Thread by @GergelSarah

Thread: “How to WriteUp Your PhD and not get Overwhelmed” I’m spending a week with one of my (co-supervised) PhD students here in India. We are discussing general principles for getting it all done, so I will share some thoughts, in case it helps others at this stage #phdchat

First #protip : Start thinking in Paragraphs. All you will do next, really, is write a few paragraphs, that’s not too bad, is it? Live and breathe your science thoughts as paragraphs, or tidy bundles of thoughts that go well together

Don’t forget that a paragraph is really only 5-6 sentences. All these sentences should support and develop your topic sentence.

Your topic sentence should be short snappy direct and powerful. Put it first. If it’s buried, move it up. If it’s long and wordy, simplify.

Don’t edit too much on first draft of a paragraph, just get the ideas on the page. Barf and Buff (thanks @duffy_ma )

If you really have trouble with that BLANK WHITE PAGE, try voice to text software for a first draft of a paragraph. Most people won’t overthink when they run their mouth.

Day 2: Writing a paper INTRODUCTION! Remember this section can be a mere 5 paragraphs if writing up in manuscript format (which hopefully you are).

And remember, 5 paragraphs x 6 sentences each, means you only need to write 30 SENTENCES and your introduction to your first manuscript is done. That seems doable, right?

Your 5 paragraphs can be:
P1: Set up a general/global/fundamental/universal problem or question
P2: still pretty general but reign it in a bit towards your own focus, mention key gaps
P3: what does reader need to know to appreciate how great your first research question is?

P4: what does reader need to know to see how great your 2nd research question is?
P5: give brief arc of paper to come, state explicit RESEARCH QUESTIONS especially if you didn’t mention them in P3/P4. Brief expectations could be mentioned here along with why your work is novel

Voilà. Now you have your #shittyfirstdraft of your Introduction. Decide on a schedule for each paragraph...can you barf draft one paragraph per day? Then you can be done in one week then start editing and filling in citations.

Day 2 thoughts on How to WriteUp Your PhD and not get Overwhelmed: the METHODS section!

The chronological order in which you actually DID the thing (or figured OUT the thing) may not be the best order to EXPLAIN the thing to the reader—especially when your methods are complex and morphed over time. Consider explaining the easier stuff first.

I like to see the Methods section start with a one paragraph general overview of entire methods. Will help clarify ideal logical flow for the entirety of the METHODS section, but I also like to know in advance what the key pieces are and how they will fit together.

Next paragraph can introduce your study site. Tell me why it’s an amazingly great choice to answer your Research Qu! Don’t just parrot precipitation and terrain stats!

Depending on how much you undertook primary data collection vs made use of existing data, you’ll need at least one paragraph on GeoSpatial Data and one on Field Data, for example. Make it clear what was downloaded directly vs subsequently derived vs directly measured by you...

...however if you MAKE A TABLE of all dependent and independent variables you will analyze, you can probably lose some text. And later you can easily make this Table into bullet points in PowerPoint for your, ahem, defense presentation.

This table should have units of measure for all variables and the scale and sources of data. If you also include the rationale for each variable and/or how you expect it to relate to your research question, then you’re really on top of your game.

If you have complex methods, consider losing a day of your life making a Methods flowchart fig. A picture says a thousand words & you’ll definitely need it for your defense PowerPoint (It may help with Methods overview paragr mentioned earlier). PS You may drink while doing this.

It’s hard to say how many paragraphs your Methods should be, but if you have 5+ paragraphs add a subheader or two (eg Data Collection, Statistical Analysis) Don’t forget the word DATA is plural.

Last & most important METHODS #protip Don’t forget to explain WHY you did each thing. Just listing sequential methods? Don’t do it. Instead, try “In order to understand how thing A impacts thing B, I used method C which is useful for ....” Expect to be asked WHY you did the thing

Continuing on Day 3: Seven Steps to a Better RESULTS section #shittyfirstdraft #phdchat #writeshopwithSarah

This next bit assumes you’ve already done a bunch of analysis...perhaps too much analysis...and started making heaps of figures but are drowning in data and thus having trouble narrowing it all and getting started. Maybe you are freaking out.

Your first step is STOOPIDLY easy. Make a Table. Summarize your variables (mean/median, range, etc). Make a column w/ nicknames, abbreviations or transformations. Don’t forget to label N (sample sizes) especially if you have nested data or you grouped stuff.

You may also wish to write a brief paragraph discussing this table. Then go have lunch or at least chocolate.

Step 2 is also pretty easy. Hopefully you’ve already done some very basic, yet monsterously huge, exploratory correlation matrices, maybe with simple bivariate plots. Organize it all into an Appendix.

Reviewers rarely want to actually see such an Appendix, but your advisor and maybe that one crusty committee member will. YOU will want to refer to it often to understand your data and look for collinearity among your variables used for any fancier statistics.

Step 3: Make a subheader that simply restates the first Research Question (from your Introduction). Add 1-3 figures maximum (or tables, maps, images) answering your question.

If none of your figs really answer your Qu, you have a few options:
- Go make them, silly!
- Tweak Research Q to reflect what you DID answer
- If you decide you need totally new Research Q, consider if it can be another separate chapter. Milk these data for all they are worth!

Step 4 will be to merely redo Step 3 but for the second research question listed in your Introduction. HOWEVER, if Step 3 is hard and you’re stalled out and your advisor is busy, consider presenting some options to your smart labmates.

Realistically, you will probably start maxing out once you exceed 5-7 figures in your chapter as some will be redundant and not essential to your questions. Consider dropping them or putting them in the Appendix.

Steps 5 and 6 are pretty interchangeable. Do them in whatever order works for you. Go back-and-forth as needed.

Step 5 is really about futzing around with your final selection of figures so they don’t look like total shite. Make the most important aspects most visible in the figure (bigger, bolder, highlighted, in colour, whatever). Think very hard about this so the reader doesn’t have to.

Step 6 Write figure captions that “stand-alone” to convey a message. Some reviewers will go straight to your figures without reading much. Label N and p values etc.

Step 7 is just writing a paragraph (or two) for each of these two sections. Keep it dry, that is, do not add conjecture, literature from others, any speculation or creative adjectives. Just stick to the facts.

Apply treats and cat videos at each step or brain will cease function. If advisor doesn’t really know what you’ve been up to let her know around now. I enjoy getting an intermediate (comprehensible, yet imperfect) draft of Intro/Methods/Results before delving into the Discussion