For students: tips and science nuggets

Below is a developing list of useful and delightful things that I’ve stumbled across on my adventures in developmental psychology and linguistics. If any of it sparks a question, let me know. My inner geek will be delighted to engage with yours.

Where I think it’ll help I’ve annotated the items below with UG and PG, although if any of my academic advisees *want* to learn R or read about the p-value debate then I heartily salute you and will do all I can to help.

Writing and communicating

baby-writingCommunicating your science clearly is fundamentally important to getting your work published and read in the scientific community, and equally importantly, to feeding back to members of the public. If you find writing challenging (a lot of us do!), take advantage of university resources like writers in residence or training courses. Writing’s one of my favourite bits of my job, and I’m convinced that’s made things easier.

English Communication for Scientists from Nature Education (specifically Unit 2) offers some very sensible advice on how to structure a paper and what kind of style is most effective for writing about your research (TL;DR: keep it simple!). The e-book isn’t specifically aimed at psychologists, so I wouldn’t suggest following its guidelines to the letter, but it’s a very helpful general guide, My favourite quote: “If you can tell a friend about your work, you are off to a good start.”

Barbara Sarnecka (Professor of Cognitive Sciences, UC Irvine) has written a fantastic writing workshop, published online via series of blog posts. She talks about why science communication is important and gives clear, helpful tips about how to plan, keep the momentum going. I wish this had been around when I was battling with my thesis!

Purdue Online Writing Lab has loads of tips on writing in APA 6th format, including sample papers. It got me through my PhD thesis, and it’ll get you through whatever paper or assignment you’re writing too.

The Writer’s Diet helped me trim the unnecessary adjectives and adverbs which determinedly, sullenly, truculently – nay, threateningly! – alas lurked darkly within my first papers.

Zotero is a free bibliographic management package. You can use it with Google Scholar and MS Word to cite papers in a range of different formats and automatically create a formatted bibliography. It’s also a great way of keeping track of your reading, because it lets you keep a list of all the papers you’ve read. I use it because it’s free and backs up online, so in the unthinkable eventuality I’m more than 100 metres away from my own computer, I can log in online. Ask me if you need help getting started. There are similar resources: Endnote, Mendeley, Papers and others. They’re all incredibly useful and do more or less the same thing. I wish I’d started using one during my undergraduate degree.

Stats

UG

2749521.largeKhan Academy really helped stats go “click” when I was an impostor syndrome-riddled PhD student. It’s an online collection of short, clearly-explained videos that deal with the fundamental intuitions about the stats you’ll use in you dissertations and theses. Try the “Statistics and Probability” section.

PG

Bodo Winter‘s tutorials on linear mixed effects models helped me get the intuition of how these analyses work.

MetaLab is an incredibly useful resource for developmentalists which provides easy-to-use effect size estimates derived from meta-analyses of developmental data, as well as clear explanations of why data sharing and meta analysis are so important to good developmental science.

Currently there’s a fascinating debate in psychology concerning the utility of the p < .05 cutoff for statistical significance. In a nutshell, it’s too easy to achieve this level of significance by running multiple different analyses on the same dataset until we get the result we’re after; this “p-hacking” has led to difficulty in replicating the results of psychological studies. Here is an interesting debate about a recent paper advocating the reduction of the cutoff to p < .005.

R Statistics

PG

pirate.jpg
Arrrrgh!

I’ll be honest: R made me cry during my Master’s, my PhD, and yesterday. But it gives you wings. Using SPSS means you can squeeze some numbers into a computerbox and get ten pages of new numbers out with some awkward-looking graphs. Using R means you can write tailor-made analyses and produce beautiful visualisations. If you’re not already experienced in programming it’s a steep learning curve at first, but utterly worth it. If you can find a course to attend early on in your studies, do it! Here are some resources I’ve found invaluable on my Quest to be a Competent Psychologist. Overall though, Auntie Google usually has the answer. Oh, and R is constantly updated by leading statisticians, free, open source, and generally wonderfully left-wing. Winner.

R Studio is where it’s at. You can also use R-commander, which is a more GUI-based interface, but if you do that then you’re less likely to learn how to code in R (disclaimer: I’m still learning!).

R For Data Science is the best free resource I know for beginning to learn R.

 

R cheatsheets provide at-a-glance summaries of some of the most useful functions in common R packages.

Cookbook for R has saved me more times than I can count. It contains simple instructions for common things you need to do. I’ve found it particularly useful for plotting.

R how-tos for some of the incredibly useful things you can do in R from Martin Frigaard.

Other useful/fun (promise!) resources

Useful general tools/tips

CHILDES is an incredible online child language resource. It contains transcripts of conversations between children an adults from a range of ages, backgrounds and abilities. The corpora are transcribed in CLAN, a notation system that allows the database to be full searchable. The database has been the basis of hundreds of influential papers in language acquisition. More recently, researchers have written accessible R packages and online tools which make it easy to dive in and have a play with the data. The fact CHILDES exists means you can do child language research without having a lab and participants. It’s pretty jaw-dropping when you think about it. Thanks, Professor Brian MacWhinney!

Evernote is an online, searchable notebook that syncs across devices. I have two boxes of paper notes under my bed from my PhD, and sometimes I have to brave the spiders and actually read them. Don’t be me. Again, there are alternatives like MS OneNote. Whatever you use, I can’t stress enough how incredibly, life-savingly helpful it is to have an online lab notebook. But no identifiable data, obviously!

How to email your instructor for the first time gives you tips on how to write emails to your (very busy) lecturers effectively. It’s difficult to know how formal to be, and how emails should be structured. Although I think it should be treated flexibly, this guide will help.

Blogs

Professor Dorothy Bishop is a world leader in research into language impairment. Her blog is a valuable resource for anyone interested in the cutting-edge in understanding language disorders, as well as how to conduct science well.

Professor Lynne Murphy blogs about the intricate and frequently very funny differences between US and British English at Separated By A Common Language.

The Speech Talk blog has loads of fascinating articles about accents and dialects in English, with plenty of audio clips. Especially interesting if, let’s say, you’re a trainee Speech and Language Therapist in the North-West of England 😉

AI Weirdness. Janelle Shane trains neural networks on big datasets of languages (mostly) and lets them gibber. Proof if you need it that the robots aren’t coming to get us for a long time.