In this paper, we put social measurement on trial: providing two perspectives arguing why measurement in the social and in the physical sciences are incompatible and counter with two perspectives supporting compatibility. For the case 'against', we first argue that there is a lack of definition in the social sciences. Thus, while measurement in the physical sciences is supported by empirical evidence, calibrated instruments, and predictive theory that work together to test the quantitative nature of properties, measurement in the social sciences, in the main, rests on a vague, discretionary definition of measurement that places hardly any restrictions on empirical data, does not require calibrated instruments, and rarely articulates predictive theories. The second argument for the case 'against' introduces the problem associated with psychometrics, including different approaches, methodologies, criteria for success and failure, and considerations as to what counts as measurement. Making the first case 'for', we highlight practical principles for improved social measurement including units, laws, theory, and metrology. The second argument 'for' introduces the exemplar of the Lexile Framework for reading that exploits metrological principles and parallels the paths taken by, for example, thermometry. We conclude by proposing a way forward potentially applicable to both physical and social measurement, in which inferences are modelled in terms of a measurement system, where specifically the output of the instrument in response to probing the object ('entity') is a performance metric, i.e. how well the set-up performs the assessment.