The best points in your PhD are the ones when you are engaged in a variety of activities – writing papers, reading articles, attending seminars, juggling thoughts – when every day is a little bit different. It feels that for too many months, whenever anyone has asked what I’m working on the answer has been “video coding”. I have been too deeply embroiled in video coding that I don’t think I’ve explained it particularly well to anyone. To my friends and family, this is where my mind has been for the last few months.
Video coding is the way that I extract data from the video footage that I collected at Wamba, DRC. The way that a researcher codes video varies, depending on the behaviour that they’re looking at, but it roughly goes like this:
1) Watch a video clip at normal speed, looking for the desired behaviour (in my case, a gesture). Pause the video when you see one.
2) In a special program (I use Filemaker Pro), create a new sheet for that behaviour instance. You will already have designed a coding sheet with categories for information that you want to know. It makes inputting data much simpler than using a spreadsheet (but in the end, it will spit out a spreadsheet for you!)
3) Re-watch the behaviour in slow motion as many times as it takes to fill in your data sheet with all the desired information. For example, I write down the name of the bonobo who gestured, what gesture type they used, what body part they used, etc.
4) If you’re coding a sequence of behaviours, such as a gesture sequence, then each behaviour in the sequence takes a different coding sheet. Usually, you will have a field or box that tells you which place in the sequence the particular behaviour occurred.
5) Work your way through the entire video clip to catch all instances of the desired behaviour.
6) Repeat for all of your video clips
Video coding is very repetitive. It’s also something that you get better at with time. For these reasons, it’s important to re-code your earlier videos, to make sure that the coding quality matches the later videos, now that you’ve become a coding wizard! Re-coding simply means re-watching the behaviours that you’ve coded (not the whole video clip! The aim is to check what you’ve already got, not try to find new behaviours), checking that you still agree with the information you entered, and changing it if not. Sometimes you add a field/box partway through the coding process (I added another ‘time’ criteria), so you need to add that to the older behaviours.
It is also important to have a second coder, another person to code part of the video footage, to ensure that your coding has high inter-observer reliability (do different people agree on your coding?) Depending on the complexity of your behaviour, you might have a naïve observer (someone who has no familiarity with the topic) or an expert. Because gesturing is quite complicated, my second coder will be someone who is already very familiar with gestural communication video coding (you know who you are!)
It would not be possible to study gestural communication without video coding. There is no way that I would be able to follow the bonobos in the forest and write down all of the information I need to know about a gesture as it happened! Video coding is therefore a necessary evil. Anyone who does it knows how tedious it is, but it has the remarkable power of allowing us to re-live events in order to accurately describe the minutiae of the behaviour. Now that I’m at the end looking back, I can appreciate it.