Principle INvestigator

A principle investigator is someone with the big idea and the means to get it. They ask the interesting questions and then allocate funding to achieve answers for those questions. They manage their grad students to do the bulk of paper writing, data collection, and figure creation. They scan over first drafts of work using their eye for detail and assign corrections and sometimes rewrite entire sections. They also keep their lab on task creating milestones, assigning deadlines, and revising every word written. The principle investigator is the smartest person of your group. There is nothing wrong with admitting someone is definitely the smartest person in the group...

THe Historian

One of the hardest parts of writing a scientific article is owning up the the fact that you're not special, you're not the smartest person, there are no wow I figured it out moments, and, that for the most part, you contribute very little. But! Science is the sum of its parts so every little bit helps. This is why the first step for solving a research question is to solve as much as you can with other people's work. This requires digging through archives and being very smart about keyword/topic searches. A hopeless search of just vague keywords will waste significant time. If you have a well posed and specific question though you can play a recursive game where you ask yourself

function solve(problem, bibliography)

if problem in paper in bibliography

return answer

else

find subproblem in problem

return solve(subproblem, bibliography)

Eventually you can get the big problem down into manageable chunks. A literature search is a tool. Use it to solve as many of these sub problems as possible so you don't do unnecessary work. Not properly using a literature search would be like trying to code from scratch machine learning algorithms in Python instead of just using...import tensorflow as ts. i.e. Completely unnecessary work that is probably wrong.

The literature search also helps you understand the history of the problem you are working on, and history is a great way to start an epic tale.

THe ANALYST

Okay so you have ticked off as many of your sub problems as possible with other peoples work. (Awesome less work, totally dig it) Now there are bound to be some sub-problems that seemingly aren't answered to the best of your knowledge. This is where you step in. Now be warned the most common mistake is that you haven't subdivided your main problem enough and you are still trying to answer a solved problem, so a good planning phase with the P.I. is critical.

These analysis questions usually should be fairly simple that are just fill in small blanks. If you're experimental then it should be as simple as,

"I need to measure <parameters> and see how they effect <quantity of interest>."

If it is theoretical they could be,

"I need calculate <quantity of interest> to see its dependence on <parameters>."

Notice the reversal of <quantity of interest> and <parameters>. That is the defining separation of theoretical vs. experimental. Due to the remote nature of this course probably everyone will have to do a theoretical paper.

THe artist

Now we come the awful fact that people hate reading, can't understand what you're saying on a first read, and really like pictures, graphs, diagrams, and sketches. So one of the key things to accept is that other scientists will probably only read you're abstract, look at your figures, and maybe read the captions. So somehow you have to condense all your hard work into a picture book.

This is hard. You have to really go through a process where you ask yourself how can I convince someone I'm right with picture. What graphs tell my story best? What variables should I plot? What diagrams should I make? How can I make this caption help guide the reader? There are many many more questions to ask yourself. This is an art.

The stupidest thing I could do is to give you some sort of "checklist" that you could tick off and feel like you hit all the marks. The only way you can develop a knack for figure creation is by looking at other published figures and noticing what you appreciate about them. The same goes for writing as well, the only way you will get better at writing scientific articles is by reading scientific articles. Then once you're good enough you can just retire to reading the abstract and figures...

THe EDITOR

Detail and formatting are the ugly beast of scientific writing. They are the literal first thing you will be corrected on when you get revision back from a journal. LaTeX makes this easier but not automatic. Attention to the very precise details of a journal template is an extremely laborious talent but a necessary one. Can you arrange the text and figures so the first time it is referenced in the text it is on the same page? Can you get all the text in the figure axis to match the body font-size? Are there too many colors? Does the text follow in logical order. The most common mistake early technical writers make is talking about something they haven't defined or introduced yet. This is like walking into a conversation amongst a friend group that goes back years and all they talk about are inside jokes.

For the most part everything is sitting in your head in this giant web of ideas that makes sense only in your head. You have to translate that nasty mess of a web into a linear cascade of ideas. This is unfortunately summed up with one sentence I think undermines just how important it is

"Know you're audience."

There is a lot more I think than just this sentence. You seriously have to imagine you are the person you intend to read the document reading the document. You spent all this time getting to know all the details of your work so well and now you have to forget all of it. Sounds like an impossible task? Well it is, and few are masters at it. This is why there are only a small percentage of famous scientists. They simply could write better than the others that did the same stuff.

Make sure all nomenclature is defined before its used to explain things, be very diligent that you take the active voice instead of the passive, if a figure or equation is in the text make sure it is referenced or else cut it, revise revise revise and then for good luck revise again. Most importantly read the final document out load and listen to yourself. Do you sound ridiculous? Fix it.