Reducing academic cheating through growth mindset: an intervention study and a mechanism analysis

ERCT Criteria Breakdown

• Level 1 Criteria

  • C

    Class-level RCT

    • Randomization occurred at the individual participant level within sessions (not class- or school-level), and the tutoring exception does not apply.
    • "Within each session, participants were randomized 1:1 to the growth-mindset intervention or the active control using a computer-generated block randomization list."
    • Relevant Quotes: 1) "We randomly assigned them to the intervention (n = 60) or control group (n = 60)." (p. 3) 2) "Within each session, participants were randomized 1:1 to the growth-mindset intervention or the active control using a computer-generated block randomization list." (p. 3) 3) "All experimental sessions were conducted in the same classroom and scheduled within a consistent evening time window (4:00 p.m. to 9:00 p.m.)." (p. 3) Detailed Analysis: Criterion C requires that randomization be implemented at the class level (or stronger, such as school/site), to reduce contamination between treatment and control participants. The paper states that participants were randomized "within each session" using a 1:1 allocation, and that sessions were run in the same classroom. This indicates individual-level assignment rather than randomizing intact classes or institutions. The ERCT exception allowing student-level randomization applies to one-to-one tutoring/personal teaching interventions. Here the intervention is a self-guided online growth-mindset program, not tutoring. Criterion C is not met because randomization was at the individual level within sessions rather than at the class (or school/site) level, with no tutoring exception applying.
  • E

    Exam-based Assessment

    • Outcomes are cheating behavior measured via an experimental logic-puzzle paradigm and self-report, not standardized exam- based educational achievement assessments.
    • "The logic test, adapted from Malesky et al. (2022), consisted of 10 questions targeting figural reasoning, verbal reasoning, and applied problem-solving."
    • Relevant Quotes: 1) "Cheating behavior was objectively assessed through experimental methods." (p. 3) 2) "The logic test, adapted from Malesky et al. (2022), consisted of 10 questions targeting figural reasoning, verbal reasoning, and applied problem-solving." (p. 5) 3) "As participants scored their own answers without monitoring, the difference between the experimenter’s recorded scores and participants’ self-reported scores was used as an indicator of cheating behavior." (p. 5) Detailed Analysis: Criterion E requires standardized, widely recognized exam-based assessments of educational outcomes (e.g., state/national standardized achievement tests). In Study 1, the main outcome is cheating operationalized via an experimental logic-puzzle task plus a self-grading opportunity. In Study 2, cheating is measured via a self-report questionnaire about cheating frequency. Neither is a standardized academic achievement exam, and the logic puzzles are used to elicit cheating behavior rather than to measure standardized learning outcomes. Criterion E is not met because the paper does not use a standardized exam-based academic achievement assessment.
  • T

    Term Duration

    • The intervention and main outcome were completed within a single short experimental session (minutes), not tracked for at least one academic term after the intervention began.
    • "The entire web-based procedure took approximately 30 min, with the intervention lasting around 25 min."
    • Relevant Quotes: 1) "The entire web-based procedure took approximately 30 min, with the intervention lasting around 25 min." (p. 3) 2) "The researcher sent pre-prepared logic puzzles to the participants, asking them to complete the test within 10 min." (p. 3) 3) "Additionally, the test questions were presented in two different orders across experimental sessions, and all sessions were completed within 2 days to minimize the chance of answer sharing." (p. 4) Detailed Analysis: Criterion T requires that outcomes be measured at least one full academic term (roughly 3–4 months) after the intervention begins (i.e., at least term-long tracking since intervention start). Here, the intervention is a brief online program (~25 minutes) and the cheating outcome is measured immediately afterward in the same session (with the puzzle test lasting 10 minutes). Across cohorts, all sessions were completed within 2 days, confirming the study does not track participants for months. Criterion T is not met because the intervention-to-measurement interval is minutes (and, across cohorts, at most days), far shorter than an academic term.
  • D

    Documented Control Group

    • The control condition is described in detail (content, structure, matching, and group sizes with baseline comparisons), enabling clear treatment-control comparison.
    • "The control group completed a structurally parallel online program consisting of four modules that were matched to the intervention in length, presentation format, and level of engagement."
    • Relevant Quotes: 1) "We randomly assigned them to the intervention (n = 60) or control group (n = 60)." (p. 3) 2) "At baseline, no significant differences were found between students in the control and intervention groups in gender (χ2 = 0.85, p = 0.36), age (t (118) = 0.25, p = 0.80), and grade (χ2 = 0.44, p = 0.93) (see Table 1)." (p. 3) 3) "The control group completed a structurally parallel online program consisting of four modules that were matched to the intervention in length, presentation format, and level of engagement." (p. 5) Detailed Analysis: Criterion D requires a well-documented control group, including who is in the group, baseline comparability information, and a clear description of what the control group received. The paper reports equal-sized intervention and control groups, provides baseline comparisons on gender, age, and grade, and describes the control content as an active control matched to the intervention in length, format, and engagement. Criterion D is met because the control group is clearly documented (composition, baseline comparability, and control activities).

• Level 2 Criteria

  • S

    School-level RCT

    • Participants were recruited from a single university and randomized individually; schools/sites were not randomized.
    • "A recruitment notice was posted at a public university in Shandong Province, eastern China, and 120 university students enrolled in the experiment."
    • Relevant Quotes: 1) "A recruitment notice was posted at a public university in Shandong Province, eastern China, and 120 university students enrolled in the experiment." (p. 3) 2) "Within each session, participants were randomized 1:1 to the growth-mindset intervention or the active control using a computer-generated block randomization list." (p. 3) Detailed Analysis: Criterion S requires randomization at the school/site (or equivalent institutional unit) level. The study recruited students from one university and randomized participants individually within sessions. There is no evidence of multiple institutions being randomized or of institutional assignment to treatment/control. Criterion S is not met because the randomization unit is the individual participant, not the school/site.
  • I

    Independent Conduct

    • The trial sessions were administered by the study team (a single primary researcher), without documented independent conduct by a third-party evaluator.
    • "All sessions were administered by the same primary researcher."
    • Relevant Quotes: 1) "All sessions were administered by the same primary researcher." (p. 3) 2) "Three research assistants supported logistics but remained outside the classroom once the session began." (p. 3) Detailed Analysis: Criterion I requires that the evaluation be conducted independently from intervention designers/implementers (e.g., an external evaluation team), to reduce bias in implementation and assessment. The paper states that all sessions were administered by the same primary researcher, supported by research assistants. It does not describe any external evaluator or third-party organization that independently conducted implementation, data collection, or analysis. Criterion I is not met because independent conduct is not documented; administration appears to be performed by the study team.
  • Y

    Year Duration

    • Outcomes were measured immediately within a single session (and all cohorts completed within 2 days), far short of 75% of an academic year; also, since T is not met, Y is not met.
    • "The entire web-based procedure took approximately 30 min, with the intervention lasting around 25 min."
    • Relevant Quotes: 1) "The entire web-based procedure took approximately 30 min, with the intervention lasting around 25 min." (p. 3) 2) "Additionally, the test questions were presented in two different orders across experimental sessions, and all sessions were completed within 2 days to minimize the chance of answer sharing." (p. 4) Detailed Analysis: Criterion Y requires outcomes to be measured at least 75% of an academic year after the intervention begins. Here, both intervention and outcome measurement occur immediately within a single session, and across cohorts the entire study completes within 2 days. This is far shorter than an academic year. Additionally, per ERCT dependency rules, if criterion T is not met then criterion Y is not met. T is not met because the intervention-to-measurement interval is far shorter than a term. Criterion Y is not met because the study does not track outcomes for anything approaching an academic year (and T is not met).
  • B

    Balanced Control Group

    • The control condition is an active, structurally parallel program matched to the intervention in length, format, and engagement, indicating balanced time and activity inputs across groups.
    • "The control group completed a structurally parallel online program consisting of four modules that were matched to the intervention in length, presentation format, and level of engagement."
    • Relevant Quotes: 1) "The intervention involved a single self-guided online session lasting about 25 min." (p. 4) 2) "The control group completed a structurally parallel online program consisting of four modules that were matched to the intervention in length, presentation format, and level of engagement." (p. 5) 3) "Participants were informed that upon completing the test, they would be ranked based on their performance, with the first place receiving a first prize (RMB 30, approximately USD 4), the second and third places receiving second prizes (RMB 20, approximately USD 3), and no monetary prize for the remaining participants." (p. 4) Detailed Analysis: Criterion B evaluates whether time, resources, and support are balanced between intervention and control, unless extra resources are explicitly the treatment variable. The paper explicitly states the control program is "structurally parallel" and matched to the intervention in "length, presentation format, and level of engagement," which directly addresses time-on-task and engagement balance. Both groups also appear to undergo the same cheating-assessment procedure (including the same prize incentives), suggesting no differential resource allocation during measurement. There is no indication that one group received extra instructional time, materials, or budget beyond what is integral to the online module itself, and the online module dosage is intentionally matched. Criterion B is met because the control condition is designed as a matched active control with comparable time and engagement inputs.

• Level 3 Criteria

  • R

    Reproduced

    • No independent replication of this specific RCT (growth mindset intervention reducing cheating) by other authors was identified, and the paper itself frames the study as providing first causal evidence.
    • "The present study addresses this gap by providing the first causal evidence, via a randomized controlled trial, for the effectiveness of growth mindset interventions in reducing cheating."
    • Relevant Quotes: 1) "However, causal evidence for its impact on academic cheating remains scarce." (p. 2) 2) "The present study addresses this gap by providing the first causal evidence, via a randomized controlled trial, for the effectiveness of growth mindset interventions in reducing cheating." (p. 7) Detailed Analysis: Criterion R requires independent reproduction: a replication conducted by a different research team, in a different context, published in a peer-reviewed outlet, that reproduces the central experimental claim. The paper does not report that its specific RCT has been independently replicated. To re-check this criterion with internet access, targeted searches were conducted for (a) replication attempts citing this DOI/title and (b) other RCTs explicitly replicating this growth-mindset-cheating intervention claim. No such independent replication papers were identified in the search results available at the time of this ERCT check. The authors’ statement that they provide "the first causal evidence" for reducing cheating via a growth mindset intervention is also consistent with the absence of prior independent replications for this specific claim. Criterion R is not met because independent replication evidence for this study/claim was not found.
  • A

    All-subject Exams

    • Because criterion E is not met (no standardized exam-based achievement assessment), criterion A is also not met; the study does not assess all core subjects via standardized exams.
    • Relevant Quotes: 1) "The logic test, adapted from Malesky et al. (2022), consisted of 10 questions targeting figural reasoning, verbal reasoning, and applied problem-solving." (p. 5) 2) "Exam cheating was assessed using the exam cheating subscale from the Academic Dishonesty Questionnaire developed by Chen (2009)." (p. 6) Detailed Analysis: Criterion A requires standardized exam-based assessments across all main school subjects, and ERCT rules specify that if criterion E is not met then criterion A is not met. This paper measures cheating behavior and mindset-related constructs rather than standardized academic achievement in any subject, so it cannot satisfy an all-subject standardized exam requirement. Criterion A is not met because criterion E is not met and no standardized exam outcomes across core subjects are reported.
  • G

    Graduation Tracking

    • The study provides only immediate/short-term measurement and no tracking to graduation; additionally, because Y is not met, G is not met.
    • "Additionally, the test questions were presented in two different orders across experimental sessions, and all sessions were completed within 2 days to minimize the chance of answer sharing."
    • Relevant Quotes: 1) "Additionally, the test questions were presented in two different orders across experimental sessions, and all sessions were completed within 2 days to minimize the chance of answer sharing." (p. 4) 2) "After completion of data collection, participants received a general debriefing regarding the purpose and procedure of the study." (p. 4) Detailed Analysis: Criterion G requires tracking participants until graduation from the relevant educational stage (here, higher education). The paper describes an experiment completed within sessions (and across cohorts within 2 days) and then debriefing after data collection, with no mention of longer-term follow-up or graduation outcomes. To re-check this criterion with internet access, searches were conducted for follow-up publications by the same author team (Chang, Bao, Zhang, Xu, Huang, Xin) reporting graduation tracking for this cohort. No such follow-up papers were identified. Additionally, per ERCT dependency rules, if criterion Y is not met then criterion G is not met. Y is not met because the study does not track outcomes for an academic year. Criterion G is not met because there is no graduation tracking (and Y is not met).
  • P

    Pre-Registered

    • No protocol pre-registration registry/ID (and no registration date showing registration before data collection) is reported.
    • Relevant Quotes: (No relevant quotes found: the paper does not mention pre-registration, a registry name, a registry ID, or a registration date.) Detailed Analysis: Criterion P requires an explicit pre-registration statement, including a registry reference (ideally an ID/link) and evidence that registration occurred before data collection began. The paper includes ethics and data availability statements but contains no mention of pre-registration or trial registration. Searching the article full text for terms such as "prereg", "register", or "OSF" did not identify any pre-registration disclosure. Criterion P is not met because the paper does not report any pre-registered protocol details.