How strong is the evidence for accelerated expansion?

Abstract

We test the present expansion of the universe using supernova type Ia data without making any assumptions about the matter and energy content of the universe or about the parameterization of the deceleration parameter. We assume the cosmological principle to apply in a strict sense. The result strongly depends on the data set, the light-curve fitting method and the calibration of the absolute magnitude used for the test, indicating strong systematic errors. Nevertheless, in a spatially flat universe there is at least a 5 sigma evidence for acceleration which drops to 1.8 sigma in an open universe.