Threshold value on similarity index from plagiarism software: a viewpoint

Threshold value on similarity index from plagiarism software: a viewpoint [This question is frequently asked by my Independent Study students]


Student Question: What is the upper tolerance limit on the similarity index assigned to an independent study report?

My answer: The similarity index is only an alarm signal to prompt the marker to study the detailed software report. In my view, A 20% similarity score and below on an independent study report is not alarming. My practice is that an index value below 22% is not alarming. So, 22% is a kind of informal threshold alarm value for my usual plagiarism checking. [I do read all the detailed plagiarism reports on students' assignments anyway.] 

Final judgement on plagiarism status is based on the detailed plagiarism software report (e.g. the highlighted colored zones in the Independent Study report and the sources of the contents indicated by the plagiarism software), not on the similarity index per se. For example, if there is a big chuck of colored content indicated by the plagiarism software and the indicated source is from an academic journal article, another assignment report from another university, or from the public domain on the Internet, etc..., it is a strongly suggested case of plagiarism.