Comments on: Statistics is the study of uncertainty http://hea-www.harvard.edu/AstroStat/slog/2008/statistics-is-the-study-of-uncertainty/ Weaving together Astronomy+Statistics+Computer Science+Engineering+Intrumentation, far beyond the growing borders Fri, 01 Jun 2012 18:47:52 +0000 hourly 1 http://wordpress.org/?v=3.4 By: Simon Vaughan http://hea-www.harvard.edu/AstroStat/slog/2008/statistics-is-the-study-of-uncertainty/comment-page-1/#comment-181 Simon Vaughan Sat, 05 Apr 2008 12:14:30 +0000 http://hea-www.harvard.edu/AstroStat/slog/2008/statistics-is-the-study-of-uncertainty/#comment-181 Thanks for the link to the Lindley article, I found it and the proceeding discussion very stimulating, even though I couldn't follow all the arguments in detail. This paper also made me appreciate for the first time the difference in usage of the term 'model' by statisticians and astronomers ('clients'). As an astronomer I automatically interpret the word 'model' as referring to the physics of the system under study, and specifically the way this affects its electromagentic emission (since usually we are passive observers of light from cosmic sources). But 'model' seems to be used by statisticians to describe the assignments of probability functions needed for statistical analysis and inference. Astronomers usually take the latter for granted, assuming Poisson or Gauss-Normal distributions, often without explicitly stating so. Maybe we can agree of a terminology offensive to neither statisticans nor astronomers that will allow the two meanings to be used without confusion (in astro-statistics papers)? Any suggestions? How about probability-model and physical-model as a first attempt? Thanks for the link to the Lindley article, I found it and the proceeding discussion very stimulating, even though I couldn’t follow all the arguments in detail. This paper also made me appreciate for the first time the difference in usage of the term ‘model’ by statisticians and astronomers (‘clients’). As an astronomer I automatically interpret the word ‘model’ as referring to the physics of the system under study, and specifically the way this affects its electromagentic emission (since usually we are passive observers of light from cosmic sources). But ‘model’ seems to be used by statisticians to describe the assignments of probability functions needed for statistical analysis and inference. Astronomers usually take the latter for granted, assuming Poisson or Gauss-Normal distributions, often without explicitly stating so. Maybe we can agree of a terminology offensive to neither statisticans nor astronomers that will allow the two meanings to be used without confusion (in astro-statistics papers)? Any suggestions? How about probability-model and physical-model as a first attempt?

]]>