ESRA 2025 Preliminary Glance Program

All time references are in CEST

Sensitive and indirect questioning
Session Organiser	Dr Daniel Seddig (KFN)
Time	Thursday 17 July, 13:45 - 15:00
Room	Ruppert 005

-

Papers

The Use and Pitfalls of Indirect Questioning Techniques: A Systematic Review of Methodological Advances and Research Applications

Dr Thomas Krause (University of Stuttgart) - Presenting Author
Mrs Clara Groß (University of Stuttgart)
Professor Susanne Vogl (University of Stuttgart)

Surveys investigating personal, sensitive, or legally delicate topics face considerable challenges when relying solely on direct-question formats. Indirect questioning techniques (IQTs) offer a critical methodological alternative, helping mitigate data integrity issues caused by social desirability bias and response distortion. This systematic review follows the PRISMA guidelines, analyzing more than 600 studies published between 2019 and 2024 to examine IQT methodologies across diverse research contexts.
Key techniques, including various implementations of the Randomized Response Technique (RRT), List Experiments, and the Crosswise Model, are evaluated in terms of their applications, effectiveness, limitations, and contextual adaptability. Despite decades of use, the expanded adoption of these techniques has triggered renewed discussions about their reliability and context-specific performance, as recent evaluation studies have highlighted persistent methodological challenges and unresolved issues.
Our review addresses essential research questions such as: Which IQTs are most frequently applied in current studies, and which scientific disciplines use these techniques? What sensitive topics and populations are commonly examined using these methods? How do individual characteristics like age, gender, education, or cultural background influence IQT effectiveness? In addition, we identify and summarize potential biases inherent in these techniques based on existing literature, with particular attention to contexts where discrepancies between direct and indirect questioning are most pronounced.
This systematic literature review is conducted with the support of large language models (LLMs), enabling efficient data extraction and synthesis. All LLM-generated results are cross-validated by human coders to ensure accuracy and reliability. By consolidating recent empirical evidence, the review provides a comprehensive assessment of IQT strengths, limitations, and contextual applicability, advancing survey research methodologies to investigate sensitive topics.

Assessing Sensitive Workplace Issues Using the Crosswise Model: Evidence from Banking

Dr Nilgün Özgül (Hacettepe University) - Presenting Author

This study aims to introduce Indirect Questioning Techniques (IQTs) and demonstrate how IQTs are implemented in surveys investigating sensitive behaviors. IQTs are essential for addressing sensitive topics, such as drug addiction, sexual behavior, tax evasion, abortion, illegal hunting, cheating, and so on, as they minimize social desirability bias and encourage truthful responses. Among IQTs, such as the Crosswise Model, are widely discussed in the literature. The Crosswise Model is increasingly favored for its simplicity, participant trust and ability to obtain reliable data.
For this purpose, an online survey was conducted with two groups of private banking employees. The first group was asked a direct question about mobbing, while the second group responded to the same question using the crosswise model. The Crosswise model, designed for privacy-sensitive topics, enables participants to respond securely without directly revealing sensitive information. The results demonstrated similar mobbing prevalences for both groups, suggesting that the crosswise model obtains reliable estimates comparable to direct questioning. Moreover, the findings provide guidance for researchers aiming to improve data reliability and reduce bias in studies on sensitive topics, particularly in online survey settings.

Robustness of Item Count Models in Analyzing Sensitive Questions

Dr Barbara Kowalczyk (SGH Warsaw School of Economics) - Presenting Author
Dr Robert Wieczorkowski (Statistics Poland)

Item count techniques are statistical methods of indirect questioning that are broadly used for dealing with sensitive questions. These techniques require the use of some control variables while the variable under study remains latent, i.e. it is not directly observable. Methodology and theory of item count methods is continuously evolving. In order to ensure high degree of privacy protection and efficiency of the estimation, important questions regarding the type of control variable and method of estimation must be answered. For many years in research practice to estimate the sensitive population proportion moment-based estimators had been widely used. However, in the modern statistical methodology of the item count techniques the problem is treated as a problem of incomplete data and therefore maximum likelihood estimators via expectation-maximization algorithm are employed to address the sensitive latent variable. This parametric approach has definitely many advantages but also introduces some new problems to item count models, especially regarding the choice of the control variable and theoretical assumptions about its distribution. In the paper we analyze the problem of robustness of various item count models to different violations in data distribution. We conduct a comprehensive Monte Carlo simulation study and examine the consequences of violations of theoretical assumptions in the modelling of the latent sensitive variable under study.