Computer simulation in data analysis: A case study from particle physics.

The paper presents a case study of the data analysis in the CDHS scattering experiment of particle physics performed in 1983. The case study compares the function of computer simulation in the data analysis with recent philosophical work on the role of simulations in high energy physics (HEP) and the theory-ladenness of the data. In the data analysis of CDHS, computer simulations entered an iterative process of probabilistic data correction. The computer simulation was a crucial ingredient of the data analysis that served to increase the accuracy of the measurement. The way in which simulation was used corresponds in a certain sense to the function of "models as mediators" (Morgan and Morrison), by mediating knowledge about measurement errors and the way of correcting them. I argue that this use of simulation did not give rise to a vicious circle of adjusting data to theory and vice versa but only to a weak, or benign, theory-ladenness of the data compatible with scientific realism. In the publication of the CDHS results, the measurement outcomes are called "observed data", indicating a realist attitude of the physicists towards the measured quantities which does not exactly fit in with entity realism or theory realism.