This article employs an interpretive approach, and in the light of contributions to this symposium by Butler and McAllister, and McLean et al., holds that metrics of research ‘quality’ are socially constructed and hence are as ‘subjective’ as peer review. Thus it rejects the use of stand-alone metrics as an ‘objective’ basis to inform funding allocations. Rather, the optimum method of ‘quality’ assessment is a panel-based exercise with expert judgement informed by a range of discipline-sensitive metrics and peer review of publications. The article maintains that the politics of metrics of political science conceals interests about the foundations of social scientific knowledge, and so the dispute over metrics and peer review is a metaphor for the conflicting epistemological preferences of UK political scientists. It is also argued that metrics-led assessment subjects political science to ‘Gradgrinding’ on two fronts: that political science departments amount to less than the sum of their parts, and the audit culture strips the discipline of its humanism.