We did not write 17025 as an accreditation standard. We wrote it as a laboratory QA/QC standard against which labs could establish their own competence. Accreditation against ISO/IEC 17025 is a large, but secondary outcome of that effort.
Although implementation of 17025 may not have the same impact on the advancement of science that pure R&D does, it has, over the last 14 years, established a consistency of measurement that has allowed laboratories to improve the consistency and validity of their measurements and this has assisted in the advancement of measurement techniques - therefore advancement of science.
Hard to prove? Maybe. Maybe not. Implementation of 17025 in environmental laboratories in Canada has been tracked with examination of its effect on their ability to produce technically valid results (as described in the Introduction to 17025) since 1997. Have a look at
This last data set has been continuously collected since 2003. Enhanced trustworthiness of results has to have some impact on the development of measurement techniques. Advancing the science too? I think so, but some may disagree.
Results that are trustworthy help legislators develop better policy and assist in the enforcement of appropriate regulations. Anyone seeking information regarding the impact of 17025 on the consistency or validity of laboratory test results should also consider this data.