Relevance & Research Question
Web tracking has become a critical method for studying digital behavior. However, most data collection efforts in this area face significant data quality issues, particularly due to limitations in sampling design that hinder representativeness and generalizability. Additionally, factors influencing participation in web tracking are poorly understood, as is the precise definition of participation. This study addresses these gaps by (1) identifying key factors across three stages of web tracking participation—consent, tool installation, and attrition—and (2) examining whether these factors differ across probabilistic and nonprobabilistic sampling strategies.

Methods & Data
This study draws on data from the GESIS Panel.dbd, which incorporates various panel recruitment strategies. Probabilistic recruitment includes sampling from the official population register and leveraging existing studies such as the 2023 German General Social Survey (ALLBUS). Nonprobabilistic recruitment involves multiple ad campaigns on META platforms (e.g., Facebook, Instagram) and open referrals. Following recruitment, participants were invited to a web tracking study, and we conducted multiple regression analyses to compare factors influencing participation across these sampling approaches.

Results
Our findings reveal that different factors influence each stage of web tracking participation, with distinct patterns emerging for consent, installation, and attrition. Additionally, recruitment strategy significantly impacts participation, with stable differences between probabilistic and nonprobabilistic samples, indicating persistent unobserved biases and underscoring the role of sampling design.

Added Value
This study advances existing research by exploring web tracking participation factors in both probability and nonprobability samples. Our findings provide actionable insights for researchers and practitioners to enhance data quality and generalizability in web tracking studies, thereby strengthening the reliability and applicability of findings in digital behavior research.

Relevance & Research Question

Do individuals from different socioeconomic status (SES) groups use the internet differently? The digital divide extends beyond disparities in access to adequate internet access. Being online offers potential benefits, but people may differ in their knowledge, opportunities, and capabilities to take full advantage of these benefits.

Methods & Data

Using a linked dataset from the German General Social Survey (a large, probability-based survey) and respondents' web surfing behavior (GESIS Web Tracking), this study explores whether online behavior varies by SES background. Respondents participated in a web tracking study, which collected data on every individual website visit over two months following the installation of a browser plug-in. This pilot study includes more than 4 million website visits from 500 participants, with the linked data providing around 340,000 website visits from 106 respondents. The websites visited were classified into content-based categories, such as “education,” “job search,” “healthy living,” “personal finance,” alongside categories like “shopping,” “sports,” and “video gaming” using two different third-party service providers. Through regression analysis, I examine whether SES is associated with particular types of website visits and whether this relationship is moderated by first- and second-level digital divide factors, such as access to fast internet connections and digital literacy.

Results

Preliminary results based on the webtracking data set show that characteristics of participants such as their education and political interest are associated with more frequent visits of particular website types such as news websites. Linking of the data sources will be possible next week. I will be able to show results on differences in online behavior by SES at the conference.

Added Value

Differing online behavior may ultimately contribute to inequalities in education, the labor market, or financial well-being, potentially mitigating or reinforcing existing social inequalities. Thus, understanding these behavioral differences is crucial for reducing structures that exacerbate inequality. Results of this work may demonstrate the substantive value of linked survey and web tracking data.

Relevance & Research Question

Recent research demonstrated that intermediaries like Facebook, Twitter, search engines and news aggregators can broaden the diversity of news people consume. However, the personalized content users encounter on intermediaries remains a black box, as tracking tools have been limited in their ability to capture in-platform content. A central debate persists around whether algorithmically curated content diverges from preference-driven selective exposure -- where users actively choose to engage with specific news items based on their interests. Using a web tracking tool that captures the public content on Facebook and Twitter as well the content encountered on other websites, this study examines whether content exposure through intermediaries affects the diversity of news accessed by German internet users.
Methods & Data

This study examines three months of web browsing histories, including survey responses, from a sample of German internet users (N = 739) to investigate how the use of intermediaries -- and the diversity of content encountered on these platforms -- affects direct visits to news websites and the diversity of news encountered. The analysis uses random-effects within-between (REWB) models, with data hierarchically structured at individual and daily levels.
Results

Preliminary findings align with existing research, demonstrating that engagement with intermediaries is positively associated with greater diversity in news exposure. Further, results show that individuals who engage with intermediaries tend to have richer and more diverse news diets within intermediary platforms compared to the diversity encountered through direct visits to news websites. However, this increased diversity comes with a caveat: users are also more likely to encounter hyperpartisan news through these platforms.
Added Value

This study offers insights into the role of intermediaries in the dissemination of information and their potential impact for information diversity in digital environments. By providing a nuanced perspective on the mechanisms driving news diversity, it advances the field's understanding of the relationship between online intermediaries and diverse media diets.