In research design, especially in psychology, social sciences, life sciences, and physics, operationalization is a process of defining the measurement of a phenomenon that is not directly measurable, though its existence is indicated by other phenomena. It is the process of defining a fuzzy concept so as to make the theoretical concept clearly distinguishable or measurable, and to understand it in terms of empirical observations. In a wider sense, it refers to the process of specifying the extension of a concept—describing what is and is not a part of that concept. For example, in medicine, the phenomenon of health might be operationalized by one or more indicators like body mass index or tobacco smoking. Thus, some phenomena are difficult to observe directly (i.e. they are latent), but their existence can be inferred by means of their observable effects. Sometimes, when multiple alternative or even competing, operationalization for the same phenomenon are available, one can repeat the analysis with all of the operationalizations one after the other, to see if the results are impacted by different operationalizations. This is often called conducting a robustness check. If the results are (substantially) unchanged, the results are said to be robust against certain alternative operationalizations of the checked variables. The concept of operationalization was first presented by the British physicist N. R. Campbell in his 'Physics: The Elements' (Cambridge, 1920). This concept next spread to humanities and social sciences. It remains in use in physics.