I don't think we can generalize the findings from a survey in US universities only. The variation in the findings are not just the result of the nature of the discipline but due to many different factors including variation in departments. According to the article, in highly competitive departments—those in which students have to compete for departmental resources as well as faculty time and attention—graduate students are significantly more likely to observe research and other types of misconduct by their peers and faculty. And this competitive department may vary across universities, not to speak about states and countries. The two most significant disciplinary differences in questionable research practices are in the use of university resources for outside consulting or other personal purposes and in the inappropriate assignment of authorship. In your article you you attributed the hierarchical structure of civil engineering leading to false authorship by junior scientists trying to please their seniors. However, the orginal article says that although inappropriate assignment of authorship to research papers by faculty was reported most frequently by faculty in civil engineering, but students reported the greatest exposure in microbiology. What does that mean? It's hard to pin point that one on any one reason. Also do you know for sure that such a heirarchical nature in this discipline exists in civil engineering in US universities and not in biology? Also the article clearly reports that disciplinary differences in overlooking others' use of flawed data or questionable interpretations of data are largely insignificant so we can't really say that Biology students don't know math and that's why they don't interpret data better. This is defintely an interesting thought and area of study and it may be a good idea for you to explore it further and discover corelations. One way you could do it is by surveying reports of frauds in a particular year and see variations by disciplines. Then compare such data across countries, years etc. to see if there is actual variation or is it just co-incidental. Don't just go by one survey in one country. Good luck!