A researcher at Umeå University has introduced statistical methods that could help scientists decide how much confidence to place in their measurements. The work, by Mohammad Reza Seydi, tackles two basic questions: how much can a measurement be trusted, and how much data must be collected before conclusions are dependable?
Seydi developed and tested the methods specifically to give researchers better answers than existing approaches. By improving the way data are analysed, the methods are meant to reduce the risk that findings are built on shaky statistical ground.
The broader goal, according to the report from Phys.org, is more reliable research findings. The methods are not described as a single fix, but as a way to make statistical practice more robust.