- Research article
- Open Access
- Open Peer Review
Who would students ask for help in academic cheating? Cross-sectional study of medical students in Croatia
© Đogaš et al.; licensee BioMed Central. 2014
- Received: 15 May 2014
- Accepted: 16 December 2014
- Published: 30 December 2014
Academic cheating does not happen as an isolated action of an individual but is most often a collaborative practice. As there are few studies that looked at who are collaborators in cheating, we investigated medical students’ readiness to engage others in academic dishonest behaviours.
In a cross-sectional survey study in Zagreb, Croatia, 592 medical students from the first, 3rd and 6th (final) study year anonymously answered a survey of readiness to ask family, friends, colleagues or strangers for help in 4 different forms of academic cheating or for 2 personal material favours. Stepwise multiple linear regression models (MLR) were used to evaluate potential factors influencing propensity for engaging others in these two types of behaviour.
Many students would ask another person for help in academic cheating, from 88.8% to 26.9% depending on a cheating behaviour. Students would most often ask a family member or friend for help in academic cheating. The same “helpers” were identified for non-academic related behaviour – asking for personal material favours. More respondents, however, would include three or four persons for asking help in academic cheating than for routine material favours. Score on material favours survey was the strongest positive predictor of readiness for asking help in academic cheating (stepwise MLR model; beta = 0.308, P < 0.0001) followed by extrinsic motivation (compensation) and male gender, whereas intrinsic motivation, year of study and grade point average were weak negative predictors.
Our study indicates that medical students are willing to engage more than one person in either close or distant relationships in academic cheating. In order to develop effective preventive measures to deter cheating at medical academic institutions, factors surrounding students’ preference towards academic cheating rather than routine favours should be further investigated.
- Medical Student
- Intrinsic Motivation
- Grade Point Average
- Extrinsic Motivation
- Academic Dishonesty
There is growing evidence that academic dishonesty is widespread in medical and health care schools worldwide -. This has a detrimental effect on medical practice because students who cheat during medical school follow the same behavioural pattern later on in their work with patients -. Finally, social predictors such as socioeconomic environment and educational system have been shown to have an important influence, most notably in post-communist countries in Europe . We and other research groups in Croatia have demonstrated that academic dishonesty was acceptable behaviour among medical students ,- and that they came to medical schools ready to cheat , indicating that different familial and cultural values acquired long before medical school have had an impact on student behaviour.
Academic cheating most often is a collaborative practice as opposed to an isolated act by an individual. However, there are few studies that examined factors that looked at the persons who are actually involved or would be potentially engaged in collaborative academic cheating. Surveys that investigated attitudes towards engaging others in cheating use general terms such as a “colleague”, “peer”, “another student”, “someone else”, and sometimes a “friend” ,,,, although it can be hypothesized that the view of those involved in collaborative cheating would be different for less conspicuous than for more serious academic transgressions. For example, it could be considered risky to collaborate with strangers in serious cheating behaviour, which would be asked only from individuals in a close relationship. The aim of the present study was to address the following research questions: 1) what is the readiness of medical students to engage others in academic dishonest behaviours; 2) who would students ask for help in academic cheating; and 3) is it more acceptable to ask more individuals for help with routine material favours than it is to seek help with academic cheating? The willingness for including others in academic cheating was also correlated with individual factors described as potential predictors of student cheating : gender, year of study, academic success and motivation for study/work.
We performed a cross-sectional study at the University of Zagreb School of Medicine in Croatia in 2008. The study population included medical students attending the 1st, 3rd and 6th (last) year of study. They were asked to voluntarily complete an anonymous, self-administered questionnaire about their propensity for cheating and motivation. The questionnaire was distributed to a convenience sample of all students present at lectures or seminars when the highest student presence was expected. The same approach and introductory lines by investigators were used for all groups tested. The study was approved by the Ethics Committee of the University of Zagreb School of Medicine.
The questionnaire consisted of three parts: general data, questionnaire about students’ propensity to ask for favours and the Work Preference Inventory (WPI) .
General questions included age, gender, grade point average (GPA) at the final year of secondary school for the 1st year students or GPA at the previous year for the 3rd and 6th year students.
The questionnaire tested students’ willingness to ask for unethical favours in academic situations and their readiness to ask for material personal favours in everyday situations from 4 different types of individuals: 1) family member or close relative; 2) best friend; 3) colleague; and 4) stranger.
Academic favours were chosen from each of the four specific clusters from the hierarchical cluster analysis of 11 self-reported cheating behaviours at the same school : ask another person to sign the lecture attendance sheet instead of the respondent (A1); ask another person to let the respondent copy answers during a written exam (A2); ask another person to send answers to the respondent during a test, using a cell phone (A3); and ask another person to use private connections with the examiner to arrange a passing exam grade with the examiner (A4). The favours were listed from the least to the most serious transgression.
Personal material favours presented to the respondents were: lending the respondent €50 for three days (P1) and lending a car for a day (P2).
A pilot test was conducted for the part of the questionnaire that related to favours was piloted with 4 educators and students from the University of Zagreb School of Medicine in order to improve the clarity, consistency and face validity of the questionnaire.
The third part of the questionnaire was the Work Preference Inventory (WPI) for assessing individual differences in intrinsic and extrinsic motivational orientations of college students . The WPI consists of 30 statements in which a respondent is asked to choose a single statement that expresses to what extent that statement described her/him: 1. Never or almost never; 2. Sometimes; 3. Often; and 4. Always or almost always. The questionnaire was scored on two primary scales (intrinsic and extrinsic motivational orientation, each with 15 statements), each subdivided into 2 subscales: outward and compensation for extrinsic motivation and enjoyment and challenge for intrinsic motivation (10 and 5 statements, respectively). The score range for each statement was from 1 (completely disagree) to 4 (completely agree).
The frequencies of positive answers indicating students’ willingness to ask others for help in academic cheating and their readiness to ask the same persons for personal favours were analyzed separately for each favour. Average scores per item were generated by summing positive answers for academic or personal favours and dividing them by the number of items in the scale, so that the overall scores had the possible range from 0 to 1 indicating the average per-item fraction of positive answers. Differences in the mean per-item score of academic favour requests between genders and study years were tested by Student’s t-test and ANOVA, respectively, whereas nonparametric counterparts, Mann-Whitney and Kruskal-Wallis tests, were used to estimate gender and year of study median differences in per-item score for personal favours. The binomial test was used to compare the expected with observed frequency of positive answers.
Motivation scores were expressed as mean (± standard deviation, SD) per-item scores on each of the two scales (intrinsic and extrinsic) and four subscales (enjoyment, challenge, outward, compensation). Summative scores on the scales were divided by the number of items in each scale, so that the overall scores had a possible range from 1 to 4, making comparisons between the scales possible.17 The mean scores for an item in the motivation scale were tested between genders using Student’s t-test, whereas differences among the years of study were tested with ANOVA and subsequent Bonferroni post-hoc analysis. Cronbach’s alpha was used to estimate internal consistency of all scales and subscales used in the study.
Stepwise multiple linear regression models of the summed score favour requests with gender, year of study, repetition of the year, grade point average and motivational subscores as independent variables, were used to evaluate potential factors influencing propensity for engaging others in cheating. In the model of personal favours, score data were first logarithmically transformed to correct for non-normality of error distribution.
Statistical analyses were performed using MedCalc, version 12.5 (MedCalc Software, Ostend, Belgium). The level of significance was set at P < 0.05.
Characteristics of the study population
Grade point average:
4.8 ± 0.2
4.0 ± 0.5
4.1 ± 0.4
Asking for academic and personal favours
We first assessed the internal consistency of the academic cheating survey. Overall Cronbach’s α for 16 items in the academic cheating survey (4 cheating situations × 4 types of persons to be engaged in the cheating act) was 0.850, 95% confidence interval [CI] 0.832 – 0.868, demonstrating a high level of internal consistency. For the personal favours survey, the overall Cronbach’s α was 0.723, 95% CI 0.687 – 0.756 due to a smaller number of items (n = 8) and also indicated a high level of consistency.
Additionally, in both surveys all items had acceptable item-total correlations of rpb ≥ 0.317 and no item had a Chronbach’s α-if-item-deleted value greater than the overall α.
Observed and expected proportions of responders with intention to engage different number of persons in a cheating behaviour
Observed proportions, No. (%)* for engaging:
Proportion expected by chance
A1. Sign lecture attendance sheet
A2. Let copy answers during test exam
A3 Send answers by cell phone during test exam
A4. Use personal connection to pass exam
P1. Lend car for a day
P2. Lend € for three days
For personal material favours, the respondents would predominantly ask one or two persons. For a choice of single person it was predominantly “family” (16.7% and 13.7% for P1 and P2, respectively), whereas for a choice of two persons it was exclusively family or friend (Table 2, Figure 1). Only a single respondent chose somebody else for lending a car.
The proportion of students who would ask for unethical academic favours depended on the seriousness of the offence: 88.8% of them were ready to ask at least one person (either in close or distant relationship) to sign a lecture attendance sheet in their place, 83.9% would ask someone to copy test answers during an exam, 50.0% would ask for test answers sent by a cell phone, and 26.9% would use someone’s connection with the teacher to pass an exam. The percentage of students who would ask someone for a non-academic favour (83.0% for lending a car and 81.7% for lending money) was similar to that of asking for the least offensive academic transgressions.
Who are the cheaters?
Average per-item scores* for unethical requests (mean ± standard deviation) with respect to study years, type of requests, and gender of medical students
P -value (gender comparison) †
0.42 ± 0.27
0.35 ± 0.23
0.45 ± 0.24
0.43 ± 0.21‡
0.37 ± 0.21
0.39 ± 0.21
0.45 ± 0.24
0.37 ± 0.17
0.40 ± 0.26
0.38 ± 0.22
0.42 ± 0.16
0.34 ± 0.17
Respondents’ motivation for work/study
The total sample of medical students had higher intrinsic than extrinsic motivation (mean item score on 1–4 summary scale 3.07 ± 0.36 vs. 2.67 ± 0.36, respectively). Female students had higher intrinsic motivation than male students (3.09 ± 0.34 vs. 3.02 ± 0.39, P = 0.0048, t-test), due to higher scores on the Enjoyment subscale (3.34 ± 0.35 vs. 3.20 ± 0.41, P < 0.001). There were no differences in their extrinsic motivation, but male students had higher scores on the Compensation subscale (2.72 ± 0.52 vs. 2.56 ± 0.49, P < 0.001). Across study years, intrinsic motivation significantly decreased in the 3rd as opposed to the 1st year of study (2.99 ± 0.41 vs. 3.11 ± 0.31, ANOVA post-hoc P = 0.002), mostly because of a significant decrease in the Enjoyment subscale, but increased again at the end of medical school. The overall extrinsic motivation did not change over the years, although there was a significant decrease in the Outward subscale in higher years relative the 1st year of study (2.77 ± 0.39 vs. 2.66 ± 0.41 vs. 2.63 ± 0.36, respectively, ANOVA P = 0.001, post hoc tests for comparison with 1st year P from 0.003 to 0.009).
Predictors of willingness to ask favours
Predictors of respondents’ (n = 592) readiness for ask for help in academic cheating or for a personal favour, ordered by the predictor strength from strongest to weakest
Academic cheating adjusted R2 = 18.0%
Score on personal favours scale
3.7 x 10−8
Higher readiness, more requests
4.1 x 10−4
Higher enjoyment, fewer requests
Extrinsic motivation – Compensation
Higher compensation, more requests
Grade point average (GPA)
Higher GPA, fewer requests
Male gender, more requests
Year of study
Higher year of study, fewer requests
Personal favours adjusted R2 = 3.6%
Male gender, more requests
Year of study
Higher year of study, fewer requests
Our study demonstrated that medical students were willing to involve others in academic cheating. Personal material favours, such as asking to borrow some money or a car, were predominantly reserved for family and friends, similar to the most serious offence in academic cheating – asking someone to use a personal connection with the examiner to arrange a passing exam grade for the respondent. There was a clear gradation of academic misbehaviours for which students would ask persons outside of the narrow circle of family and friends. Students had no problems in engaging strangers in academic cheating, especially for favours they considered less serious, such as signing a lecture sheet or copying answers during a test. There were no major differences across study years and genders.
The limitations of our study include the self-reported nature of the data collected on cheating and the fact that we asked them about their intentions and not actual conduct, as it is not clear whether asking about future intentions is a more sensitive indicator of cheating than directly requesting a report on past cheating . Also, the results of the survey may have been affected by a potential overlap in the meaning (a “friend” who could also be a “colleague”). In the Croatian language, in which the survey was performed, the difference is clear and respondents had the ability to select the “friend” response to describe their friendly relationship with a fellow medical student. The responders’ choices for personal favours (exclusively family or friend) indicated that these categories were not confused. Furthermore, the study was performed in a single medical school, limiting the generalizability of our findings. However, the response rate across all study years in this study was high, and the responses obtained about their willingness to ask others for help in cheating were similar to reported prevalence of admitted cheating behaviour in the same and other medical schools in Croatia over the last decade ,. For example, the prevalence of admitted academic cheating of medical students for the four academic misbehaviours that were included in this study were 89.1% for signing for someone else on the attendance sheet, 52.2% for copying answers during a test and 1.6% for using private connections to pass an exam . The consistency of data over time confirms that the phenomenon of high academic cheating in the observed setting is real and well-established, confirming the relevance of the results of the current study on students’ willingness to ask for help in cheating to a real-world situation at similar academic institutions.
The major predictor for engaging others in academic cheating was the score on readiness to ask someone for a personal favour, which explained a half of the total variance in the regression model that also included intrinsic and extrinsic motivation, gender, grade point average and year of study. On the other hand, the only positive predictor for personal favours was male gender.
Taken together, these results indicate that academic cheating is perceived by medical students as an activity mostly outside of one’s personal accountability. In an environment that is very permissive to academic cheating and cheating in general ,-, students’ choice of persons who they would be ready to involve in cheating demonstrates their awareness of the acceptability of the behaviour. For smaller academic transgressions they would not mind asking a favour from anyone available, including strangers. As the seriousness of cheating increases, they were more selective in their choice of individuals, being aware of the transgression and the consequences for the other person. It is important to emphasize that our study did not address academic collusion, which is considered as a cheating behaviour when students engage in collaboration and is defined as an “unacceptable level of shared work in the final assignment” . All of the behaviours used in our study survey were clear transgressions of individual behaviours and specified in the curricular norms of the institution .
Social and educational system predictors of academic cheating are particularly strong in post-communist countries in Europe, not only in medicine but also in other disciplines, such as economy and business, leading to drastic differences in cheating: the prevalence of academic cheating in these disciplines was 88% in Eastern European countries compared with 5% in Scandinavian countries . A recent study from Iran, using a randomized response technique to control for responses to sensitive issues, demonstrated that 93% of medical students impersonated an absent student during a mandatory class . We showed in our previous report from 2004 that 94% of the students from the same medical school as in the current study admitted to cheating at least once during their studies . Comparing the data from this study to those from our 2004 study, students’ attitudes towards cheating have become even more lax. Academic misbehaviour offered to respondents were chosen from 4 very distinct clusters of academic transgressions, based on reported behaviour, approval of such behaviour, perception about its frequency and willingness to report cheating. In the current study, the prevalence of possible engagement of others in the least serious offenses (signing an attendance sheet and copying during a written exam) was similar (88.8% and 83.9%), whereas these differences were higher for their respective clusters in the previous study (87% and 56%, respectively for the perception that such behaviours are common). A multi-campus study of medical schools in Croatia showed that not only did almost all of the freshmen admit to cheating, but most of them enrolled with cheating experience from high schools . Croatian medical students also exerted pressures on and engaged others in asking their teachers for undeserved academic benefits, as 45% of the medical teachers reported experiences of pressures intended to ensure that the student obtains a passing or a better grade on an exam .
While previous studies suggested that men generally self-reported more cheating than women , this study and other studies from Croatia , did not find major gender differences. In our study, female students had less willingness in engaging others in academic cheating only in the 1st year of study and asked fewer personal favours. However, gender was a weak predictive factor in multiple linear regression model, with male students being more willing to ask either for a personal favour or help in cheating. The lack of major gender differences can be explained by long-standing female predominance at medical schools in Croatia ,.
On the other hand, the only change that occurred during the course of the medical curriculum was that female students showed greater willingness for engaging others in academic cheating in the 3rd year then when they began medical education. This may be due to the fact that female students often adopt the male pattern of behaviour when engaged in a traditionally male predominant profession , particularly in the setting of clinical teaching at university hospitals, which starts in the 3rd year and is dominated by men, particularly in leadership positions . The 3rd year, where we observed the highest propensity for engaging others in cheating, was also characterized by a significant decrease in intrinsic motivation of the students in our study. This finding may be related to the pressures to be successful while facing the difficulties of medical education and the strong hierarchy of clinical work, which may have affected their behaviour. Our previous study of moral reasoning in medical students showed that moral reasoning deteriorated in the third year of the curriculum, when students enter the clinical rotations .
The finding that the study year, gender, GPA and motivation were weak predictors of the willingness to engage others in cheating in the multiple linear regression model indicates that collaboration in cheating at medical schools is predominantly influenced by the contextual and social settings ,. This means that interventions aiming at improving individual integrity of the students, such as those related to their academic success or fostering of intrinsic motivation for study and work, cannot be effective without structural changes in the institutions, educational system and society in general. Students are also aware of this, as recently reported by Henning et al. , who performed a qualitative study of reasons for cheating and suggestions for strategies to decrease cheating among medical and pharmacy students. The great majority of students (91%) expected and were willing to respond to cheating control by external bodies, from parents and teachers to institutions as well as agencies and governments .
Our study demonstrated that medical students are willing to ask others for help in academic cheating. Academic cheating is a part of a wider human network that supports corruption in societies , and translates into corruption in the health sector which can then lead to a perception of lower quality health care . In the situation of increased migrations of health workers within Europe and globally , the social aspect of willingness to engage others in academic cheating and the consequences of such behaviour are not only the problems of individual schools or countries but need to be addressed at the societal level.
The authors thank Prof. Zoran Đogaš, MD, PhD, and Shelly Pranić, MS, from the University of Split School of Medicine for critical reading and language editing of the manuscript.
The study was supported by the research grant from the Ministry of Science and Technology of the Republic of Croatia, No. 216-1080314-0245 to Matko Marušić. The publication of the article was funded from the University of Split Excellence Award to Ana Marušić. The sponsors had no role in study design, as well as the collection, analysis, and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication.
- Baldwin DC, Daugherty SR, Rowley BD, Schwarz MD: Cheating in medical school: a survey of second-year students at 31 schools. Acad Med. 1996, 71: 267-273. 10.1097/00001888-199603000-00020.View ArticleGoogle Scholar
- Desalegn AA, Berhan A: Cheating on examinations and its predictors among undergraduate students at Hawassa University College of Medicine and Health Science, Hawassa, Ethiopia. BMC Med Educ. 2014, 14: 89-10.1186/1472-6920-14-89.View ArticleGoogle Scholar
- Kukolja Taradi S, Taradi M, Dogas Z: Croatian medical students see academic dishonesty as an acceptable behaviour: a cross-sectional multicampus study. J Med Ethics. 2012, 38: 376-379. 10.1136/medethics-2011-100015.View ArticleGoogle Scholar
- Montuno E, Davidson A, Iwasaki K, Jones S, Martin J, Brooks D, Brooks D, Gibson BE, Mori B: Academic dishonesty among physical therapy students: a descriptive study. Physiother Can. 2012, 64: 245-254. 10.3138/ptc.2011-13.View ArticleGoogle Scholar
- Henning MA, Ram S, Malpas P, Sisley R, Thompson A, Hawken SJ: Reasons for academic honesty and dishonesty with solutions: a study of pharmacy and medical students in New Zealand. J Med Ethics. 2014, 40: 702-709. 10.1136/medethics-2013-101420.View ArticleGoogle Scholar
- Gitanjali B: Academic dishonesty in Indian medical colleges. J Postgrad Med. 2004, 50: 281-284.Google Scholar
- Turren JF, Staik IM, Gilbert DK, Small WC, Brling JW: Undergraduate Academic Cheating as a Risk Factor for Future Professional Misconduct. Proceedings of the First ORI Research Conference on Research Integrity: 19–20 November 2000; Bethesda. Edited by: Steneck NH, Scheetz MD. 2002, Office of Research Integrity, Bethesda, Maryland, USA, 42-49.Google Scholar
- McCabe DL, Trevino LK, Butterfield KD: Cheating in academic institutions: a decade of research. Ethics Behav. 2001, 11: 219-232. 10.1207/S15327019EB1103_2.View ArticleGoogle Scholar
- Glick SM: Cheating at medical school. BMJ. 2001, 322: 250-251. 10.1136/bmj.322.7281.250.View ArticleGoogle Scholar
- Teixeira AAC, Rocha MF: Cheating by economics and business undergraduate students: an exploratory international assessment. High Educ. 2010, 59: 663-701. 10.1007/s10734-009-9274-1.View ArticleGoogle Scholar
- Hrabak M, Vujaklija A, Vodopivec I, Hren D, Marusic M, Marusic A: Academic misconduct among medical students in a post-communist country. Med Educ. 2004, 38: 276-285. 10.1111/j.1365-2923.2004.01766.x.View ArticleGoogle Scholar
- Bilic-Zulle L, Frkovic V, Turk T, Azman J, Petrovecki M: Prevalence of plagiarism among medical students. Croat Med J. 2005, 46: 126-131.Google Scholar
- Hren D, Vujaklija A, Ivanisevic R, Knezevic J, Marusic M, Marusic A: Students’ moral reasoning, Machiavellianism and socially desirable responding: implications for teaching ethics and research integrity. Med Educ. 2006, 40: 269-277. 10.1111/j.1365-2929.2006.02391.x.View ArticleGoogle Scholar
- Kukolja Taradi S, Taradi M, Knezevic T, Dogas Z: Students come to medical schools prepared to cheat: a multi-campus investigation. J Med Ethics. 2010, 36: 666-670. 10.1136/jme.2010.035410.View ArticleGoogle Scholar
- Hejri SM, Zendehdel K, Asghari F, Fotouhi A, Rashidian A: Academic disintegrity among medical students: a randomised response technique study. Med Educ. 2013, 47: 144-153. 10.1111/medu.12085.View ArticleGoogle Scholar
- Henning MA, Ram S, Malpas P, Shulruf B, Kelly F, Hawken SJ: Academic dishonesty and ethical reasoning: pharmacy and medical school students in New Zealand. Med Teach. 2013, 35: E1211-E1217. 10.3109/0142159X.2012.737962.View ArticleGoogle Scholar
- Amabile TM, Hill KG, Hennessey BA, Tighe EM: The Work Preference Inventory: assessing intrinsic and extrinsic motivational orientations. J Pers Soc Psychol. 1994, 66: 950-967. 10.1037/0022-35188.8.131.520.View ArticleGoogle Scholar
- Sutton A, Taylor D: Confusion about collusion: working together and academic integrity. Assess Eval High Edu. 2011, 36: 831-838. 10.1080/02602938.2010.488797.View ArticleGoogle Scholar
- Babic A, Brekalo M, Juric S, Puljak L: Pressures and interventions imposed on medical school teachers regarding students’ examination grades. Med Educ. 2013, 47: 820-823. 10.1111/medu.12219.View ArticleGoogle Scholar
- McCabe DL, Trevino LK: What we know about cheating in college. Change. 1996, 28: 28-10.1080/00091383.1996.10544253.View ArticleGoogle Scholar
- Danic A, Hadzibegovic I, Loparic M: Status of women in small academic medical communities: case study of the Zagreb University School of Medicine. Croat Med J. 2003, 44: 32-35.Google Scholar
- Hren D, Marusic M, Marusic A: Regression of moral reasoning during medical education: combined design study to evaluate the effect of clinical study years. PLoS One. 2011, 6: e17406-10.1371/journal.pone.0017406.View ArticleGoogle Scholar
- Bobak M, Murphy M, Rose R, Marmot M: Societal characteristics and health in the former communist countries of Central and Eastern Europe and the former Soviet Union: a multilevel analysis. J Epidemiol Community Health. 2007, 61: 990-996. 10.1136/jech.2006.052134.View ArticleGoogle Scholar
- Nikoloski Z, Mossialos E: Corruption, inequality and population perception of healthcare quality in Europe. BMC Health Serv Res. 2013, 13: 472-10.1186/1472-6963-13-472.View ArticleGoogle Scholar
- Costigliola V: Mobility of medical doctors in cross-border healthcare. EPMA J. 2011, 2: 333-339. 10.1007/s13167-011-0133-7.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.