Eye Tracking Methodology

Abstract

This chapter has two main aims. The first is to introduce readers to the range of eye tracking technologies currently available, and describe the basic principles on which they operate. The second is to provide readers with an understanding of the main determinants of eye tracking data quality and the ways in which this can be quantified. A greater understanding of how eye tracking technology works, and the key determinants of data quality has two important benefits. Firstly, it will improve the likelihood of researchers being able to maximise the quality of the eye tracking data they generate themselves, using eye tracking technology that is appropriate for their research goals. Secondly it will increase their ability to critically evaluate eye tracking research produced by other researchers. Holmqvist et al. (2011) identify several distinct categories of eye tracker users, including usability and media consultants as well as those interested in human-computer interaction and gaze controlled interfaces. This chapter assumes that the majority of readers are academic researchers, probably working in the fields of psychology or cognitive neuroscience and related disciplines, and as such are most likely interested in using eye tracking technology to establish point of gaze and oculomotor dynamics—and also be concerned with issues of accuracy, precision, sampling rate and timing. Section 8.2 of this chapter concerns eye tracking technology, and starts with a brief historical overview of early eye tracking techniques, followed by a description of some of the less common technologies that can still be found in research published today—albeit often in relatively niche areas. The vast majority of commercial eye trackers that are currently available are video based—and as such this approach is covered in most detail. Video-based eye tracking methodologies can be divided (broadly) into two categories: stationary, screen-based systems and mobile head-mounted systems. Clearly these two types of equipment are generally used in very different research scenarios—and differences in the technology concerned make comparisons across these categories difficult if not meaningless. As such they are treated separately in the chapter. Section 8.2 may provide some useful information for readers who are considering purchasing an eye tracker—but it is important to note that this is not its primary purpose—nor is it meant as an exhaustive description of the pros and cons of all currently available eye tracking techniques and commercial models for all possible research scenarios. Not only would such an undertaking become rapidly outdated, it would also involve comparing apples with oranges. Indeed, care has been taken to avoid mentioning specific manufacturers or models where possible. Hopefully any readers interested in purchasing an eye tracker will, after reading this chapter, be equipped with sufficient knowledge to make informed decisions as to which type of eye tracker would be most appropriate given their research goals—and be able to critically evaluate performance claims made by manufacturers and ask the right questions of sales people. Those already in possession of an eye tracker may gain a better understanding of how it works, and its capabilities and limitations, and be more confident that they are using it to its full potential. Section 8.3 considers eye tracking software—not only is software a central component of most commercially available eye trackers, and an important determinant of data quality, it is also one of the main factors determining the ease with which the technology can be used, and the range of uses to which it can be put. Again, this entire section is intentionally generic, and is not intended as an exhaustive evaluation of all currently available software. Section 8.4 addresses the other key topic of the chapter—data quality. It starts with an attempt to define important key terms such as “accuracy” and “precision”. The section then considers how such concepts might be quantified. The final part of this section is intended to offer practical advice for maximising data quality—including some basic information on the importance of setting up participants and getting a good calibration. The precise setup and calibration details will differ depending on the type of eye tracker used—so the advice contained in this section again intentionally generic and aims to outline basic principles and good practices that apply to all or most eye tracking scenarios.