Next week’s Stats Society talk (jointly held with Sydney R Users Forum) is by Statistical rockstar Bill Venables. It’s up the road at UTS: Seminar Room, Level 3, 645 Harris Street (NOTE: this is on the corner of Mary Ann St, not the main tower building) 6:00pm - 6:30pm for drinks 6:30pm - 7:30pm for talk 7:30pm - Light meal with the speaker in the seminar room (you are welcome to join) Cheers, Michael --- O what a difference R makes : Reflections on "Regression Analysis", E. J. Williams,(1959) Abstract: "If you torture the data long enough, it will confess to anything." - Author unknown. E. J. Williams’ book "Regression Analysis" Wiley, (1959), appeared just before the modern era in Statistics really began. This new age is characterised by an exponential growth in computational power, coupled with an explosion in theoretical and practical developments to take advantage of the rapidly developing computational and graphical tools available. The great virtue of Williams’ book is that it presents real (though very small) data sets, mostly in full, and always in a solid context. The analyses are motivated and the computations detailed. From a modern point of view, what is most lacking is any graphical explorations of the data. In this talk I will re-visit three of Williams’ examples and explore them from a modern standpoint, graphically and analytically. These will lead to an historical discussion of the genesis of some modern statistical methods, which when Williams wrote were just getting underway. These include transformations, generalized linear models, random effect extensions and non-linear regression. Bio: Bill Venables is an internationally renowned statistician, with interests in the development of statistical methods and related computation. He currently works with marine scientists on marine biodiversity and stock assessment issues. Dr Venables recent activities have included: biodiversity assessment, mapping, modelling and monitoring in Northern Australia, contributing to fishery sustainability by providing scientific support to management, including improved methods of stock assessment and assessment of environmental impact His books Modern Applied Statistics with S and S Programming, written with Professor Brian Ripley of the University of Oxford, United Kingdom, have become standard references for professionals and advanced students. Dr Venables joined CSIRO in 1999 and is currently a Post-Retirement Fellow. Before joining CSIRO he was Head of the Department of Statistics at the University of Adelaide, South Australia.