Part 2: how Prof. Gilderoy Lockhart provides technical support

What can happen when Prof. Gilderoy Lockhard, Order of Merlin, Third Class, Honourary Member of the Dark Force Defence League; and five-time winner of Witch Weekly's Most Charming Smile Award, became your supervisor

It was a little difficult to talk with my supervisor about my research, because he never used the required experimental techniques himself and also was not strong in the theory. But he possessed a talent to steer a conversation in a direction where he felt comfortable and we spent hours discussing his household repair skills and his family’s summer trips. When he was in a mood for a more scientific talk, he examined a growth of his h-index and remembered minutes from conferences that he attended.

Well, I had enough experience to perform lab work on my own and this part went quite smoothly. The first problem appeared when the time came to analyze the data. I had only a basic knowledge of statistics, my supervisor, however, was an author of more than 120 articles on microbial ecology and should be an expert in statistics. To my surprise, he was unable even of choosing an analysis method (PCA? CA? CCA? NMDS? * Not so many...). On my perpetual questions about which method he considered more suitable, he kept saying that Dr. X advised to use PCA and Dr. Y advised CCA and so on. Questions like why did Dr. X and Dr. Y advised this or that method, or what is the difference between the methods puzzled my supervisor enormously.

One day, trying to solve once and forever all my questions, my supervisor organized a consultancy meeting. On that meeting Ms. F, a docent from our Department, was explaining how ecological data should be analyzed:” … and then you export your Excel file ...and then you press this button ...and then you press that button...” All attempts to find out why she chose, for example, a coefficient of similarity, got no understanding... She told me in an injured voice: “I do it in an autopilot mode! I have more interesting things to do in my life than to find out which coefficient to use!”

After concussion that I got on the meeting, I buried my hope and decided to use CA method. I bought a book “Cluster Analysis for Researchers” by H. Charles Romesburg(my enormous thanks to the Author!), and spent few months in a battle with statistics all by myself.

* PCA – principal components analysis; CA – correspondence analysis; CCA – canonical correspondence analysis; NMDS – nonmetric multidimensional scaling: the tools for data analysis in community ecology.