The development of mathematical models that can reliably simulate the energy performance of a whole building or a building component with minimal discrepancy between the real and simulated data is a major aim of Building Physics science. In order to create models that accurately represent real physical phenomena it is necessity to perform tests on buildings and building components, producing real data that can be used to adjust and validate these models. If these tests are not undertaken correctly, incorrect data sets, insufficient data sets or excessively complex and expensive experiments may be performed. Thus, depending on the aim and the accuracy needed for the mathematical models, the test environment and test set up must be chosen correctly. This problem has been studied inside Subtask 2 of the Annex58 “Reliable building energy performance characterisation based on full scale dynamic measurements”. The aim was to come to a roadmap on how to measure the actual thermal performance of building components and whole buildings. This means under realistic boundary conditions (field exposure or artificial climate) and taking into account workmanship. Since there are many established methods and different Standards for different measurement purposes, the solution has been to organize the existing methods (both Standards and widely used non-Standard testing methods) into a decision tree. This decision tree begins with the question “What do you want to characterize?” and determines the context, environment, experimental design and analysis method being used by the user, terminating in a document reference. In a very simple format, following the decision tree and having a clear idea of what you need to characterize or model, you will reach an end branch of the decision tree where a testing Standard or testing method will be defined. The objective of this paper is to present the decision tree, its logic and the way it should be used.
ORCID: 
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)