The influence of misleading images on misinformation beliefs: probative value, repeated exposure, cognitive thinking styles, and political identity

The prevalence of misleading images on social media represents an increasingly serious problem of misinformation online, yet we have limited knowledge about how photos affect the perceived veracity of accompanying claims. This thesis investigated the impact of accompanying photos (nonprobative, mislabelled, and manipulated) on the credibility of claims, compared to text-only versions, across eight studies. We also examined the role of repeated photo exposure in shaping perceptions of truthfulness, and the mechanisms underlying these potential effects by considering participants’ cognitive thinking styles (deliberative and intuitive) and socio-political biases. To explore this, we used the Cognitive Reflection Test (Frederick, 2005) and Faith in Intuition scale (Epstein & Pacini, 1999) to examine the relationship between participants’ cognitive thinking style, partisanship, photo bias, and photo familiarity effects. Chapter 2 explored the impact of nonprobative and probative (specifically mislabelled) photos on perceptions of truthfulness for false claims. Nonprobative photos are only thematically related to the claim (i.e., providing no evidence in support of the claim), while mislabelled photos (i.e., genuine photos taken from one context and deceptively applied to another) misleadingly appear to depict the described event. Our findings showed that mislabelled photos increased belief in false claims over text-only claims (i.e., fabricated photo evidence effect), especially among individuals who tend to apply less cognitive effort when making judgements. Nonprobative photos did not enhance belief in either true or false news claims compared to text-only versions, thus we did not observe a ‘truthiness effect’ (Newman et al., 2012) for news claims. This suggests that the content of the image, rather than the ease of processing, plays a primary role in influencing judgements and biasing belief of associated claims. Chapter 3 investigated how cognitive thinking style may relate to belief in manipulated photos and false claims, and the effects of prior exposure to these photos on perceptions of truthfulness of novel claims. The illusory truth effect refers to the tendency for familiar stimuli to enhance processing fluency and create a sense of associated truthfulness (Hasher et al., 1977). Findings demonstrated a brief exposure to a photo was sufficient to produce illusory truth effects, even when the photo was manipulated, and its inauthenticity was evident. Susceptibility to believing false claims accompanied by manipulated photos was found to be, in part, due to a person’s failure to apply more critical thought. However, cognitive thinking style did not impact illusory truth effects. Chapter 4 examined the effects of genuine and manipulated partisan photos on misinformation belief, testing predictions drawn from two theories: Classical Reasoning and Motivated System 2 Reasoning. The Classical Reasoning account posits analytical thinking helps detect and reject misinformation (Pennycook & Rand, 2019a), while the MS2R account argues that analytical thinking can, counterintuitively, increase acceptance of politically aligned misinformation due to rationalisation based on one's political identity (Kahan, 2013). Results showed higher analytical thinkers (among Republican and Democrat supporters) were more competent in discerning between true and false claims. However, analytical thinking also appeared to amplify polarised responding, as they were more likely to believe a false politically?congruent claim over a true incongruent claim – demonstrating potential identity-protective processes at play. These findings highlight the potential influence of effortful thinking and partisan motivation in assessing political claims and photos, contributing to the ongoing discussion surrounding the reasons why individuals may accept certain claims while dismissing others. Overall, our findings offer insights into how visual misinformation may add to the development of false beliefs, and how prior beliefs and individual differences in deliberative and intuitive thinking styles may increase vulnerabilities of certain individuals to such effects. Through this investigation, we have deepened our understanding of the effects of perceptual cues and cognitive processes on belief formation, contributing to a broader understanding of how false information may spread and informing the development of evidence-based strategies to counteract misinformation. In sum, we present new evidence that lays a strong foundation for further exploration of misleading images within the context of visual misinformation – an area poised to gain significance given the growing prevalence of AI-generated images on the internet and across social media networks.