Privacy Attitudes toward Mouse-Tracking Paradata Collection

Felix Henninger, Pascal J Kieslich, Amanda Fernández-Fontelo, Sonja Greven, Frauke Kreuter

Abstract

Mouse-tracking paradata offers deep insights into respondent behavior but raises significant privacy concerns. This study employs a vignette experiment (N=1,498) to examine how the structure of consent requests and the stated motivations for data collection affect reported willingness to participate (RWTP). We find that while mandatory "joint" consent for both survey and mouse-tracking reduces overall participation compared to a survey-only baseline, it results in a higher total yield of paradata than an optional "separate" opt-in. Surprisingly, detailed justifications for tracking often decreased participation, suggesting that transparency can inadvertently sensitize users to privacy risks.

Introduction

In online research, "paradata"—data about the data collection process—is a valuable resource. Mouse-tracking, which records cursor trajectories, allows researchers to identify confusing questions, detect hesitation, and infer cognitive or emotional states. However, because it captures involuntary motor behavior, mouse-tracking is more invasive than traditional metrics. This raises a critical methodological hurdle: does asking for explicit consent bias the sample by driving away privacy-conscious respondents? This research investigates how consent structure and stated motivations influence a person's reported willingness to participate (RWTP).

Methods

The researchers conducted a vignette experiment with 1,498 participants representative of the German population (recruited via Respondi AG). Participants were presented with hypothetical scenarios to measure their attitudes toward survey participation and mouse-tracking.

The study employed a 2x5 experimental design:

  1. Consent Structure:
    • Joint Consent: Mouse-tracking was a mandatory part of the survey.
  2. Separate Consent: Participants first decided on survey participation, then were asked to opt-in to mouse-tracking.
  3. Motivation: Five levels of justification for tracking, ranging from "None" to specific goals like "Improving the survey."
Experimental flow for Joint vs. Separate consent conditions.
Overview of the 10 experimental conditions based on consent elicitation and purpose.

Willingness was measured on an 11-point Likert scale (0–10), analyzed both as a continuous variable and a binary "yes/no" choice.

Results

The Bundling Advantage

There is a clear trade-off between participant quantity and data completeness. To maximize the total amount of mouse-tracking data, making it a mandatory "joint" part of the survey is most effective. While mentioning mouse-tracking reduced RWTP (86% for survey-only vs. 75% for joint consent), the "Separate Consent" group had a lower total data yield. In that group, 86% agreed to the survey, but only 72% of those respondents opted into tracking, resulting in a total yield of only 62%.

The Transparency Paradox

Contrary to the assumption that explaining data usage increases trust, providing specific reasons (e.g., "improving the survey") often led to lower RWTP compared to providing no reason. This suggests a "sensitization effect," where detailed justifications highlight privacy risks that participants might not have otherwise considered. This aligns with Nissenbaum’s Theory of Contextual Integrity: if the data flow (mouse movements) feels out of place for the context (a simple survey), it triggers a privacy defense mechanism.

Conclusion

The study highlights a strategic choice for researchers: optional opt-ins preserve higher survey participation but sacrifice significant amounts of paradata. Mandatory bundling ensures higher data yield but at the cost of a slightly smaller, potentially biased sample. Furthermore, researchers should be cautious with justifications; over-explaining may inadvertently discourage participation by making the tracking feel more intrusive. Future work should explore whether actual behavior mirrors these reported attitudes and if privacy-preserving technologies (like local feature extraction) can mitigate these concerns.

Summary updated

Abstract

Survey participants' mouse movements provide a rich, unobtrusive source of paradata, and offer insight into the response process beyond the observed answers. However, the use of mouse-tracking may require participants' explicit consent that their movements are recorded and analyzed. Thus, the fundamental question arises how this affects the willingness of participants to take part in a survey at all -- if prospective respondents are reluctant to complete the survey if additional measures are collected, paradata collection may do more harm than good. Previous research has found that other paradata collection modes reduce the willingness to participate, and that this decrease may be influenced by the specific motivation provided to participants for collecting the data. However, the effects of mouse movement collection on survey consent and participation have not been addressed so far.In a vignette experiment, we show that willingness to participate in a survey decreased when mouse-tracking was part of the overall consent. However, a larger proportion of the sample was willing to both take part and provide mouse-tracking data when these decisions were combined, compared to an independent opt-in to paradata collection, separated from the decision to complete the study. This indicates that survey practitioners may face a trade-off between maximizing their overall participation rate and maximizing the number of participants that also provide mouse-tracking data. Explaining motivations for paradata collection did not have a positive effect and, in some cases, even reduced participants' willingness to participate.
BibTeX
@inproceedings{henninger2023privacy,
  author = {Felix Henninger and Pascal J Kieslich and Amanda Fernández-Fontelo and Sonja Greven and Frauke Kreuter},
  title = {{Privacy Attitudes toward Mouse-Tracking Paradata Collection}},
  booktitle = {Public Opinion Quarterly},
  year = {2023},
  url = {https://openreview.net/forum?id=XmT4cXoBlM},
}
Figure