Pre‐election perceptions of politicians and their promises as a function of the reference group match between speaker and listener

The Categorization‐Attribution‐Matching or CAM model of persuasion was tested in a field experiment, using an actual election to provide a natural manipulation of reference group similarity‐dissimilarity between speaker and audience. Just prior to the election, 344 Ss rated a set of campaign proposals, which were attributed either to the incumbent Prime Minister of Australia, or to the Leader of the Opposition. Half of the Ss received an unsupported, and half received a supported version of the message, thereby manipulating information content. Ss also rated the speaker, and indicated how they intended to vote. As predicted, speakers and speeches were rated as more credible and persuasive respectively, in the conditions where the political affiliation between source and target was matched, than in the mismatched conditions. And as predicted, the provision of additional information did not make a speaker more credible or a message more persuasive, when reference group salience was high.