Dear Sir, It is possible that the philosophy of consensus science may question, to some extent, the basis for both analytical chemistry and metrology [1]. Change of practices and procedures may be needed; genuine scientific results may be lost and theories that rely on data out of statistical control may incorrectly be proven true. There is an obligation of science to make fundamental changes, and consensus science may be a useful example of scientific methodology. There is no evidence in modern analytical chemistry for the idea that experts are able to distinguish between reliable and unreliable analytical results. The experts have no access to raw data of inter-laboratory comparisons (ILCs) and they have no access to raw data of scientific publications. Therefore, they are left with comparing highly processed data, which is unfortunate for comparisons in analytical sciences. Scientists are subjective when they propose a model and should remain objective when they compare theory to experiments. To the best of my knowledge, the staff never imposes contributions to the uncertainty. Laboratory staff prepare everything carefully and correctly in the laboratory, and even in cases where errors are committed, these errors are small and insignificant in the general picture of uncertainty. The main contribution to uncertainty is the apparatus long-time precision, which has been overlooked in uncertainty budgets [2]. If this contribution were included, would it increase the likelihood of correspondence between results in ILCs? Presumably, a university or a company would never employ incompetent staff to operate advanced apparatus. Accordingly, it is reasonable to assume that all results of ILCs belong to the same distribution. The analytical capability of laboratories cannot be checked unless the universal level of uncertainty of the involved method is known. It may safely be assumed that everyone who has invested in advanced scientific apparatus also has the competence to perform measurements. I agree that a two-step procedure for selection of qualified laboratories in advance of an ILC is not an option because that would be in conflict with consensus science. The idea of all laboratories showing equal degrees of competence is unattainable. Staff need acceptable working conditions and the results should be provided with a reasonable consumption of resources, labor, time etc. Nature does not care about ILCs and legislation. It cannot be decided by politics or debate whose fundamental constants in metrology or whose analytical results are right or wrong. We can only approach the true mechanisms of nature by performing measurements that possess a certain level of uncertainty. Hence, the method of consensus should be prevalent in science. Nothing can be decided by a single expert; results of many independent experts are required in order to maintain independence and credibility. It was interesting to me that the author of the ‘Letter to the Editor’ stated that outliers had been removed from the data set that constitutes the basis for certified reference materials (CRMs). Novel types of outliers, identified as ‘analytical outliers’, were invented for the occasion that supposedly should be distinguished from ‘statistical outliers’. It was not possible to find the technical reasons, as otherwise claimed possible. Hendrik Emons militates against better judgment: outliers were, in fact, rejected in BCR CRMs on the basis of statistical reasons and not due to technical reasons. Just a few examples: “The set of Lab 14 was rejected because its standard error exceeded s” [Bureau communautaire de reference, BCR Wholemeal flour CRM no. 189/Brown bread CRM no. 191] or “Laboratory 5 (TXRF) had an outlying variance for total Cr (Annex 1, Table 3a) and did not satisfy the criterion for This article is a response to the ‘Letter to the Editor’ to be found at DOI 10.1007/s00216‐013‐7141-5.