I’m here to share my top five tips for researchers who want to get confident with statistics. I know, I know – you’re really busy and statistics is one of those things that you’d love to get your head around at some point, but it’s just not your priority right now.
These are small hacks that you’ll be able to fit into your normal work so that you can learn on the go. Before I share those tips though, let’s start with a pep talk…
You are more than capable of using statistics
In my experience, most people are capable of using and understanding statistics, even though a lot of them don’t feel like they can. Most people feel like it’s too complicated, too techy, too mathematical.
Why do they feel like that? As a statistician, it pains me to say this, but it’s often the fault of statisticians. Maybe you took a stats module as part of your degree and it left you feeling more confused. Maybe you’ve asked the central stats team at your university for advice and walked away feeling a bit belittled and none the wiser. Maybe you felt like you’d nailed it, only to have a stats reviewer slate the analysis in your paper.
Most statisticians don’t do it on purpose. When they speak statistics all day, they often forget that those concepts aren’t familiar to everybody. They often forget that they’ve got the benefit of years of experience and learning in that area.
I’m sorry if you’ve had an experience like that and it’s knocked your confidence. Statistics isn’t rocket science. It just takes an understanding of some basic ideas, intuitive thinking and the right tools. This is a fab YouTube video that really focusses on those ideas.
I’d love to help to build your confidence back up. I hope these tips will start you on that journey, because you most definitely can do this.
Here’s how you can…
1. Know that we all Google stuff
Does it ever feel like everybody knows how to handle their data, while you feel overwhelmed? Rest assured that we don’t all know what we’re doing all of the time. Simply knowing that can really help with imposterish feelings.
Statistics is a massive subject. It’s wide and it can go deep, so you might end up understanding some areas well, while others remain a mystery. The trick is to:
- Have a good understanding of the basic principles.
- Get to know the topics you use regularly really well.
- Know when something requires you to get more knowledge (either by yourself or from someone else).
- Know where you can get reliable information on this.
So, take a deep breath and make peace with the fact that not even the experts know it all. I’d honestly struggle to do my job without Google (and I’m pretty awesome at what I do!).
Some good places to start if you’re looking for advice are my Facebook group (The Health Researcher Hub), The Analysis Factor, and Laerd Statistics. I also love the UCLA website for step by step walk-throughs of different statistical programmes because they also show you how to interpret the results.
2. Read the statistics analysis sections of papers
When we read papers, most of us are guilty of skimming over the details and jumping straight to the juicy results.
Try to start a new habit here. Still go to the results first if that’s what you like to do but then go back and read the statistics analysis section in the Methods.
This is helpful in a few ways:
- You’ll start to notice patterns in the types of analysis that were used.
- It will make the analysis feel less intimidating.
- You’ll start to be able to see what makes a good statistics analysis section for when you come to write your own.
If you find the terminology in the analysis section baffling then this website is a great starting point for explaining the phrases you’re likely to come across a lot.
3. Use statistics as often as you can
Like when you learn a new language, if you don’t practice the new statistics that you’ve learned you won’t become familiar with them. So, instead of asking somebody else to do it, start by having a go yourself, and get help if you need to.
The more you practice, the easier it gets. It’s that simple.
[Ed: I really like the Not So Standard Variations podcast, by Roger Peng (John Hopkins) and Hilary Parker (Etsy, Stitch Fix). Podcasts like this are a good way to decode statistics.]
4. Practice talking about statistics
When you try to explain something, you often have to understand it really well yourself before you can make it understandable to other people. By talking about statistics, you’ll find out which bits you do understand well. And you’ll also figure out which bits you’re less confident about so then you know you need to find out more about those.
My top tip for this is, while you’re learning statistics, to buddy up with somebody else who is learning, too. You can explain to one another what analysis you’ve been doing knowing that the other person won’t mind you making mistakes. You might also be able to figure out gaps in knowledge together, and you’ll probably learn faster because you’ll hear about what they’ve been working on as well.
If you’re on the hunt for a buddy, you could ask around in your department or find a virtual buddy by asking in a Facebook group like mine. Following accounts like Significance Magazine on Twitter can also help you to earwig in on stats conversations before you’re ready to join in yourself.
5. Ask questions
If you don’t understand or know something then ask! It’s the quickest way to move forward so you can crack on with your analysis. It can take a bit of confidence but asking publicly can often help others who are unsure about the same thing as well. If you’re studying, don’t be afraid to ask questions of your lecturer!
If you’re not studying then there’s some great forums out there so have a look around for one that suits the way that you like to learn. Aside from my Facebook group, here are two of my faves:
Getting started really is that simple. Clear away the imposter feelings then get reading and talking about statistics, and mostly importantly get stuck in and have a go!
Author Bio: Danielle Bodicoat works with academic health researchers helping them to get confident with using statistics to analyse their data.