Prob/Stats Evolution of Data Analysis Techniques: A Journey through History

[brainpushups]

Is anyone aware of any book(s) that presents the history of data analysis techniques? I'm most interested in how scientists during the enlightenment dealt with uncertainties and how techniques for dealing with uncertainty developed over time.

 

[brainpushups]

Thanks for asking...
Lately I've been reading a bit about the history of science. There are many good books that detail the development of theory. For example, Rene Dugas' A History of Mechanics which was published back in the 1950s does an excellent job of detailing the history of the development of the principle of virtual velocity/work and many other aspects of mechanics - from correct conclusions reached by incorrect procedures to how the same conclusions were reached with better developed mathematics.

There are also many good books on data analysis, but none that I have found detail anything about history. I am simply curious about the history of the rigor of measurement. The language of error analysis must have developed in tandem with the formalization of statistics, but I have been unable to locate any books that detail the history of these ideas.

I recognize that deductive reasoning still played a major role in the development of early Newtonian physics, but scientists at the time also based conclusions on empirical evidence. Even Kepler's rejection of circular orbits was based on data not fitting a model. What kind of criteria did early scientists use to claim that data was in support of their hypotheses? How did the formalization of statistics change this?

I suppose I could try to locate original papers and see what some famous scientists did...perhaps I will if no such source exists.

 

[mfb]

I suppose I could try to locate original papers and see what some famous scientists did...perhaps I will if no such source exists.

I don't know about historic papers (although I never saw very detailed data analysis there), but the statistical methods used are still improving. If you compare papers from 1980 with modern papers in particle physics, for example, there is a huge difference. Many things that were neglected back then are taken into account now. And the increasing complexity of the detectors and physics models does not make that easier.