February 11, 2019

"PLS-Based Model Selection" - forthcoming in JAIS

Our paper "PLS-Based Model Selection: The Role of Alternative Explanations in Information Systems Research" by Sharma et al, (forthcoming in JAIS) examines the use of information criteria for model selection, such as AIC, BIC and related metrics, in the context of PLS-SEM models (partial least squares path models). Our co-author Marko Sarstedt presents the main points and results in a slick 10-min video.


February 7, 2019

Behavioral Big Data Research in Healthcare: Talk at ISB

Today I'll deliver a talk as part of the 2019 SRITNE Distinguished Speakers series on Behavioral Big Data Research in Healthcare: Challenges and Opportunities

Where: Indian School of Business, AC4 MLT

When: Feb 7, 2019, 3pm

October 28, 2018

Repurposing trees for causal research: talk at BU

I'll be giving a talk on Monday, Oct 29, 2018, at Boston University's Questrom School of Business on Repurposing trees for causal research


Classification & Regression Trees ("trees") and their variants are popular predictive tools used in many machine learning applications and predictive research. While studying causal effects and structures is central to research in many areas, trees are not commonly used in causal-explanatory research. In this talk I will describe special uses of trees that we developed for tackling two causal-explanatory issues: self selection and confounder detection. For self selection, we develop a novel tree-based approach adjusting for observable self-selection bias in intervention studies, thereby creating a useful tool for analysis of observational impact studies as well as post-analysis of experimental data which scales for big data. For tackling confounders, we use trees for automated detection of potential Simpson's paradoxes in data with few or many potential confounding variables, and even with large samples (big data). Our approach relies on the tree structure and the location of the cause vs. the confounders in the tree. I will illustrate these approaches on applications in eGov, labor economics, and healthcare. 

Relevant papers:

  • Yahav, Shmueli, and Mani (2016). "A Tree-Based Approach for Addressing Self-Selection in Impact Studies with Big Data," MIS Quarterly, (40: 4) pp.819-848.
  • Shmueli and Yahav (2018), "The Forest or the Trees? Tackling Simpson’s Paradox with Classification Trees", Production and Operations Management, vol 27 no 4, pp. 696-716.

October 28, 2018

Poster at MIT CODE 2018

Ali Tafti and I presented a poster on Controlling or losing control? Conditioning on covariates in randomized experiments guided by causal structure at the 2018 Conference on Digital Experimentation (CODE) at MIT.  We use Pearl's causal diagrams to show which variables can, should, or should not be conditioned on, and illustrate it on 4 well-known digital experiments.

June 15, 2018

Talk at Università di Padova: "Explaining, Predicting, Describing"

I'll be giving a seminar talk today at University of Padova's Department of Statistical Science (Il Dipartimento di Scienze Statistiche dell'Università di Padova) titled Statistical Modeling in 3D: Describing, Explaining and Predicting. In this talk I extend beyond "explain or predict" to also compare and contrast with "describe".

Where: University of Padova, Aula Benvenuti, Campus S. Caterina 

When: June 15, 2018, 12:30pm