Knowledge of misconduct amid North-African post-graduate dental students: A cross sectional study

Study design

The study was conducted following the guidelines established by the STROBE statement.³⁶ This cross-sectional study was conducted at the Faculty of Dental Medicine of Monastir (FDMM), Tunisia, from May 2, 2022, to October 30, 2022. The research protocol was determined to be exempt from human subjects’ research review by the FDMM Ethics Committee, and therefore, formal approval was not required. For the “paper method” (Figure 1), informed written consent was obtained from each participating student. For the “electronic method” (Figure 1), the fact that the participant responded to our email and sent us the filled questionnaire as an attached file indicates his/her consent to participate in the study. The survey was administered anonymously, ensuring confidentiality. Detailed information regarding the study objectives was provided to the students, ensuring their understanding of the research purposes.

Figure 1. Study flowchart.

TAOM: Tunisian Association of Oral Medicine. PGDSs: Post-Graduated Dental Students.

Study population

The study population consisted of all PGDSs enrolled at the FDMM during the 2022 academic year. The FDMM is the sole institution in Tunisia that offers dental medicine education, admitting approximately 30 to 40 new PGDSs each year. PGDSs are residents who have completed their 5-years of medical studies, and passed the Tunisian national exam to obtain a specialist diploma. For the specific year under investigation (i.e.; 2022), the total number of PGDSs was 147, distributed as follows: 38 in the 1^st year, 33 in the 2^nd year, 38 in the 3^rd year, and 38 in the 4^th year.

All 147 PGDSs were invited to participate in the survey, and a five-step recruitment process, outlined in Figure 1, was implemented. To summarize, the following three recruitment methods were utilized: i) Individual emails were sent to all 147 PGDSs; ii) Announcements were disseminated through the Facebook pages of the 14 medical dental departments of the FMDM; and iii) Convenience sampling was employed, where questionnaires were distributed to PGDSs who had attended the medical congress of the Tunisian Association of Oral Medicine (n=32) or were acquainted with authors involved in the study (i.e.; 15 colleagues chosen from the 147 PGDSs ‘list and who had not taken the survey before). During steps 1 to 3, which involved the electronic enrolment method, 59 PGDSs were recruited. In steps 4 and 5, which involved the paper-based enrolment method, an additional 47 PGDSs were recruited. During the paper-based enrolment method (i.e.; steps 4 and 5), two researchers (RK and MK in the authors’ list) asked the PGDSs if they had taken the survey before. If they had not, the researchers administered the questionnaire and collected the surveys from the PGDSs upon completion.

Sample size calculation

The sample size was estimated using the following formula³⁷: n = [(Z _α/2) ² × P × (1 - P) × D]/E ²; where « n » is the required number of students, « P » is the proportion of the main event of interest (i.e.; KoM low-level), « E » is the margin of error, « Z _α/2 » is the normal deviate for a two-tailed alternative hypothesis at the desired level of significance, and « D » represents the design, which is equal to 1 for simple random sampling.

According to a study conducted in Saudi Arabia,¹⁹ it was found that 17.9% of dental students (p = 0.179) were unaware of the meaning of the term “plagiarism” among a sample of 246 participants. Assuming a confidence interval (CI) of 99.9% (Z _α/2 = 3.29) and a « E » of 0.065, the calculated sample size (n) was determined to be 376 students [376 = [(3.29) ² × 0.179 × (1 – 0.179) × 1]/0.065 ²]. However, since our descriptive study involves a finite population (i.e.; a population with a limited number of students), a correction was applied to the sample size.¹⁹ The number of students to be included after correction is given by the formula n’ = (N x n)/(N + n), where “ n ’” represents the corrected sample size, “N” is the size of the finite population (which is 147 in our study), and “n” is the sample size calculated from the standard formula (n=376). Therefore, the corrected minimum sample size (n’) is equal to 106 students [106 = (147 x 376)/(147+376)].

Applied questionnaire

The study employed a questionnaire written in French (See extended data Appendix A³⁸ for the French version and Appendix B³⁹ for the English version), and participants were not given a time limit for completing it. The questionnaire was divided into two sections. The first focused on gathering general characteristics of the students, including age (in years), sex (i.e.; male/female), department such as biomaterials, dental anatomy, general anatomy, microbiology, oral physiology, oral radiology, dental surgery, endodontics, fixed prosthodontics, maxillofacial prosthesis, orthodontics, pedodontics, periodontology, and removable prosthodontics, discipline (i.e.; differentiating between basic sciences and clinical sciences), post-graduate level (i.e.; 1^st year, 2^nd year, 3^rd year, 4^th year), experience level (arbitrarily defined as non-experienced for 1^st and 2^nd years, and experienced for 3^rd and 4^th years), response modality (i.e.; electronic or paper), and dental doctor graduation (i.e.; yes, not yet).

The second section of the questionnaire focused on the French survey, which comprised 11 questions related to KoM.³⁵ The survey was an online quiz accessible on the Laval University website.³⁵ Each question provided three answer choices (i.e.; yes, no, may be), with one point awarded for each correct answer.³⁵ A total score was calculated for each participant and based on this score, three KoM’ levels were identified:

For statistical purposes, two objective KoM’ levels were considered: “excellent or acceptable” and “low”.⁴

An additional question regarding KoM self-assessment by the PGDSs was added at the beginning of the questionnaire⁴: how do you evaluate your KoM? Is it “low”, “acceptable” or “excellent”? Two subjective KoM’ levels were retained: “excellent or acceptable” and “low”.⁴

The 11 questions explore the following four forms of misconduct: Verbatim plagiarism (questions 1 to 3), paraphrasing (question 4), publication ethics (questions 5 to 8), and cheating in classrooms (questions 9 to 11). Scores for each question (i.e.; 11 questions) and for the four forms of misconduct were calculated. The percentages of participants who responded correctly to each question (i.e.; 11 questions) and to all questions of each form of misconduct were calculated.

Why we have chosen the Laval University quiz²³?

While various tools have been established and recommended to assess plagiarism, particularly in terms of, attitudes, acceptance, and practice,^{16,23,24,35,40–44} evaluating KoM is also important, especially among PGDSs. Assessing KoM helps promote academic integrity, ensures ethical research practices, cultivates critical thinking skills, and enhances the quality and originality of scholarly work. Our study appears to be the first to apply a virtual quiz to assess students’ KoM.³⁵ We opted for the Laval University quiz³⁵ to evaluate KoM among Tunisian PGDSs for several reasons. First, the quiz was previously used in a local study involving UHDs.⁴ Second, all available surveys used to assess understanding; attitudes, practices, and knowledge related to plagiarism were in English and lacked a certified French version.^{16,23,24,40–44} Since French is the second language in Tunisia, and since medical teaching is in French, the French Laval quiz appeared to be more suitable for Tunisian PGDSs without the need for translation or validation. Third, the Laval quiz³⁵ is not time-consuming, requiring thoughtful responses, with an average of 15 minutes needed to answer the 11 questions.

Statistical analysis

The distribution of variables was analyzed using the Kolmogorov-Smirnov test.²⁸ The latter involves comparing the cumulative distribution functions of the observed data with those of a theoretical distribution, typically a standard distribution like the normal or uniform distribution.²⁸ It assesses whether the observed data follow the specified distribution or not.²⁸ Quantitative data were presented as mean ± standard deviation (SD) with 95% CI and range (minimum-maximum). Categorical data were presented as relative frequencies. The two-sided Chi-2 test was used to compare the percentages of PGDSs based on subjective (self-assessment) and objective (survey) assessments of KoM. All statistical procedures were performed using a statistical software (StatSoft, Inc. (2014). STATISTICA (data analysis software system), version 12. www.statsoft.com, RRID: SCR_014213). The significance level was set at 0.05.

KoM’ level

The KoM mean score was low at 4.4±1.8, with 85.85% of PGDSs showing a low-level of KoM. Of concern was the finding that a significant percentage of PGDSs underestimated their KoM level (Table 2). These alarming results closely resemble those reported in a previous study conducted among Tunisian UHDs.⁴ Using the same French survey,³⁵ Khemiss et al.⁴ reported that UHDs had a low score of KoM (mean±SD: 5.4±1.9), with 74% of them demonstrating low-level of KoM. Our findings suggest that Tunisian PGDSs have not received adequate guidance regarding plagiarism and misconduct, a growing issue with the widespread availability of Internet resources and artificial intelligence.⁴⁵ This low-level of KoM may be a primary reason for the prevalence of plagiarism in North-African states, such as Tunisia.⁴ The findings also confirm a lack of KoM among both academics and students in Tunisian universities.

The observed low mean score of KoM (4.4±1.8 out of 11 questions) was intermediate compared to studies conducted among postgraduate and undergraduate dental students (Table 5). On one hand, our score was consistent with an Iranian study involving 452 PGDSs, where the mean number of correct answers to 25 plagiarism knowledge questions was low at 6.9±3.3¹⁷ (Table 5). On the other hand, our score contradicted findings from a Saudi Arabian study involving 246 dental students, where the mean plagiarism knowledge scores, assessed through 15 questions,^16,40 were high at 10.4±1.1 for males and 10.5±1.2 for females³⁶ (Table 5).

The high percentage of PGDSs displaying a low-level of KoM (85.85%) was also intermediate compared to percentages reported in the literature (Table 5). On one hand, our percentage aligned with findings from some studies^13,15,16,18 (Table 5). For instance, 53% and 69% of 1153 dental students from 62 American and Canadian dental schools believed they should not be responsible for monitoring the academic integrity of other students and did not consider cheating a serious problem at the university.¹³ In a Brazilian study involving 186 dental students, 78%, 74%, 69%, 48%, and 44% of students reported a lack of respect for image copyrights, a lack of understanding of direct quotes and paraphrasing, ignorance about the different types of plagiarism, and the belief that authors can be wrongly referenced, respectively.¹⁵ An Indian study of 248 dental master’s degree students revealed that 51%, 49%, and 49% of students did not recognize plagiarism as a punishable offense, were unaware of its adverse effects on the community’s health, and did not know that software is available for plagiarism checking, respectively.¹⁶ Similarly, an Iranian study involving 102 dental students, found that 77%, 56%, 53%, 48%, and 45% of students believed using previous method descriptions was justified, self-plagiarism was not punishable, plagiarized parts of a paper could be ignored if they had scientific value, using one’s own previously published work without citation was allowed, and copying with a colleague’s permission was not wrong, respectively.¹⁸ On the other hand, the percentage in our study differed from that in some reported studies^14–16 (Table 5). For instance, among Australian undergraduate students, 56% of nine postgraduate students considered plagiarism a problem, and 90% of 103 first-year students in the Bachelor of Oral Health (BOralH) and Bachelor of Dental Science (BDSc) program reported that their school had clear guidelines for dealing with plagiarism.¹⁴ In addition, 88%, 84%, and 74% of Indian dental master’s degree students were familiar with the term “plagiarism”, recognized it as an act of academic malpractice, and were aware of the mandatory dental council of India point system for scientific publications in academics.¹⁶ Some authors also reported that 100% of Brazilian students enrolled in a dentistry course were aware that plagiarism constitutes a crime.¹⁵

A significant percentage of PGDSs underestimated their low-level of KoM. While 62.26% of PGDSs considered their KoM level to be low, objective evaluation increased this percentage to 85.85% (Table 2). No previous study has evaluated this issue in dental students. Our result was similar to that reported for Tunisian UHDs, where 44.8% of them perceived their KoM level as low, but objective evaluation raised this percentage to 74.0%.⁴ Our finding was partially aligned with another study comprising doctoral candidates at the Faculty of Medicine and comparing attitudes towards scientific misconduct and self-reported behavior (e.g.; plagiarized publication).⁴⁶ This discrepancy between subjective and objective evaluations can be explained by the Kruger-Dunning effect,⁴⁷ a cognitive bias in which individuals with low ability or knowledge in a particular domain tend to overestimate their competence in that domain.⁴⁸ The overestimation of KoM observed in our study aligns with results reported in various subject areas, such as logical reasoning aptitude among university students and clinical practice among specialist physicians.^47–50

It appears that 100%, 100%%, 99.1%, and 63.2% of our PGDSs have low-levels of knowledge concerning verbatim plagiarism, publication ethics, cheating in classrooms, and paraphrasing (Table 3). The percentages of medical and dental students with low-levels of knowledge concerning the above-cited four forms of misconduct vary depending on the specific study and population sampled (Table 5). However, several studies have reported concerning levels of ignorance or misunderstanding in these areas. First, approximately 53.2% of medical students were unaware of what constitutes verbatim plagiarism.⁵¹ This lack of knowledge indicates a significant gap in understanding academic integrity. Second, around 50% of health science students lacked sufficient knowledge of publication ethics key concepts, including authorship and conflicts of interest.⁵² Third, in a survey of dental students, 29% displayed a limited understanding of what constitutes cheating.⁵³ Finally, many students struggle with understanding proper paraphrasing, with studies indicating that 30-50% of medical and dental students could not accurately paraphrase academic content without inadvertently plagiarizing.⁵⁴

Which factors influence the KoM level of Tunisian PGDS?

The two groups of PGDSs with “low” and “acceptable or excellent” levels of KoM showed comparable characteristics, including age, sex, discipline, post-graduate level, experience level, graduation, and response modalities (Table 4). To the best of the authors’ knowledge, while several studies have tried to identify the influencing factors of plagiarism and misconduct in higher education,^{14,17,55–61} no previous specific study has aimed to identify the factors that influence the KoM’ level of PGDS. Our finding aligns with a 2017 study that evaluated the role of plagiarism in 761 conference abstracts written by graduate students and early- to late-career faculty from approximately 70 countries.⁵⁵ The study examined patterns of plagiarism and misconduct among professional academic writers and identified that demographic categories such as geographical location of the current institution, geographical location of the institution where the highest degree was obtained, rank, position, citizenship information, sex, and discipline were not consistent indicators of text-matching.⁵⁵ However, it appears that the post-graduate level, academic year, and students’ age influence the KoM’ level of dental students^14,17 (Table 5). First, one study¹⁴ reported that the percentages of Australian undergraduate students who considered plagiarism a problem were 56%, 63%, and 83% among post-graduate students, fifth-year students of the BDSc program, and first-year students of the BOralH and BDSc program, respectively. Second, in the same study,¹⁴ the percentages of undergraduate students who believed their school had clear guidelines for dealing with plagiarism were 90%, 56%, and 44% among first-year students of the BOralH and BDSc program, post-graduate students, and fifth-year students of the BDSc program, respectively. Third, an Indian study¹⁷ identified that KoM increased as the age and academic year of dental post-graduate students progressed. In contrast to our study, some authors⁵⁶ reported that first-year medical dental students exhibited the highest negative attitude towards plagiarism, followed by second-year and then third-year students. Fourth, Perkins et al.⁵⁸ identified other factors that influence KoM, including English language abilities, personality traits, and student experience and education on plagiarism. Finally, in an empirical study,⁶² hospitality students advanced six key factors in order to explain cases of extreme plagiarism including a poor time management, fear of failure, improve grade point average, personal/family problems, poor level of English, and unsure about referencing and plagiarism policy.

How to explain the low-level of KoM among Tunisian PGDSs?

Four factors may contribute to the observed low-level of KoM amid Tunisian PGDSs. First, there appears to be a perception of permissiveness towards plagiarism within dental medical faculties, with institution leaders exhibiting tolerance towards plagiarists, despite Tunisia implementing initiatives to combat this issue.⁶³ Although a decree was introduced in 2008 allowing universities to autonomously sanction plagiarists, research integrity remains under-discussed, and there is a lack of initiatives to address this concern.⁴ The absence of comprehensive policies may explain the low-level of KoM among PGDSs. Sixteen years after the decree’s publication,⁶³ it is crucial to evaluate whether policies against plagiarism and misconduct in the medical field are being effectively implemented in Tunisia. Second, a significant reason for the prevalence of plagiarism and misconduct is the inadequate emphasis by institutions on addressing these problems.⁶⁴ The absence of mandatory training modules on research and medical writing within medical curricula may account for the observed low-level of KoM.⁶⁵ Writing a thesis, article, or dissertation is a complex task for medical students, requiring training and practice similar to any other academic skill.^20,66,67 PGDSs require support from qualified supervisors to improve the quality of their academic and scientific work.^66,67 Qualified supervisors play a vital role in ensuring the integrity of manuscripts produced by their students, as responsibility for suspected plagiarism/misconduct is shared between senior researchers and young scientists.⁶⁸ Our results in term of low-level of KoM are expected since some previous Tunisian studies including UHDs from different specialties including dentistry, reported a low-level of awareness regarding misconduct.^4,7 These results reinforce the notion that academics who themselves lack knowledge or understanding in a particular subject are less likely to effectively teach or transmit that knowledge to their students. The adage “One who lacks something cannot give it” aptly applies. Third, the low-level of KoM may be attributed to the lack of plagiarism/misconduct detection software tools in medical faculties.⁶⁹ The implementation of plagiarism detection tools has the potential to reduce the prevalence of plagiarism/misconduct within institutions.^4,70,71 However, it is important to consider different perspectives on the effectiveness of these tools. While Stappenbelt and Rowles.⁷² identified that allowing students to check their own work can be beneficial and improve their skills, Youmans⁷³ reported that the availability of such software for students does not necessarily reduce plagiarism/misconduct in their own work. Additionally, Rodafinos⁷⁴ concluded that expecting students to self-check their work might inadvertently encourage them to “cheat” by avoiding matches with the original sources. Since 2022, the FDMM has implemented plagiarism checker software (i.e.; Turnitin), providing all teachers with access to this tool. This initiative aims to effectively raise awareness about plagiarism/misconduct among students, encouraging them to use the software to check their theses and articles prior to submission. The faculty authorities are committed to promoting awareness of plagiarism/misconduct and provide training to prevent their occurrence. Ford and Hughes¹⁴ reported that dental students and staff expressed support for the use of Turnitin across all courses. However, a significant proportion (44%) of PGDSs expressed uncertainty regarding their stance on the use of Turnitin.¹⁴ However, it is important to note that existing plagiarism/misconduct detection software tools are often limited to specific languages and can only identify copied words, while figures and other content may not be adequately compared.^4,68,75,76

Recommendations

Misunderstanding plagiarism/misconduct is a common occurrence, and the lack of clarity surrounding the concept makes KoM essential to prevent it. However, it should be noted that KoM alone is not sufficient. As highlighted by Iloh et al.³¹ despite the awareness of plagiarism/misconduct among Nigerian students, instances of violations were still detected. Therefore, it is crucial to focus on enhancing researchers’ knowledge and attitudes towards plagiarism/misconduct, particularly among junior researchers. According to Poorolajal et al.⁷⁹ there is a significant association between the prevalence of plagiarism/misconduct and an increase in scores of knowledge and attitude towards plagiarism/misconduct, with a decrease of 13% and 16% per one-unit increase, respectively. This highlights the importance of researchers’ awareness regarding plagiarism/misconduct. Therefore, it is crucial for scholastic programs to focus on enhancing knowledge and improving the attitude of researchers, particularly junior ones, towards plagiarism/misconduct. Achieving the following three goals, as proposed by Muthanna,⁹⁰ should be the target: i) Develop skills in creative, critical, and scientific thinking; ii) Encourage creative writing, translation, and publication across various disciplines; and iii) Advance knowledge through scientific research in all fields. To address these concerns, it is recommended to organize workshops, seminars, and scientific writing courses specifically focused on plagiarism/misconduct.^20,90 Interactive seminars have shown effectiveness in addressing this issue, as suggested by Marshal et al.⁹¹ Dedicated modules on medical research ethics and writing, and learning outcomes for educational courses on plagiarism/misconduct should be integrated into undergraduate medical curricula in North-Africa to proactively prevent plagiarism/misconduct.⁴ Combatting plagiarism/misconduct is a shared responsibility, and each medical school should establish its own institutional policies to effectively address this growing phenomenon.⁹⁰

Underlying data

Zenodo: Excel data of the 106 Tunisian post-graduate dental students [Data set]. DOI: https://doi.org/10.5281/zenodo.13212820.⁹²

The project contains the following underlying data:

Extended data

Zenodo: Questionnaire (version française): état de connaissance en matière de plagiat: enquête auprès des résidents en médecine dentaire. DOI: https://zenodo.org/records/10992113.³⁸

The project contains the following extended data:

[Appendix A: Copy of the French questionnaire] (Applied questionnaire).³⁸

[Appendix B: Copy of the translated English questionnaire].³⁹

Reporting guidelines

Zenodo: STROBE checklist for “Knowledge of misconduct amid North-African post-graduate dental students: A cross sectional study”. Doi: https://zenodo.org/records/10692067.⁹³

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).