Scientific Integrity
Effective October 2023
The University of Arts Linz is committed to compliance with the ‘Guidelines for Good Scientific Practice’. It is a member of the Austrian Agency for Scientific Integrity (OeAWI). This network has developed practical guidelines which also apply to the University of Arts Linz. The guidelines are presented comprehensibly in a document which can be found at oeawi.at/richtlinien.
The most serious violations of research integrity include:
a) fabrication of research results and data
b) falsification of research results and data
c) plagiarism and self-plagiarism
According to the Universities Act 2002 (§ 51 Abs. 2 Z 31) plagiarism occurs when texts, contents or ideas are adopted and presented as one's own. This includes, in particular, appropriation and use of text passages, theories, hypotheses, findings or data through direct, paraphrased or translated adoption without appropriate indication and citation of sources and authors. This also includes self-plagiarism, i.e. the reuse of one's own texts without specifying their first publication.
This also includes content produced with the assistance of AI. In this connection, Austrian copyright law stipulates that users do not gain authorship of output generated by AI systems. Thus, content generated by AI is not considered an independent achievement.
The use of AI tools for scientific work is permitted, but each use must be documented in detail in accordance with these principles. Documentation includes the name of the AI tool, the type of use and the scope of use.
In case of a confirmed serious violation of the guidelines for good scientific and artistic practice, the Rectorate will take according measures. They may range from an order to write a new thesis on a different topic to the revocation of the academic degree. Procedures for cases of misconduct are set out in §15 of the statutory provisions of the University of Arts Linz: Satzung_und_Richtlinien/Studienrechtliche_Bestimmungen.pdf
Plagiarism check by Docoloc
Since winter semester 2022/23, the University of Arts Linz has carried out plagiarism assessment of academic theses. For this purpose, the university uses the software Docoloc. In justified cases of suspicion, seminar papers or not yet completed theses may also be assessed.
The procedure for final theses is as follows: Students submit the thesis as a PDF file to the Department of Studies. Subsequently, the department retrieves the assessment report. These Docoloc reports are then sent to the respective supervisor for further assessment. Hence, there is no ’automatic assessment’. Supervisors continue to play an important role, there is still room for formal experimenting.
Notes for supervisors regarding the handling of reports
A plagiarism search engine such as Docoloc can find and record matches between an examined document and other documents. A qualified assessment as to whether or not an analysed document contains plagiarism cannot be made achieved through software alone. Ultimately, a decision can only be made by a person who examines the document and the Docoloc report as well. The report provides valuable information on matches that can substantiate a suspicion of plagiarism. In order to make the use of a Docoloc report easier, detailed information is provided below on how the report is generated and how the data can be interpreted.
In case of suspected plagiarism in a seminar paper, it is also possible to carry out a plagiarism check. Please contact Andre Zogholy: andre.zogholy@kunstuni-linz.at.
Number of matches found
In addition to general information on the assessed document, the header of the report includes the number of checked and found sentences in the document and their ratio in percentage. This percentage is an initial indication of whether and in what proportion matches were found in the document. However, this does not make it possible to say whether plagiarism has occurred. On the one hand, this value refers to the entire document and does not indicate, for example, whether matches found are randomly distributed across the entire document. Moreover, Docoloc cannot discern whether a match is simply a regular quotation placed under inverted commas. Nevertheless, experience shows that, usually, percentages below 10% do not constitute plagiarism. In these cases, the software often only finds coincidental matches as well as matches in references or bibliographies. Still, plagiarism cannot be ruled out, even with a value of 10% of sentences found. The higher the percentage, the more matches were found and the more likely it is that plagiarism occurs in parts of the document.
Distribution of matching sentences
The distribution of matches found, illustrated in the overview bar of the report, provides further indication of whether the checked document might be plagiarised. If matches are concentrated in one area of the document, it is more likely to be partially plagiarised. The reference list also provides valuable information for plagiarism checking. A list containing only few documents with many matching sentences means that many passages were found in only a few reference documents. In such cases, probability of plagiarism is higher. If, on the other hand, the list only contains documents in which only few matching sentences were found, probability of plagiarism is lower.
Matching sentences
However, not every match is plagiarism. Individual sentences can match coincidentally. Only when several consecutive sentences are identical, the definition of plagiarism applies. It should also be noted that Docoloc does not recognise text citations. As the conventions of citation are very diverse across all disciplines, reliable automatic recognition of cited passages is practically impossible. If cited passages are found in other sources, they are also labelled as matches in the report. Such cases require a decision as to whether the matching text passage is a valid citation or not.
