Here you can find slides, syntax, and data for my 2 hour webinar on network analysis, hosted by the Transcontinental Computational Psychiatry Workgroup (TCPW). No need to look at these before the webinar.
Analysis of mental health data is usually based on sum-scores of symptoms or the estimation of factor models. Both types of analyses disregard direct associations among symptoms that are well-understood in clinical practice: mental disorders can be conceptualized as vicious circles of problems that are hard to escape. A novel research framework, the network perspective on psychopathology, understands mental disorders as complex networks of interacting symptoms. Despite its comparably recent debut, with conceptual foundations in 2008 and empirical foundations in 2010, the framework has received considerable attention and recognition in recent years.
In this webinar, we will use R to learn about (1) network estimation, (2) network inference, and (3) network stability in cross-sectional data. Regarding network estimation, the state-of-the-art network model for cross-sectional data is the pairwise Markov Random Field or regularized partial correlation network that estimates the conditional dependence relations among items. We will learn to estimate appropriate network models for our data: the Ising Model for binary data, and the Gaussian Graphical Model for metric data. In this first section, we will also cover regularization methods that avoid the estimation of false positive associations in networks. The second topic, network inference, covers graph theoretical measures such as centrality that allow us to interpret networks. What symptoms are most connected with other symptoms? Finally, network stability allows us to gain insight into the robustness of our networks. If time allows, we will conclude the webinar with advanced methods such as the statistical comparison of networks, and how to deal with ordinal and mixed data. Is it noteworthy that network analysis is not limited to psychopathology data, but has been employed to study other psychological constructs such as intelligence, personality traits, and attitudes.
Researchers do not need to run R during the webinar; I will show the estimation routines in R on my screen, and make the data and syntax I work with freely available so that participants can reproduce the analysis after the workshop, and use the code to work on their own data.