Disclosure decisions: the combined effects of reciprocity, comparisons, and question sequences

TL;DR

This study explores how question order, comparisons, and reciprocity affect people's willingness to share sensitive personal information.

Contribution

The research identifies optimal combinations of question sequences and framing to maximize voluntary disclosure in sensitive contexts.

Findings

Invasive questions asked later in the questionnaire lead to higher disclosure rates.

Statements justifying the purpose of information requests increase disclosure percentages.

Easy questions first help establish reciprocity and encourage more divulgence.

Abstract

The study examines how different information-collection tools influence individuals’ voluntary disclosures of sensitive, private information. The individual and combined effects of dyadic relationships comparisons, and question sequences were tested.

A 3 × 3 × 3 between-subjects main experiment was utilized. Using 27 unique online data collection methods, 1,276 participants were randomly assigned to a condition to measure actual voluntary disclosure using a pre-tested set of 18 questions with varying levels of invasiveness, covering themes such as drug use, sexual preferences, medical conditions, and consumption choices.

Findings show that invasive questions asked later in the questionnaire maximize divulgence. Using statements to justify the purpose of the information request enhances disclosure percentages. Participants would also mimic the disclosure behavior (or abstention from it) of a majority. Combinations of these factors indicate that disclosure reciprocity is best established by having easy questions asked first. The use of a triple combination was better than the use of these factors individually, except for comparison-inducing messages.

Our findings expand our understanding of how reciprocity can be established in question-based settings (i.e., through chatbots, structured interviews, medical questionnaires, etc.). Implications emerge on aspects that elevate concerns, for example, by overwhelming individuals with information on comparisons and data use. We map interactions and propose future research to further isolate combinations of disclosure techniques as they are more reflective of real-life situations than their isolated examinations.

Figures