This is a gentle reminder that registration for the StarFlame Summer Institute on Advanced Methods Training will close in less than two weeks. (Please let us know if you need more time to submit the registration.) Short descriptions of selected courses are shown below. Tuition is minimal and there is also a discount for economic hardship. For more details, please visit:
We also invite you to join our mailing list (https://forms.gle/WrwW8Cg4iGvYj5kL8) to stay updated with new course offerings, scholarship opportunities, and other events. Please feel free to share the information with others. Many thanks!
Correlation is not causation. Counterfactual causal inference is one of the most revolutionary inventions in statistics and social science research methods. Based on the potential outcomes framework, this course presents the state-of-art of causal inference methods. Selected topics include the concept of potential outcomes, randomized experiments, matching, propensity score methods, sensitivity analysis, instrumental variables, regression discontinuity, difference-in-difference, synthetic control, marginal structural models, nonbinary treatment, mediation, and interference. Examples and code will be provided. Knowledge of R (or Stata) and logistic regression is required.
Widely used in social and behavioral sciences, structural equation models (SEM) provide a highly valuable framework for integrating measurement and mediation analysis into statistical modeling. An SEM allows linking latent (or unobserved) variables to observed ones in order to address measurement error. It also permits the specification of multiple dependent variables simultaneously and an assessment of mediation based on the decomposition of total effects into direct and indirect effects. This course emphasizes the intuition behind SEMs and their applications. Examples, exercises, and code (R or Stata) will be provided. Knowledge of R (or Stata) and linear regression is required.
Bayesian analysis is revolutionary in that it can estimate complex models with no analytical solutions and also incorporate prior knowledge. This course introduces Bayesian analysis in a conceptually accessible way, with a focus on application and interpretation. Selected topics include the history of Bayesian analysis, the Bayes's theorem, the basics of likelihood theory, the Markov chain Monte Carlo methods, applications of the Bayesian methods for estimating generalized linear models and multilevel regression models, and post-estimation analysis (model diagnostics and comparisons). Examples and R code will be provided. Knowledge of R and linear regression is required.
Network analysis shifts the research focus from individual units to their connections and so brings both theoretical and methodological innovations. Interest in network analysis has EXPLODED recently, due to new advances in statistical modeling and the rapid growth of network data. This course covers the major methods to collect and analyze network data. Selected topics include basic network analysis (centrality, positions, and clustering), the exponential random graph model for modeling network formation, causal analysis of network effects, the stochastic actor‐oriented model for dynamic network analysis, and meta network analysis for combining and comparing estimates from multiple random graph models. Case studies and R code will be provided. Knowledge of R and logistic regression is required.
StarFlame (https://www.starflame.org/) aims to provide the most cutting-edge, efficient, and affordable training in advanced research methods within 10 hours.