Causal Inference Through Propensity Score Analysis

Add to your list(s) Download to your calendar using vCal

• Dr Michael Posner, University of Cambridge and Villanova University.
• Thursday 07 July 2016, 19:00-21:00
• For further details, please register at www.meetup.com/Data-Insights-Cambridge/.

If you have a question about this talk, please contact Dr Sobia Hamid.

In data science, we work with “found data” that is gathered in the process of doing business. The appeal of these data are that they are readily available and plentiful (“the four Vs of Big Data”). However, this type of data exploration presents challenges for inferring causality. Causal inference is established in clinical trials through the randomised controlled trial, which is considered the gold standard. Found or observational data, however, suffers from violations of the reliance on even distribution of groups due to subjects self-selecting into the study or into modalities based on various characteristics or circumstances.

The propensity score method is frequently used for analysis of observational data in fields including medicine, psychology, education, and survey research. It is essentially stratification on multiple factors using a single summary measure and is performed by calculating the conditional probability (propensity) of being in the treated group given a set of covariates, weighting (or sampling) the data based on these propensity scores, and then making statistical inferences using the weighted data.

Dr. Posner will provide an overview of propensity score analysis and review methods of data selection or allocation of weights. Various methods are compared using empirical analysis and via an application on sending patients to respite care and simulations are described and discussed to compare methods.

Speaker Bio: Dr Michael A Posner is an Associate Professor of Statistics in the Department of Mathematics and Statistics at Villanova University, where he has been since 2005 after completing his PhD in biostatistics at Boston University. His publications and research span the fields of statistics education research, biostatistics, public health, healthcare research, statistics and the law, educational research, and analysis of observational studies. He is currently on sabbatical at the University of Cambridge.

His research grants, totaling over $3 million, have come from the National Science Foundation, the Agency for Healthcare Research and Quality, and the Villanova Center for Nursing Research. His most recent grant is on Data Science education for undergraduate students. Dr Posner won the 2010 Villanova University Faculty Award for Innovative Teaching, the MAA ’s 2012 Alder Award for Distinguished Teaching, and the ASA ’s 2012 Waller Education Award. He is the founding director of the Center for Statistics Education, a Center for Excellence in the College of Liberal Arts and Sciences at Villanova University.

This talk is part of the Data Insights Cambridge series.

This talk is included in these lists:

• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge talks
• Chris Davis' list
• Data Insights Cambridge
• For further details, please register at www.meetup.com/Data-Insights-Cambridge/
• Interested Talks
• Trust & Technology Initiative - interesting events
• bld31
• ndk22's list
• ob366-ai4er
• rp587

Note that ex-directory lists are not shown.