supererogatory; McNamara, 2011).
One limitation is that the present research did not directly assess students' personal experiences of observing cheating incidents and deciding whether to report in real life. In addition to students' predicted actions, it would be valuable to examine their actions in real situations (e.g., Ajzen, 1991;Mesmer-Magnus & Viswesvaran, 2005;Woodzicka & LaFrance, 2001). Predicted and actual decisions may sometimes differ because decisions about reporting occur in complicated situations with informational ambiguity, multiple competing considerations, and other intervening factors (e.g., perceived behavioral control; Ajzen, 1991). The relatively simple scenarios tested in the present research identified specific factors that mattered in students' decisions to report (e.g., personal involvement in the situation, severity of cheating act), which can be combined to capture the complexity of decisions in realistic situations. Building on methods that test decisions to report in experimental settings (Jenkel & Haen, 2012;Yachison et al., 2018), future extensions of this research can investigate whether students' predictions align with their decisions in academic contexts.

Conclusion
Decisions about whether to report witnessed acts of cheating occur in complex situations, with incomplete information and conflicting considerations. Both studies indicated that decisions about whether to report are sensitive to features of the situation in which the cheating and reporting occurs. The findings have implications not only for scientific theories of socio-moral decision-making, but also for educators who seek to support academic integrity among students. Efforts to increase students' willingness to report would need to address the specific reasons and beliefs that discourage students from reporting. For example, to address uncertainty or ignorance of policies, educators could offer students clear information about the process and consequences of reporting (e.g., explicitly stating whether students are responsible for reporting). In addition, to change student decisions about reporting, educators could examine the concerns students have against reporting and adjust the assignment or classroom structure accordingly (e.g., increasing ease or anonymity of reporting).

Appendix: Summary of Types of Situations Mentioned in Study 1 and Manipulated in Study 2
To create stimuli for Study 2, participant statements in Study 1 were categorized based on the situational factors that they stated would influence their decision to report (κ = 1.00). Common factors included whether anyone was affected by the act, whether the cheater would receive an unfair grade, and whether it is the witness's business to get involved. See Table 5 for all percentages.