How to write a research paper

Last year, our department did away with a second year qualifying exam for PhD students--faculty hated grading them and after two years in the program, no one ever failed--and instead put in a research competency requirement for second year PhD students.  Here's a brief description of the requirements from our graduate school handbook:

Each student is required to complete a sole-authored research manuscript that meets high standards of scholarship and exposition. The manuscript must meet standards comparable to those used to evaluate contributed papers to major scholarly meetings though it does not have to be publishable quality. The manuscript may represent only a modest departure from published work, but it must be well conceived, justified, and communicated.

The appendix to the handbook--because all academic documents have appendices--contains a few more details about what is involved.

Yesterday, the current class of 2nd year PhD students asked me to meet with them and answer a simple question: "What the hell does that mean?" 

So I've put together some answers to questions that came up in our meeting.  First I will give you the handbook answer, then my interpretation.  Hopefully a few of you out there will find them useful.  I may add to these as I think of more.  Feel free to pile on in the comments (or ask questions that I can answer).

1) What is an acceptable topic for a 2nd year research competency paper?

Handbook: The manuscript must address a question of scholarly significance in economics, ideally in the student’s area of specialization.

Me:  Anything.  If you find it interesting, go with it.  Someone else might find it interesting too.  There's nothing worse than spending 6 months on something that bores the crap out of you.  The whole principle of academic freedom is based on the notion that you can do whatever you want and no one can do anything about it.

2) What should be in a research competency paper?

Handbook: The manuscript should clearly pose the researchable question, appropriately place it within the relevant scholarly literature, and make substantive progress towards answering the question.

Me:  An introduction, a body and a conclusion along with a bunch of technical stuff that few people will ever read.  Seriously, most research papers have a very simple to follow format: 

• Witty Title: Subtitle--subtitle optional, but usually helps to interpret the nondescript but witty title
• Abstract--cut and pasted from the introduction and conclusion
• Introduction and statement of the problem--always written after the conclusion, that way you know what question you really answered
• Background literature--the results of an extensive Google Scholar search on your keywords.  You've probably only read 50% of the papers you cite, but make sure you at least spell the authors' names correctly
• Methods or theory--Fancy mathematical equations with key steps missing to convince the reader that you are way smarter than they are.  Always refer to these steps as 'obvious' to make sure you sound as condescending as possible
• Data--a lengthy description of how you manipulated, mined and mangled the data to force it to fit into your specific problem.  Always give your variables cryptic abbreviations like INCLT100 (Income less than $100,000) or WTF (Wholesale trading factor).  This saves you at least three lines of text at the end of the paper and forces your reader to constantly flip pages back and forth where they might catch some of your genius they missed the first time
• Expected results--a summary of what you think will happen when you run your models.  Always write this section in hindsight, after you already know the outcome.  It makes you look a lot smarter when you 'forecast' the correct results.
• Results--A bunch of tables with too many numbers that few will ever look at, but dammit, it took you two months to come up with results you liked so make sure you put every single bit of work into the paper.
• Discussion--This is where you include the one sentence summary that is the only line anyone actually reads: "The results show...".  Make sure you also cut and paste this line into the abstract.
• Directions for future research--a listing of the real research questions you wanted to answer but couldn't because you came to the crushing realization that you aren't smart enough to figure out the 'obvious' equations in Econometrica or you figured out that all data sucks and there is no way to answer the real question at hand.

3) What do you mean by 'high standards of scholarship and exposition'?

Handbook: The manuscript must be free of glaring technical and expository mistakes.

Me:  Write poorly.  OK, not poorly, but like other economists.  Read a bunch of journal articles and follow their style.  Most economists have no formal training in writing and it shows--but at least it's consistent.  So instead of trying to revolutionize the way economists write, follow the leaders--take easy concepts and make them sound complicated.  Instead of writing "when prices go up, people buy less," write "Price and quantity demanded are inversely related."  Instead of writing "The next dollar is worth less than the last one," write, "Consumer preferences exhibit diminishing marginal utility of income" or "The indirect utility function is assumed to be quasi-concave in income."  Trust me, it's always better to sound smart than to write clearly.

4) Does 'sole-authored' mean I can't get any help with my paper? 

Handbook:  The student’s advisor may offer broad guidance in preparation of the manuscript, but may not contribute to the writing of the manuscript or offer specific editorial assistance.

Me:  Hell no.  We all get help with our papers.  Sole-authored means the paper was written by a tasty fish (rim shot).  Thank you, I'll be here all week.  Seriously, sole-authored just means that you came up with the idea, you ran the models and you wrote the paper.  You are free to ask for as much feedback on those steps as you would like.  If you feel guilty about the amount of help you got on a paper, thank everyone you've ever met and their mother in the first footnote.