Students and faculty of Boğaziçi University, a top university in Istanbul, Turkey, protested this week after the appointment of Melih Bulu as its new president (news coverage in English in The Guardian, The Times, and Bloomberg).
Bulu’s appointment was criticized as being more political than academic, because he is a close ally of Turkish President Erdogan. Critics also accused him of plagiarism in his published articles and PhD thesis. Bulu has a PhD in business management. During this week’s protests, his 2003 thesis suddenly disappeared from a Turkish repository site.
Here, I take a closer look at Bulu’s thesis to see if these accusations hold water.
Twitter user @06JAnk wrote a thread on Bulu’s PhD thesis, in which they showed side-by-side screenshots of the thesis and older sources.
All these comparisons show that entire paragraphs were literally copied from the source. The paragraphs sometimes end with a citation of the source, but the text is not — at least in any of the examples shown in @06JAnk’s thread — shown within quotation marks.
A closer look with SimTexter
I decided to take a closer look at Bulu’s PhD thesis. A PhD thesis is the place in which a graduate student can not only write about his or her research, but also show that they can read and understand the current literature, and rephrase it in their own words.
Here is a comparison of text from a paper by Gulati, Nohria, and Zaheer, Strat. Mgmt J (2000) (left) and the 2003 Bulu PhD thesis (right), carried out using SimTexter, developed by Prof. Dr. Debora Weber-Wulff.
Here is another comparison, this one of text by Lakhal, Martel, Oral and Montreuil, European Journal of Operational Research 118 (1999) (left) and the 2003 Bulu PhD thesis (right).
Overall view of chapters 3 and 4
Chapter 3 of Bulu’s thesis in particular contains several paragraphs that are identical to text found in papers and books published before 2003, the year in which he completed his PhD.
Here is Chapter 3, with the text that @06JAnk and I identified as being identical to older sources. In this composite image, different colors highlight different sources:
Chapter 4 is not problem-free, either. Here, again, is a composite of that chapter with the non-original text marked using different colors.
I suspect that the true amount of copied text may be even greater than indicated above. Google Scholar appears to have few papers indexed from before 2003. Some of the text appears literally in other PhD theses published after 2003, which might indicate that certain paragraphs are quite popular among graduate students.
While parts of the PhD thesis appear to be unique, several paragraphs are indeed identical to text from older papers and books. The percentage of plagiarism is hard to evaluate, because many older sources are not available as text or OCR-scanned PDFs, and are thus hard to search. But it seems that about 30-50% of the text of Chapter 3 is not original.