The case that religion alone can account for radicalism in the later eighteenth century has been persuasively argued and has received some empirical support. We test it using data from Bristol during the American crisis. Literature from the 1774 election shows that contemporaries assumed commonality of interest of poor voters with other poor voters and with the anti-establishment/radical candidates, and commonality of interest of richer voters with other rich voters and with the establishment candidates. To a lesser extent voters were encouraged to vote as loyal Anglicans or religious Dissenters. To see how this rhetoric translated into action, a poll book (1774), a city directory (1775), and rate and tax returns (1773–7) were combined by nominal record linkage. Using the linked data we found that poorer Bristolians voted preferentially for anti-ministerial candidates and particularly for the radical candidate, while richer Bristolians tended to support the ministerial candidates. The divisions were clear when voting was compared with wealth (using fiscal data), and when compared with occupation provided an appropriate classification was used. To see if radical action was sustained in Bristol, loyal addresses and petitions for conciliation of the Americans were examined. The rhetoric surrounding petitioning/addressing again assumed commonality of interest based on socio-economic circumstance. This carried over into action: petitioners for conciliation tended to be poorer people, whilst addressers, who favoured coercion, tended to be richer. The study refutes the thesis that religion alone accounts for radical behaviour, and suggests the need to examine the totality of people's experience before their political behaviour can be fully understood. [ABSTRACT FROM PUBLISHER]