IDENTIFYING FACTORS FOR ACADEMIC MISCONDUCT USING LOGISTIC REGRESSION B. Santos 1 , J. Gonçalves 2 , P.G. Almeida 3 1 University of Beira Interior, CERIS-CESUR (PORTUGAL) 2 University of Beira Interior, CITTA (PORTUGAL) 3 University of Beira Interior, GEOBIOTEC (PORTUGAL) Abstract A preliminary statistical approach to the academic conduct of a group of students at the University of Beira Interior (Portugal) was performed using data from voluntary participated surveys. Inquiries were made before and after attending training on the meaning and types of plagiarism, citation rules for academic writing and bibliography preparation. The adopted methodology includes the development and application of stated preference inquiries and the statistical analysis of the collected data. A descriptive analysis was performed and a logistic regression model was developed from the data collected before the training to quantify the student’s behaviours and establish potential associations relevant to the characterization and future monitoring of academic misconducts. The study found that 90% of the students surveyed had never attended a lecture on plagiarism and academic misconduct and that about 50% of the total were aware of a case of plagiarism committed by colleagues. Results revealed that 80% of the students had committed fraud at least once in a test and/or assignment and 83% of those were never detected. The logistic regression model revealed that students aged 18 to 20 and working students have a statistically significant effect on the Logit of the probability of committing fraud in a test and/or assignment. The results suggest that academic misconduct among students is widespread and that measures must be taken to reduce its prevalence and to ensure a merit-based education system. Keywords: Academic misconduct, Plagiarism, Logistic regression, Awareness. 1 INTRODUCTION Currently, dishonesty has become a serious problem that has affected many aspects of academic life. Academic dishonesty is a serious act that not only affects the person who is committing it but also the institution and the academic community. Previous research has demonstrated a strong link between academic dishonesty at university level and future professional misconduct in the workplace [1]. Misconduct in the academic environment may take different forms including plagiarism, multiple submission, falsification, inappropriate collaboration, completing assignments on behalf of others, cheating, copyright infraction and fake degrees. This issue and several approaches to avoid it have been studied for a long and by many. In these approaches, great emphasis is placed on penalizing those found to have misconducted [2]. Recent publications presented a different approach considering cultural and social reasoning as well as education on plagiarism and academic misconduct, targeting prevention instead of a penalty-based approach [1] [3] [4]. The existence of a conduit code at the academia and an information effort, so students understand not only the consequences but mainly what is considered academic misconduct and the nuances of plagiarism, are discussed by several authors [5] [6] [7]. In order to understand the reasons to commit academic misconduct extensive inquiry-based studies were conducted not only to students but also to faculty staff showing their different perception levels on the problem [1] [5]. So why do students commit violations of academic integrity? According to [1] (and also supported by [5] for a different student population) 38% of engineering students say it’s because of the lack of time to study or to complete the assignments (against 40% of the faculty staff stating other reasons and enhancing the mild consequences of such acts). According to this analysis students who study and have a professional activity should be those that most prominently commit academic misconduct. Teixeira [8] has found that working students are in fact less Proceedings of INTED2019 Conference 11th-13th March 2019, Valencia, Spain ISBN: 978-84-09-08619-1 4071