False or misleading information significantly threatens democracy by disrupting informed decision-making and diminishing public trust in democratic institutions. The spread of mis/disinformation is, therefore, a cybersecurity concern. Technical methods, such as AI, used to detect mis/disinformation automatically are often inadequate because they fail to consider psychological theories that may help inform the models. This research addressed this shortcoming by examining the persuasive cues evident in fake news compared with genuine news. It applied the Elaboration Likelihood Model, a Dual Process Theory, to examine distinguishable cues in posts regarding the Indigenous Voice to Parliament: 200 fake news posts and 200 genuine news posts. As predicted, fake news stories were more likely to contain the following cues: emotional appeals, repetition, celebrity endorsements, and loudness cues. However, contrary to predictions, fake news stories contained more statistics compared to genuine news. As anticipated, genuine news stories were more likely to feature rational appeals. However, contrary to hypotheses, genuine stories included more visual cues than fake news. It is concluded that both types of news attempted to persuade but in different ways. The findings, especially the significant use of loudness cues, could inform the development of LLMs (large language models) to detect fake news automatically.