Why Eye-Tracking?

Web surveys are a common self-administered mode of data collection using written language to convey information. This language is usually accompanied by visual design elements, such as numbers, symbols, and graphics. As shown by previous research, such elements of survey questions can affect response behavior because respondents sometimes use interpretive heuristics, such as the “middle means typical” and the “left and top means first” heuristics when answering survey questions. In this study, we adopted the designs and survey questions of two experiments reported in Tourangeau, Couper, and Conrad (2004). One experiment varied the position of nonsubstantive response options in relation to other substantive response options and the second experiment varied the order of the response options. We implemented both experiments in an eye-tracking study. By recording respondents’ eye movements, we are able to observe how they read question stems and response options and we are able to draw conclusions about the survey response process the questions initiate. This enables us to investigate the mechanisms underlying the two interpretive heuristics and to test the assumptions of Tourangeau et al. (2004) about the ways in which interpretive heuristics influence survey responding. The eye-tracking data reveal mixed results for the two interpretive heuristics. For the middle means typical heuristic, it remains somewhat unclear whether respondents seize on the conceptual or visual midpoint of a response scale when answering survey questions. For the left and top means first heuristic, we found that violations of the heuristic increase response effort in terms of eye fixations. These results are discussed in the context of the findings of the original studies.

Web surveys are an established data collection mode that use written language to provide information. The written language is accompanied by visual elements, such as presentation formats and shapes. However, research has shown that visual elements influence response behavior because respondents sometimes use interpretive heuristics to make sense of the visual elements. One such heuristic is the ‘left and top means first’ (LTMF) heuristic, which suggests that respondents tend to believe that a response scale consistently runs from left to right or from top to bottom. We conducted a web survey experiment to investigate how violations of the LTMF heuristic affect response behavior and data quality. For this purpose, a random half of respondents received response options that followed a consistent order and the other half received response options that followed an inconsistent order. The results reveal significantly different response distributions between the two groups. We also found that inconsistently ordered response options significantly increase response times and decrease data quality in terms of criterion validity. We, therefore, recommend using options that follow the design strategies of the LTMF heuristic

Survey researchers since Cannell have worried that respondents may take various shortcuts to reduce the effort needed to complete a survey. The evidence for such shortcuts is often indirect. For instance, preferences for earlier versus later response options have been interpreted as evidence that respondents do not read beyond the first few options. This is really only a hypothesis, however, that is not supported by direct evidence regarding the allocation of respondent attention. In the current study, we used a new method to more directly observe what respondents do and do not look at by recording their eye movements while they answered questions in a Web survey. The eye-tracking data indicate that respondents do in fact spend more time looking at the first few options in a list of response options than those at the end of the list; this helps explain their tendency to select the options presented first regardless of their content. In addition, the eye-tracking data reveal that respondents are reluctant to invest effort in reading definitions of survey concepts that are only a mouse click away or paying attention to initially hidden response options. It is clear from the eye-tracking data that some respondents are more prone to these and other cognitive shortcuts than others, providing relatively direct evidence for what had been suspected based on more conventional measures.

Eye-tracking is frequently used as an empirical method to assess attention to defined areas or features of visual fields. These range from survey forms to informational and marketing products. A common practice is to use a fixed integrated eye-tracking device with a monitor to display the visual stimulus. This allows the researcher to control presentation of the stimulus, but diminishes real-world application for paper-based documents - especially those spanning multiple pages and requiring respondent interaction with the paper document. This presentation will report on the methodology and data quality for a pretest Westat conducted on behalf of the Food and Drug Administration (FDA). The objective of the pretest was to refine procedures for collecting eye-tracking measures of attention for comparison with responses to self-reported measures of attention as well as information recall, recognition, and comprehension. Forty-one participants who had one of two medical conditions wore eye-tracking glasses while viewing a paper DTC advertisement related to their medical condition. Following the ad reading, participants answered a web questionnaire related to the advertisement and responded to debriefing probes about their experience with the eye-tracking equipment and the questionnaire. We report on lessons learned with respect to achieving successful calibration and accurate eye-tracking data. The findings will be informative for researchers interested in collecting eye-tracking data for real-world stimuli.

Eye-tracking has been used to better understand the survey response process. For instance, eye-tracking has been used to identify questions that are difficult to comprehend, how to present long lists of response options, and to measure the length of fixation on definitions in Web surveys. Grid questions have been commonly used in Web surveys as well as in other types of surveys. The literature shows respondents took less time to answer questions when they were presented in a grid than when they were presented individually across separate pages or screens. The use of grid questions, however, may also be associated with several undesirable outcomes, including higher breakoff rates, higher missing data rates, and straightlining. Relatively little is known about how children answer grid questions. This paper demonstrates how eye-tracking was used to determine the feasibility of using such questions to measure the background characteristics of students in the National Assessment of Educational Progress (NAEP) questionnaire. Fourth grade students were answered both grid and discrete (single-item per screen) versions of questions on tablet computers while wearing real-world eye tracking glasses. This study addresses four research questions related to the use of eye-tracking to test survey questions. First, we examine whether grid items require more effort to answer than discrete items for fourth grade students. Second, we investigate how the processing of sub items changes within a grid. Third, we examine how the processing of questions change over time. In order to address these research questions, we examine difference in the mean number of fixations per word and the mean duration per word for grid and discrete questions. Overall, the study finds support for the use of grid questions with fourth grade students in the NAEP. Implications for the use of eye-tracking equipment to evaluate survey questions are also discussed.

Concerns about the burden that surveys place on respondents have a long history in the survey field. A review of the existing literature shows that the term “burden” is defined loosely and that researchers measure response burden in many different ways. Some measure response burden through properties of surveys/tasks that are believe to impose response burden, such as the length of an interview. Some measure response burden through respondents’ attitudes and beliefs toward surveys, such as interest in and perceived importance of the survey. Others measure response burden through respondent behaviors (e.g., willingness to be re-interviewed) or direct respondent measures (e.g., feelings of burden). All three types of measurement are based on self-reports to survey questions, which are subject to the usual sources of reporting error due to misunderstanding of the survey questions, partial retrieval of information, biased or inaccurate judgment strategy, problems in mapping to the given response options, and more or less deliberate misreporting. In this paper, we will examine the use of eye-tracking equipment to measure burden. In eye-tracking research, task-evoked pupillary responses have been shown to be a consistent index of cognitive load and difficulty. For instance, dilated pupils are found to be indicative of higher levels of cognitive load and difficulty, in essence burden. We will create three measures—mean pupil dilation, peak pupil dilation, and latency to peak—as indicators of burden. These alternative measurements of response burden are free from errors in self-reports and are potentially stronger indicators of burden. We will evaluate the feasibility of using task evoked pupillary responses to measure burden in the survey context by comparing the three indicators to respondents’ self-reports about burden.

Web surveys make it earlier to present images to the respondents than other modes of data collection. A few studies have examined the use of pictorial examples in Web surveys and found that the characteristic of the exemplars (e.g., their frequency or typicality) has an impact on the responses that are collected (e.g., Couper, Tourangeau, and Kenyon, 2004; Tourangeau, Conrad, Couper, and Ye, 2014). Tourangeau et al. (2014) compared visual examples with pictorial examples and found that respondents tended to report more foods consumption when they got verbal examples than when they got pictorial examples. The finding suggests that the pictures may narrow the interpretation of the category of interest. However, the findings also suggested that respondents are more likely to attend to the pictorial examples than to verbal examples. However, no direct evidence of respondent attention was collected to support either argument. Using eye-tracking, the current study compared verbal examples with pictorial examples in a lab setting to examine whether respondents attend to pictures more than words, whether items with verbal examples require more effort to answer than those with pictorial examples, and how the processing of items change over time. To address these research questions, we will examine differences in mean number of fixations and the mean duration for items with pictorial examples and verbal examples. The number of fixation is related to the amount of information that a respondent is processing, while the duration of fixations is related to the amount of difficulty that the respondent is having (Ares et al., 2014). The same food consumption questions with examples used in Tourangeau et al. (2014) will be used in current study.

Holding constant other scale features, the direction in which a scale is presented has been found to affect the resulting survey answers; respondents are more likely to select a scale point closer to the start of the scale regardless of its direction, producing primacy effects (Yan, 2015). What remains understudied is the mechanism underlying this scale direction effect. Two common response processing models are offered as possible explanations for these effects: satisficing, and anchoring and adjusting. The satisficing model treats the impact of scale direction as a special case of response order effect and argues that satisficers sequentially process the rating scale and select the first option that seems reasonable. The anchoring-and-adjustment heuristic assumes that respondents start with an initial anchor (the beginning of a scale) and make adjustments to the anchor until a plausible point is reached. Since both notions predict a primacy effect, it is hard to know which notion offers a better account for scale direction effect. To learn more about what’s behind scale direction effects, we will collect eye tracking data from respondents’ as they respond to a web survey. As the eye movement data (e.g., fixation counts and fixation duration) show directly the amount of attention paid to question components, we will first characterize how respondents process a rating scale and how the processing differs respondent characteristics. Then we will explore which of the two notions account for the scale direction effect. This paper demonstrates how eye-tracking can be used to address theoretical issues related to respondents’ use of rating scales.