The bi-annual Professional Development Workshop in Quantitative Analysis Using Stata is a great success story.
The highlight of the June 2010 workshop was a two-day presentation by Yulia Marchenko on Multiple Imputation (MI), who is the chief developer of Stata's routines for multiple imputation -mi-. The MI showcased how incomplete datasets can be analysed in using more efficient methods of estimation rather than deleting the casewise observations. Yulia, demonstrated using real-world examples how MI is a simulation-based method consisting of three steps: (1) imputation creates multiple imputed (completed) datasets according to a chosen imputation model, (2) complete-data analysis performs primary analysis of interest on each of the imputed datasets, (3) pooling consolidates results from step 2 into one MI inference using Donald Rubin's combination rules.
The workshop also enjoyed the standard high quality contribution from key MEAFA members:
Richard Gerlach, Vasilis Sarafidis, Andrey Vasnev, and Demetris Christodoulou. Click here for the original announcement.
Overall, the five days were attended by 75 participants from academia, industry (e.g. The Allen Consulting Group, Access Economics, Balllarat Health Services, SBScibus) and government (e.g. Australian Institute of Family Studies, National Prescribing Service, Office of Fair and Safe Work QLD, the George Institute, NSW Department of Community Services, Cancer Council Victoria). As usual, 25-35% of the places were allocated to PhD students at a significant discount.
It is also worth noting that this workshop was audited by StataCorp's vice-president, Alan Riley.