Most theories of democratic accountability presume that voters only evaluate governments over issues over which they bear responsibility. However, recent empirical work suggests that this characterisation is too simple, with many voters selectively attributing credit and blame to fit their partisan priors. Empirically adjudicating these mechanisms remains difficult, since both share observational equivalencies, and most existing studies focus on policies particularly prone to partisan bias, like the state of the economy.

To overcome these issues we study the early phase of England’s Covid-19 vaccination rollout, a programme which was extremely popular across the political spectrum and for which the government unambigiously received credit. Using a survey experiment embedded in Wave 21 of the British Election Study Online Panel, we test whether information emphasising the Government or the National Health Service (NHS)’s role in the programme affects how voters allocate responsibility and evaluate the Government.

Our findings are three-fold. Firstly, we show that our NHS treatment has a large negative effect on perceived government responsibility. This effect, in turn, mediates a decline in government approval. Secondly, unlike previous studies, we are able to demonstrate that swing voters are actually less responsive to new information when determining responsibility. Thirdly, we find that voters’ personal experience, such as being at greater medical risk from COVID-19 or having been offered the vaccine, also affect their receptiveness to our informational primes. Our paper demonstrates that, under the right conditions, voters can impartically update their responsibility perceptions, and that this has downstream effects on government approval.