Using the looking-whilelistening procedure for second language research

In a fundamental respect, LWL is indistinguishable from other looking-time methods. Participants look at pictures (sometimes a single picture, sometimes two to four pictures, and sometimes a complex scene), and they listen to speech that has the potential to direct their attention to some aspect of the visual scene (Fernald, Pinto, Swingley, Weinberg, & McRoberts, 1998; Fernald, Zangl, Portillo, & Marchman, 2008). Like automated eye trackers and the intermodal preferential looking paradigm, LWL generates moment-to-moment data about eye gaze, but the differences exist in how eye-movement data are collected, processed, and analyzed. I will not focus here on the many merits of automated eye trackers (Tobii, EyeLink, ASL, SMI, etc.), but instead on the merits associated with LWL, which has several advantages over other options for collecting data related to eye movements. First, many labs cannot afford automated eye trackers; LWL is cheaper than most other options. Second, not every lab has personnel who are sufficiently skilled in programming to successfully implement automated eye-tracking studies; LWL requires minimal programming skills. Third, recalibration in automated eye-tracking methods sometimes leads to momentary data loss; LWL is minimally prone to data loss. This is particularly advantageous for research on young children, whose inattention-needless to say-contributes to attrition and high rates of data loss. Fourth, the technology needed for portable LWL setups is straightforward, making it a feasible means for collecting data in diverse locations outside of university communities and outside of first-world countries. However, there is one main disadvantage associated with LWL relative to automated methodologies: Coding of eye movements takes time. Using specialized software, the burden is not as great as one might imagine, but LWL data do require more postprocessing relative to automated eye trackers.