Response Item Network (ResIN) Analysis of Political Attitude Data
Half Day Short Course | Register here
2026 APSA Annual Meeting & Exhibition — Boston, MA
9:00 am – 1:00 pm
How do we identify individual political beliefs that people subjectively find most polarizing? Which cognitive phenomena are sufficient or necessary for the emergence of a structured (and so, polarized) attitude network?
ResIN (short for Response Item Networks) combines the statistical prowess of item-response theory (IRT) with the computational efficiency and ease of interpretation of belief network analysis (BNA). ResIN provides a flexible and easy to implement framework to visualize and make statistical inferences about complex inter-relationships between survey item responses. More specifically, by simulating item responses as charged particles in a latent attitude space, ResIN utilizes the location of political attitude nodes in a latent ideological space.
ResIN models thus benefit from additional information about the spatial intersections of major attitudinal cleavage lines, the relative strengths and correlations between different latent attitude factors, and the approximate position of attitude nodes that bridge different ideological camps.
Workshop participants will be given a detailed theoretical and practical introduction to belief network theory more generally, and ResIN-modeling more specifically. The goal of the applied tutorial in R is to assess the extent to which the spatial location of attitude-bridging nodes correspond to the real world ability of such attitudes to enable compromise between different ideological groups.
Background rationale: The advent of the internet has enabled people to connect, access information, and engage in free debate, leading to remarkable collective projects such as Wikipedia and the Polymath project. However, we are also witnessing a contrasting phenomenon, where individuals are becoming increasingly segregated into ideological camps often engaged in an ideological war against each other. For example, in the United States only 4% of couples are currently composed of a mix of Republicans and Democrats. This environment has shown to be heavily detrimental for collective action potential and so for thriving societies. In the past, research on collective intelligence has focused on the importance of diversity. In this workshop, we would use methods such as ResIN to study the distinction between positive and negative diversity among people’s socio-political attitudes. Under positive diversity, people hold numerous, often diffuse opinion combinations on different topics. In case of negative diversity, opinions on different topics tend to be strongly correlated (for example, in the United States, people who support gay rights tend also to be in favor of women rights and gun control). Therefore, positive attitude diversity harbors greater potential for two or more groups to find a common denominator and enable cross-group collaboration.
ResIN (short for Response Item Network) is an effective method to distinguish between positive and negative diversity. During the workshop, we will examine longitudinal datasets such as the ANES and ESS to observe how attitude networks have evolved over time, and how these structures are connected to variables such as affective polarization (i.e. the dislike between people of different ideological groups), which can limit the potential for cooperation and collective action. Using ABMs we will also explore which conditions are sufficient for producing highly structured networks and the foundations of affective polarization. For example, some preliminary studies suggest that simply the possibility of recognizing the association between two attitudes in other people, together with the ability of learning this pattern, may be enough to produce polarization. Finally, we will combine ABM and data analysis to study possible interventions. Specifically, we will focus on which attitudes may be responsible for most polarization and which types of online interaction could increase or decrease polarization. The participants of this project will receive a detailed introduction to ResIN, including how they can apply the modeling approach in their own research fields and get familiar with the computational implementation in R. The goal of the workshop will be to identify a low number of potential attitude candidates for a simple experiment on polarization reduction in the US and in one other European country. Participants are highly encouraged to bring in their own ideas about additional applications of these research methods and how to test them in novel contexts.
- APSA Annual Meeting Pre-conference Short Courses are half- or full-day events that offer diverse professional development opportunities and allow attendees to connect with scholars from various backgrounds. This year’s pre-conference short courses will be held on Wednesday, September 2, 2026, in Boston, Massachusetts. Sponsored by APSA Organized Sections, Related Groups, and other affiliated organizations. All short courses require pre-registration to attend.
- Register here for the APSA Annual Meeting & Exhibition »
Be the first to comment