Staying on track 7: ToTD 32-35

12 November 2015 / Tim Holmes

Last week it turned out I had more to say than even I expected on Tip 31, so rather than hit you with a super long entry containing the usual 5 tips, I decided to split it. So this week I wanted to catch up on Tips 32-35 before we get into the final week of tips which are all quite meaty! Today’s tips are mostly about the practical problems of data collection which are sometimes easy to overlook even if you’ve piloted your study, and if you’ve been following these tips you’ll know by now that it shouldn’t be a matter of “do I pilot?” but rather “when, how long and how many times ?”! One of the reasons for overlooking some of these considerations is that piloting is not always as realistic as it should be. I know it sounds harsh, but you should try to make the pilot as difficult for yourself as it can possibly be, and that certainly means taking a trip to where you will be collecting the data at the time of day you will be collecting it. Here at AI we have developed a bit of a mantra when it comes to data collection and that is “control what you can, and expect what you can’t” – hopefully some of these tips will help a little with that!

Tip 32: Glasses need power – make sure you can bring them back from the dead!

This really shouldn’t be a tip that needs explaining, but for those of you who don’t collect eye-tracking data on a regular basis, this is really about getting to know your equipment and practicing your process for recording data. As far as I’m aware, all mobile wireless eye-trackers require power to run the forward facing scene camera, the infra-red eye cameras and to transmit the data via WiFi to a remote laptop or tablet for live viewing. The problem is, these batteries can discharge their power quite quickly and, moreover, unevenly! The power that a system requires can vary over the duration of a recording because of, for example, temperature changes in the environment or frequency of attempts required to connect to the remote viewing device. When you’re already focussed on running a study and looking after your participant, keeping an eye on the battery indicator is something that can be easily forgotten if it’s not well rehearsed behaviour so if necessary write a reminder on your hand, stick a post-it on your notepad or even set an alarm on your phone that corresponds to 70% of the battery duration. If you’re mobile, always carry spare batteries with you, and make sure that wherever you’re using as your base location for the day can provide somewhere for you to plug in a battery charger. Finally get into the habit of rotating the batteries regularly during a day of data collection – don’t wait for the 10% panic, change it between participants for a fresh battery and get the old one back on charge. Knowledge might be power, but without power there will be no insight!

Tip 33: Qual or quant? It can be both!

Back when I was doing my PhD, the idea of eye-tracking as a qualitative methodology was quite alien to me. I was all about collecting masses of data from more and more participants and then analysing the heck out of it before running statistical tests of power and significance. Today that’s still my preferred way of working, but it doesn’t always have to be like that. One of the great things about eye-movements is how tightly they are tied to cognitive processes and especially decision making. We all do them, so seeing someone else’s is also intuitively quite enlightening, to the extent that even very small samples of participants can be worth tracking if only to gain a high level insight into their patterns of attention. Moreover, someone who is trained in observing those patterns can use them to construct meaningful interviews, or deep-dives, that explore that behaviour through different methods. Now, I need to be careful here, because in Tip 36 which I just posted on Twitter I mention the use of think aloud protocols and will write a whole lot more about these in the next blog, which will hopefully encourage you not to rely too heavily on such methods. But for now, from a purely qualitative perspective, eye-tracking has have much to offer without some of the sample size constraints I alluded to in Tip 24, and can be a great tool for illustrating behaviour to others by allowing them to see through another’s eyes.

Tip 34: When using remote live view with mobile eye-tracking, get to know your WiFi!

So, you’ve piloted your study, you’ve recruited appropriately, you’ve checked your battery power and you’ve taken care calibrating the tracker before sending your participant off to do some well-defined task. In fact, you’ve been doing this all morning and not having any issues … until the stakeholders show up and suddenly you can’t get the live-view working for them to see how brilliantly you’ve designed their research. Can you tell I speak from experience here?! You try everything, turning the glasses off and on, switching the memory, bringing the remote viewer closer to the participant, all to no avail – what else could be going on?

Well, as I mentioned in Tip 32 above, most live view capability with mobile eye-tracking relies on WiFi, which has distinct limitations in terms of range and something called line-of-sight. What this means is the figures the manufacturer has given you for WiFi range are almost certainly idealised and do not apply in the real-world. OK, but you piloted right? And this didn’t happen then, so surely it must be something else? Trust me, I feel your pain here, because it doesn’t automatically follow that because you are using high-grade neuroscientific technology to collect data that you are an expert in wireless data networks – I’m not! Well it turns out that WiFi can fall foul to bandwidth issues because of the amount traffic on the network, just like your broadband at home on Black Friday! For this reason, many systems provide dedicated WiFi networks that won’t be shared by other devices in the vicinity, and on the surface this seems like it should resolve all issues. But WiFi networks rely on radio channels and multiple networks can run on the same channel bringing with them the same bandwidth issues. If you can select the WiFi channel for your system, then you have a possible solution by changing channel and, if you want to be proactive, not using channel 1 which the most commonly used channel for WiFi devices. Remember you can always run a quick WiFi network scan from your mobile phone to get a sense of just how many networks there are in the vicinity of your data collection. If you’re device has a hardcoded (fixed) channel then I’m afraid you’re options are severely limited, but if the problem suddenly occurred when all the stakeholders showed up, maybe suggest they all turn their WiFi off because they might be the very reason you’re having trouble!

Tip 35: With desktop systems you need to limit participant movement – think beyond the chin rest!

Back in the bad old days of eye-tracking, you would have been lucky get away with just a chin rest as the only violation of your participant’s personal space. Very precise eye-trackers often required the participants to bite on a bar as well to stabilise their head movement, and some tracker and studies still require this today. But if you’re lucky enough to be able to work with a tracker that doesn’t require a chin rest, does it mean you shouldn’t use one? And moreover, does it mean that you don’t have to think about other ways your participant could move?

Well of course the answer to these questions is both “no”! Even trackers with large head-boxes (the amount of space in which the head can move while the eyes are still detected by the tracker) will benefit from a stabilised head, and so the only time you should dispense with the chin rest is when your participants would not be comfortable using one as is the case with some age-groups or clinical conditions, or when the act of using the chin rest would severely damage the ecological validity of the study. Having been to plenty of commercial research labs, where chinrests are typically not used, there is one other really important consideration that is frequently overlooked and that’s because it right under their bottoms! The chair you use for eye-tracking can be your best friend or your worst enemy because a badly selected chair will encourage participant movement through discomfort, inappropriate viewing angle or simply through its ability to rotate, recline, or roll across the floor! Because you will have participants of all different shapes and sizes the best thing is to have the eye-tracker and screen on a desk that is adjustable rather than a chair which can be moved every which way. If that’s too expensive, then a height adjustable chair makes sense, but don’t let it rotate and lock the wheels before you start calibration!

I’m into the final week of this block of 20 tips and this week are them some doozies, so follow me on Twitter to keep up, or check back here next week for the final instalment in this series.


Leave a Reply