For me, the hardest part of a piece of scholarly writing is the conclusion or discussion section, the "so what?" that comes at the end. For the uninitiated among you, most papers in my field look like this:
Title (must include a colon, otherwise it's not scholarly)
Abstract (say what you did in 75 words)
Lit Review (all the previous studies on the topic, and theory, may include hypotheses or research questions)
Method (how you did your study, how you measured stuff)
Results (what you found, statistical tests, etc.)
Discussion (sigh, yeah, that "so what?" part)
I tend to write the Results first because, for me, it's more fun, and I am a God of SPSS, a skill mastered in grad school on a mainframe with a 300-baud modem and now perfected with a top-end system with multiple screens. While doing Results I'll mess with the Literature Review, but in my head I have the whole paper from start to finish, I just tend to write it in a hodge-podge fashion.
Except for that Discussion thing. That I put off till the end.
Why? In part because I know what the study is about, but I'm so close that I have a hard time backing off and writing the big picture, the "so what?" of the study, especially without repeating myself. We all know the routine: say what you're gonna say, say it, and they say what you said. That leads to academic suckiness.
Is there a trick to not sucking in that last section? Maybe I should write it first? Maybe I should finish the study, let it sit for a week, then go back and read it, write the grand "what it all means." Also in that section you need to list the weaknesses or limitations of your research and, let's face it, no one enjoys that either. Limitations? Me?
Funny thing is, this paper is full of big concept ramifications. It's all about democracy, about elections, about winners and the consent of the losers. And Fox News. And Karl Rove. And tables full of multiple regression analyses. And more Fox News. So it should write itself, except that never in my life has something ever written itself.
An editor told me long ago -- if writing comes easy, you're not doing it right.