In the last month, I got to give 3 workshops on fairly different topics, and we made all materials available now. I also followed Lisa’s example (thanks for the tip!) and created an OSF website specifically for my talks and workshops.
This blog summarizes the last workshops, and provides links to all materials.
1. Formalizing theories
Together with Don Robinaugh, we taught the workshop “Improving psychological science by formalizing psychological theories: The value of computational modeling” at SIPS 2019. It was a fantastic experience to see so many people care about formalizing theories. The workshop was largely based on Don’s and my experiences when learning about formal models for our recent paper on Panic Disorder; I summarized the paper in a few tweets here.
Here is the abstract of the workshop:
Most psychological theories are “verbal models”: narrative explanations of a psychological phenomenon. Because of the vagaries of language, these theories are vulnerable to hidden assumptions and shortcomings. Moreover, they create fertile ground for questionable research practices. In this workshop, we provide an introduction to theory development through computational modeling. First, we review the advantages of modeling and show how it fosters transparent and cumulative science. We then propose ideas for how to develop and evaluate formalized theories. We provide examples throughout, focusing especially on a computational model of Panic Disorder. Our hope is that participants leave with an appreciation of the importance of formalized psychological theories and ideas for how to develop and evaluate formalized theories in their own work. We will illustrate the process of generating a computational model in R using difference equations. Although prior knowledge in these areas is not necessary to understand the main points of the workshop, it would likely facilitate understanding for parts of the workshop. Importantly, this is not a workshops were we will work hands-on in R, so no laptops are required.
2. Questionable measurement practices and how to avoid them
Like last year, Jessica Flake and I taught another iteration of our measurement schmeasurement workshop at SIPS, with updated materials. The workshop was based on our preprint, which we are currently revising for publication.
Here we are, preparing the workshop the day before:
In this workshop, we will define questionable measurement practices and explain how they make it impossible to evaluate a wide range of potential validity threats to a study’s conclusions. We will demonstrate that psychology is plagued by a Measurement Schmeasurement attitude by showing that QMPs are common, offer a stunning source of researcher degrees of freedom, pose a serious threat to cumulative psychological science, but are largely ignored. We will go through examples and a set of questions that researchers and consumers of scientific research can consider to identify and avoid QMPs. We will discuss how measurement transparency promotes rigorous research, allows for thorough evaluations of a study’s inferences, and is necessary for meaningful replication studies. The workshop is based on our preprint; we have written up the rationale for the workshop in the APS Observer piece “measurement matters“; and curated a measurement resource list for folks who want to learn more about measurement.
3. Introduction to network modeling
Finally, I taught my 15th network analysis workshop, this time at VCU in Richmond, where I’m currently doing a mini sabbatical.
The workshop starts with a conceptual introduction on why items in psychological data tend to co-occur, and what this implies about the constructs such as mental disorders, cognitive abilities, personality, and attitudes. This is followed by an introduction to social and psychological network models; an overview of the network literature in psychopathology (the field where network psychometric models have been used most over the last years); and a summary of important topics (centrality, comorbidity, early warning signals). The main group of statistical models we learn are network models in cross-sectional data. We will use the free statistical environment R to learn the basics about (1) network estimation, (2) network inference, and (3) network accuracy and replicability. We will do so in both lectures and practicals. Please bring your laptops, make sure to have RStudio installed and running; a basic understanding of R is suggested. The last day of the workshop will cover advanced topics and methods, such as network comparisons, modeling of different types of variables, and considerations about causality. We will largely work with freely available data I provide, but if you have your own data you would like to investigate, feel free to bring it along. Note that consecutive sections in the workshop build on each other.
You can find slides, data, syntax, exercises, and a full video of the 3 days here.
Feel free to use and re-use any and all of the materials from these workshops freely. Giving us a shoutout when re-using the materials would be appreciated!