Melbourne School of Psychological Sciences - Theses
Permanent URI for this collection
Search Results
1 - 10 of 345
-
ItemThree Essays on the Psychological Micro-Foundations of Foreign Policy Attitudes( 2023-06)This dissertation investigates the impact of deep-seated psychological constructs on individuals’ attitudes towards foreign policy. It consists of three research streams exploring personality traits, social dominance orientation (SDO), and disgust sensitivity. The findings provide valuable insights into the micro-foundations of foreign policy attitudes and their psychological underpinnings. The first research stream focuses on personality traits, examining their lower-level aspects/facets and their connection to foreign policy attitudes. Cooperative Internationalism (CI) is primarily influenced by openness, orderliness, enthusiasm, and compassion, with volatility acting as a limitation. Militant Internationalism (MI) is rooted in assertiveness and volatility, while politeness constrains MI attitudes. Isolationism (ISO) is negatively associated with Agreeableness and positively with Neuroticism, showing resistance to isolationism rooted in personality aspects favouring social engagement. Global Justice (GJ) shares a personality profile with CI but exhibits a negative effect of politeness due to its association with traditional moral values. The second research stream investigates SDO and its influence on foreign policy attitudes. Social Dominance Theory aligns with the findings, as the dominance dimension (SDO-D) underlies attitudes favouring the use of force abroad. It motivates dominant and aggressive foreign policies, but surprisingly, it is also associated with cooperative and redistributive stances. SDO-D predicts isolationism, highlighting its role in preserving power. The anti-egalitarian dimension (SDO-E) drives cooperative foreign policy attitudes and opposes redistributive strategies. The third research stream examines the influence of disgust sensitivity on foreign policy attitudes. Pathogen disgust sensitivity does not uniquely predict conservative foreign policy attitudes, contrary to prevailing notions. Sexual disgust, however, is associated with militant attitudes, potentially driven by a desire to protect societal norms. Individuals with high disgust sensitivity may perceive cooperative attitudes as risky, but the benefits of cooperation often outweigh their concerns. Overall, this dissertation contributes to understanding the complex interplay between deep-seated psychological constructs and foreign policy attitudes. Personality traits, SDO, and disgust sensitivity offer valuable insights into the formation of attitudes towards world politics. We provided the first integrative and systematic exploration into the relationships between psychological features underlying and foreign policy attitudes.
-
ItemThe fusion-secure base hypothesis( 2023-07)Strong group commitments – such as identity fusion, a powerful form of group alignment – have historically been considered reliable predictors of intergroup violence. However, these aggressive outcomes are largely conditional on outgroup threat perception, with the effect of identity fusion on benign intergroup relationships underexplored. This dissertation outlines the fusion-secure base hypothesis, which leverages attachment theory, social identity theory, and identity fusion theory to argue that strong group commitments may engender trusting and cooperative intergroup relationships with non-threatening outgroups. However, when an outgroup is perceived as a threat, identity fusion will predict the violent outcomes for which it is well known. Chapter I introduces the theory via an historical example, the 1914 Christmas Truce, and presents the general structure of the dissertation. Chapter II outlines the theory in in detail, considers existing evidence, and dictates a research agenda, which is then pursued in Chapters III & IV. Chapter III describes a series of empirical studies that test the association between identity fusion and a willingness to trust and interact with outgroup members, and whether perceptions of a secure base mediate the relationship, with an internal meta-analysis supporting the unique suitability of fusion to engender positive intergroup relations. Chapter IV tests perceptions of outgroup threat as a moderator of the fusion-secure base hypothesis. This chapter explores intergroup relations in a complex and oft-violent social ecosystem, the Bangsamoro Autonomous Region in Muslim Mindanao (BARMM) region of the Philippines, by examining numerous intergroup dyadic relationships. Finally, Chapter V reflects on future research directions and considers the practical and theoretical implications of the theory. In summary, this dissertation proposes and explores a novel theory, the fusion-secure base hypothesis, and presents considerable theoretical and empirical evidence in its favour. This suggests that strong group commitments can help or hinder intergroup relations, depending on the context, and may represent both a cause and cure of sectarian violence.
-
ItemThe Mental Health-Related Barriers and Benefits to Exercise( 2023-05)Leisure-time exercise is a well-established and highly recommended treatment for symptoms of anxiety and depression, however, such symptoms can act as a barrier toward leisure-time exercise engagement. Thus, this thesis aimed to examine how symptoms of anxiety or depression and leisure-time exercise may interact, using quantitative methods to explore potential predictors and correlates of mental health-related barriers and benefits to leisure-time exercise. As a necessary first step toward the quantification of mental health-related barriers and benefits to exercise, this thesis includes the development and psychometric validation of the Mental health-related barriers and benefits to EXercise (MEX) scale in healthy adults, as reported in Chapter 1. Further psychometric validation of the MEX was achieved in Chapter 2, where the mental health-related barriers and benefits to exercise were shown as divergent factors from more ‘generalised’ exercise barriers and benefits items in an exploratory factor analysis (e.g., physical, lifestyle, or logistical-related barriers and benefits to exercise). Chapter 2 additionally explored affective distress profiles within the general population, and examined these profiles on differences in levels of both generalised and mental health-related barriers and benefits to exercise. Results showed that groups with higher levels of psychopathology had higher levels of both mental health-related barriers to exercise and generalised exercise barriers, and lower levels of physical activity and of generalised exercise benefits. However, the affective groups did not differ in levels of mental health-related benefits to exercise. Chapter 3 examined differences in mental health-related barriers and benefits to exercise in a sample of adults with chronic pain, compared to a demographically matched sample of adults without chronic pain. The sample of adults with chronic pain reported significantly higher mental health-related barriers to exercise and anxiety, depression, and stress-like symptoms than adults without chronic pain. Additionally, individuals with chronic pain also reported lower leisure-time exercise engagement than those without chronic pain. There was no difference in levels of reported mental health-related benefits to exercise between the groups. However, a multiple regression showed that mental-health related benefits were less predictive of leisure-time exercise in those with chronic pain relative to those without. Meanwhile, the mental health-related barriers to exercise negatively predicted exercise engagement with a moderate effect size for both groups. Taken together, this thesis provides critical insight into the relationships between mental health-related barriers and benefits to exercise and affective symptoms, and between the barriers and benefits and leisure-time exercise engagement in both the general population and in adults with chronic pain. Overall, the results of this thesis suggest that the measurement of mental health-related barriers and benefits to exercise may provide important information in personalised exercise prescription for adults with mild-to-moderate affective distress in the general population, and for adults with chronic pain.
-
ItemPredicting and improving the use of CPAP therapy in individuals with obstructive sleep apnoea( 2023-02)Obstructive sleep apnoea (OSA) is a common chronic sleep disorder characterised by the repetitive narrowing or collapse of the upper airway during sleep. Consequences of untreated OSA include fragmented sleep, intermittent hypoxia, memory deficits, increased risk of cardiovascular events and workplace accidents, and changes in mood, particularly depression. Comorbid depression is highly prevalent in OSA samples, and it is difficult to diagnose one condition in the presence of the other due to symptom overlap. Continuous positive airway pressure (CPAP) is the treatment of choice for moderate to severe OSA. However, CPAP use is suboptimal in many individuals. As such, there remains a significant burden of OSA on the individual and the healthcare system. The work presented in this thesis was part of a larger randomised controlled trial, the CPAP for OSA and Depression (COSAD) trial. The three experimental studies reported herein (Chapters 3-5) stemmed from the COSAD study design. Patients recently diagnosed with OSA and referred to the Austin Health sleep laboratory for CPAP implementation were recruited for this trial. Study 1 examined predictors of CPAP use up to 4 months post CPAP initiation. Predictors included patient and disease characteristics, sex, age, BMI, OSA severity, psychological variables (anxiety and depression), behaviour change variables (decisional balance and processes of change, and self-efficacy, which is comprised of 3 constructs; risk perception, outcome expectancies and CPAP self-efficacy), the first week of CPAP use, and wait time to CPAP initiation. A delay of 95 days to CPAP initiation was not associated with any differences in CPAP use at 1 week, 1 month and 4 months. Female sex, processes of change and self-efficacy were associated with CPAP use at 1 week. Average nightly CPAP use at 1 week was associated with CPAP use at 1 month and 4 months. No other patient or disease characteristics or psychological variables were associated with CPAP use up to 4 months. Study 2 investigated the effect of treating OSA with CPAP therapy on mood and sleepiness. Mood, including depression, anxiety and stress, were measured at baseline and 4 months and were compared between an intervention group (participants underwent 4 months of CPAP therapy for OSA) and a waitlist control group (participants were not treated for OSA). Stress improved in females with 4 months of CPAP therapy compared to females in the control group. Sleepiness, but not mood, improved after 4 months of CPAP therapy compared to control. Study 3 investigated the effect of a multi-dimensional intervention on CPAP use and self-efficacy compared to treatment as usual. CPAP use and self-efficacy improved in the intervention group compared to treatment as usual across the 4-month trial. These findings support previous work that early CPAP use predicts subsequent CPAP use, that CPAP use improves daytime sleepiness and that an intervention can increase CPAP use. Furthermore, this work has advanced the field by demonstrating that a delay to CPAP initiation may not effect subsequent CPAP use. Additionally, there may be sex differences in stress symptoms when treating OSA with CPAP therapy. Implications for future studies include the importance of the early period of CPAP use and the effect of wait time to therapy initiation on subsequent CPAP use.
-
ItemExploring the associations of diet quality with mental and gut health in the context of the mediating potential of the gut microbiota in adult women( 2023-03)Background: A rapidly growing evidence base indicates that habitual dietary quality is associated with the common mental disorders. At the same time, the relationships between diet and mental health may be influenced by the co-occurrence of highly prevalent gastrointestinal conditions, such as irritable bowel syndrome (IBS), whose current first-line treatments include dietary alterations which may impact individuals’ dietary habits. It is also increasingly recognised that the gut microbiota (i.e., the community of microorganisms inhabiting the gastrointestinal tract) may be involved in the pathophysiology of internalising disorders and IBS. As the gut microbiota is heavily influenced by habitual diet, a theoretical framework has been proposed whereby the gut microbiota mediates the associations between diet quality and anxiety/depression psychopathology. Accordingly, this thesis conducted a detailed cross-sectional examination of the associations of several diet indices in adult females with clinically diagnosed anxiety and/or depressive disorders (both at the disorder and symptom level) while considering IBS comorbidity (Aim 1). Cross-sectional associations between diet indices and the gut microbiota were also investigated, with consideration of the moderating influence of IBS and internalising disorders (Aim 2). Finally, the proposed theoretical mediation model was examined to assess the relationships between diet indices and anxiety/depression outcomes, and whether this was mediated by the gut microbiota (Aim 3). Methods: This thesis used data from the Bugs and Brains study cohort, a mixed-methods study of 158 adult females with anxiety/depression, IBS, comorbidity, or with no history of any of these conditions. Participants were recruited based on strict inclusion/exclusion criteria related to medical/psychiatric comorbidities, medication use, and lifestyle behaviours (e.g., drinking and/or substance use). Participants attended a semi-structured psychiatric interview, provided a stool sample for 16S rRNA gene sequencing, had their anthropometrics measured, and completed validated questionnaires assessing demographic, lifestyle, diet, psychiatric, and health related factors. Dietary quality was assessed using two measures of whole-of-diet intake (i.e., the ModiMedDiet [MMD] and the Healthy Eating Index for Australian Adults-2013 [HEIFA-2013]) and a measure of the proportion of ultra-processed foods (UPFs) consumed per day. Measures of gut microbiota diversity and community composition (i.e., alpha and beta diversity, respectively) were estimated for all participants. Data analysis was performed using R (v4.2.0), combining a series of multivariate logistic/linear regression analyses and mediation analyses using the PROCESS model. Results: In the first empirical chapter (Aim 1), logistic regressions revealed that all three diet indices predicted odds for an internalising disorder, with the proportion of UPFs most strongly associated with higher odds for internalising disorders. Linear regressions also indicated that the HEIFA-2013 and UPF intake were marginally associated with lower and higher depression symptoms, respectively, while no influence due to IBS comorbidity was evident in any of the analyses. In the second empirical chapter (Aim 2), linear regressions indicated that the HEIFA-2013 was consistently directly associated with higher microbial diversity, and there was evidence for a direct association between the HEIFA-2013 and alpha diversity which was moderated by IBS diagnosis. Meanwhile, the MMD and HEIFA-2013, as well as IBS diagnosis, were directly associated with a gut microbiota composition characterised by more positive loadings of anti-inflammatory bacterial genera (e.g., Ruminococcaceae UCG-002, Christensenellaceae R-7 group, Coprococcus 2), and negative loadings of potentially pathogenic pro-inflammatory genera (e.g., Ruminococcus gnavus group, Flavonifractor, Eggerthella). Finally, in the third empirical chapter (Aim 3), mediation analyses indicated no evidence to suggest that the gut microbiota mediated the associations between diet quality and anxiety/depression outcomes. However, several noteworthy moderation findings were indicated in relation to alpha diversity and depression symptoms. Conclusion: The current thesis adds to the existing literature by revealing clear links between dietary quality and clinical internalising disorders in a rigorous cohort with strict inclusion/exclusion criteria that was independent of key confounding factors, such as age, BMI, and host sex. Findings highlighted that UPFs warrant particular attention in relation to increased odds for anxiety and/or depressive disorders in the present analyses, given the rapid increase and uptake of these foods globally. Furthermore, while the gut microbiota was closely linked with overall dietary quality, no evidence was found to support it as a mediator of the relationships between diet quality and mental health. Several reasons for these unexpected results are discussed, including the complex and multifactorial relationship between habitual diet and anxiety/depression as it relates to underlying biological mechanisms, as well as methodological and conceptual considerations in the trialogue of diet, the gut microbiota, and mental health. Due to the methodological rigour of the Bugs and Brains study, the current thesis was well placed to examine the cross-sectional relationships presented herein. However, further longitudinal research that combines taxonomic and functional outputs is needed to gain mechanistic understanding of the relationships between diet, internalising disorders, and the gut microbiota.
-
ItemSleep and PTSD: bi-directional relationship and underlying mechanism( 2023-04)Post-traumatic stress disorder (PTSD) is a debilitating and enduring disorder that a small but significant number of people develop following exposure to a traumatic event. A common feature of PTSD is disrupted sleep including insomnia and nightmares. Ongoing difficulties in sleeping prevent the sleep-dependent adaptive processing of traumatic memories. Further, sleep disruptions prevent recovery, and perpetuate the disorder when established. This thesis aimed to investigate the bi-directional association between changes in sleep and PTSD symptoms as well as between sleep and underlying fear memory processes. In part I, I examined the immediate effect of changes in sleep on fluctuations of PTSD symptoms and vice versa. Here I used both subjective and, for the first time, objective measures of sleep (study 1). This study found differences in the association between sleep and daytime PTSD symptoms depending on the sleep measurement method. Additionally, preliminary evidence suggests sex-specificity in the association between night-time sleep and daytime PTSD symptoms. Part II focused on the role of sleep in fear conditioning and extinction learning – the experimental model of PTSD development and treatment. Extinction learning and extinction recall (i.e., the ability to learn and remember that previously dangerous stimuli are not threatening anymore) are thought to be impaired in PTSD and impacted by sleep. Particularly rapid eye movement (REM) sleep has been suggested as the sleep stage most important for processing emotional memories. First, a systematic review summarized the available literature assessing the effect of REM and other sleep stages using a meta-analytic approach (study 2). The overwhelming majority of research highlighted the importance of REM sleep, but the meta-analysis did not find the expected REM sleep effect on extinction recall. Following this and the finding from part I, the third and last study investigated the effect of subjectively reported sleep on fear conditioning and extinction learning (study 3). Again, no effect of subjective sleep indices on extinction recall was found. However, both studies found that there were differences between sexes and outcomes were dependent on the diagnostic status (study 2) or rather PTSD severity (study 3). Together, this thesis filled important gaps in the literature and highlighted that sleep is important in the expression of PTSD symptoms in day-to-day life as well as in the mechanism underling PTSD such as fear conditioning and extinction learning. However, the effect of sleep may vary depending on key associated factors. Specifically, the relationship may differ depending on the sleep measurement method (objective and subjective), the samples studied (clinical and control populations) and sex (men and women).
-
ItemDeconstructing discourse: the window to the temporal lobe epilepsy patient( 2023-04)This thesis aims to understand high-level language dysfunction among patients with temporal lobe epilepsy (TLE). Disturbances of language in temporal lobe epilepsy are commonly identified on the basis of anecdotal clinical reports, proffered as a cognitive complaint, and are identifiable on formal neuropsychological assessment at a single-word level. Despite this, there is a limited understanding of how this relates to dysfunction at the level of discourse in this patient population, where language is characteristically verbose or ‘circumstantial’. The nature of the neurolinguistic impairment in TLE has not yet been fully characterised. This thesis therefore intends to elucidate the micro- and macrolinguistic basis of circumstantiality in TLE by describing the nature of their discourse in diverse linguistic contexts. Examining the field through the lens of psycholinguistic models of language production, neuropsychological function, and neuroanatomical frameworks, this thesis aims to develop a deep clinical understanding of language in these patients. It seeks to devise a theoretical model of the mechanisms underlying high-level language dysfunction. Language samples were audio-recorded, transcribed verbatim, and coded for key discourse features in order to conduct a detailed multi-level discourse analysis. The first study (Chapter 7) investigated the influence of elicitation structure on language output. This involved eliciting monologic output in two contrasting conditions. One was highly structured and constrained, relating to the description of the Cookie Theft cartoon, while the other was minimally structured and unconstrained, asking participants to provide a spontaneous description of a typical day in their lives. A constrained context gave rise to microlinguistic disturbances in individuals with TLE, reducing fluency, with more pauses and fillers. Under an unconstrained context, as anticipated, classical aspects of verbosity emerge in those with TLE, manifesting as longer speaking time, a longer duration of pauses, and a higher proportion of repetitive or redundant statements. Macrolinguistic elements such as coherence and informativeness are widely impacted, particularly disturbing language formulation. Discourse in unconstrained conditions aligns closely with the clinician’s anecdotal reporting of verbosity in TLE which often does not map onto psychometric assessment. This speaks to the demands of naturalistic language contexts where patients are asked open-ended questions. I conclude that these patterns suggest that different discourse impairments emerge in TLE as a function of differing linguistic challenges and reflect a dynamic linguistic system taking shape under specific contextual conditions. The second study (Chapter 8) aimed to replicate and extend the seminal work of Field and colleagues (2000) regarding narrative discourse production in TLE. Participants were asked to tell a cartoon story on five immediately consecutive occasions. When compared to healthy controls who produce increasingly concise and coherent versions of the narrative across repetitions, individuals with TLE fail to do so. The narratives produced by individuals with TLE are less informative overall. They produce fewer novel units and introduce more content that is repetitive, extraneous, and does not progress the narrative. Their narratives are ultimately less fluent, less cohesive, and less coherent relative to healthy controls. There is a failure to compress discourse over repetitions. These findings suggest that individuals with TLE do not benefit from repeated engagement with a narrative in the same way that neurologically normal individuals do. I conclude that disturbances to social cognition and ultimately pragmatics in TLE might underpin inefficiencies in their communication. The study presented in Chapter 9 is the first to consider how individuals with TLE produce discourse for personal reminiscence. This study involved eliciting a personal memory on four immediately consecutive occasions. The uniqueness of autobiographical memory systems make them a particularly interesting area of research for discourse production. Personal memories are highly specific and localised in time. It is presumed that the nature of autobiographical memory will produce a different psycholinguistic phenotype in TLE. As anticipated, a distinct pattern of dysfunction emerged in TLE when compared with healthy controls. Healthy controls did not compress their output volume across repetitions and instead produced more novelty and a more coherent and refined account over time. Among individuals with TLE, discourse disturbances in the retelling of personal memories suggested an issue of quality not quantity. Under these circumstances, they tended to consistently tell a vague story across repetitions, with disturbances to fluency, cohesion, and coherence. This reflects a reduced capacity to refine a coherent mental representation which likely relates to impoverished autobiographical memory traces and the cognitive demands of recalling these specific personal events. Traditionally, language has been examined in an overly simplified and insensitive manner via measures of lexical retrieval, or based on language assessments specifically designed for evaluating aphasia. Neither of these avenues are applied appropriately when it comes to understanding high-level language impairments which are more subtle in their presentation. The detailed multi-level discourse analysis procedures applied presently provide a considerably more nuanced understanding of language in TLE than is possible with standardised assessments which focus predominantly on the single-word level. The thesis is that the mechanisms of circumstantiality and discourse compression are context-dependent. Circumstantiality is subserved by capacity limitations and semantic processing impairments across contexts, and amplified by pragmatic disturbances and deprioritisation of frontal networks when the content is personal. The research reported here highlights the importance of examining and understanding language in naturalistic, unconstrained contexts and provides a unique opportunity to examine the profound and pervasive impact of TLE on discourse which impacts daily communication, social and occupational functioning, and penetrates multiple levels of neurolinguistic function. These findings speak to the importance of revisiting historical notions via novel approaches that offer alternative perspectives, or windows, into patients with TLE.
-
ItemMetacognition in decision making: Exploring age-related changes in confidence( 2022-11)Metacognition is a fundamental human function that supports goal-directed behaviour. By constantly monitoring and evaluating our decisions we are able to detect errors when they occur and adjust the behaviour accordingly. Metacognitive evaluations can be expressed in ratings of decision confidence or error detection reports. Humans are generally capable of forming well-calibrated estimates of their own performance, yet metacognitive abilities have been shown to be specifically affected by healthy ageing. However, the mechanisms underlying this decline remain poorly understood. This thesis aims to investigate the cognitive processes of age-related changes in perceptual metacognitive performance by combining approaches from the fields of error monitoring and decision confidence. For this, we developed a new paradigm for studying the metacognitive evaluation of errors and correct responses that was feasible for adults of all ages. While recording an electroencephalogram (EEG) and response force, a sample of 65 healthy adults from 20 to 76 years made a series of decisions in a modified version of the Flanker task and subsequently indicated how confident they felt about their decision on a four-point scale. Across two studies, conducted in the same large sample, I addressed three specific research questions: first, how is metacognitive performance affected by healthy ageing? Second, what are factors contributing to the observed decline in metacognitive performance? And third, how does an age-related decline in metacognitive performance affect subsequent behaviour? The analysis of behavioural data (Study 1a) showed that metacognitive accuracy declined significantly with older age and that this decline could not be explained by the decline in task performance alone. Independent of age, however, participants adjusted their performance according to their metacognitive evaluation of their previous decision and responded more cautiously after reporting low confidence. The analysis of electrophysiological data (Study 1b) focussed on the modulation of two correlates of error monitoring by confidence and age. The results indicated that the error/correct positivity (Pe/c), a component discussed as a marker of error detection and decision confidence, scaled with reported confidence in errors but did not show the expected modulation by age. The amplitude of the error/correct negativity (Ne/c), a marker of early error monitoring processes, also scaled with reported confidence in errors, but in contrast, was less sensitive to variations in confidence with older age. Finally, Study 2 investigated the effect of age on the relationship between confidence and two response parameters of the initial decision: response time and response force. We replicated a widely reported negative relationship between confidence and response time. Importantly, we showed, for the first time, that confidence was also negatively related to fine-grained changes in peak force, which was intuitively exerted by the participants. Notably, these associations were dependent on the accuracy of the response and changed markedly across age: the relationship between confidence and response time was only found in correct responses and was pronounced with older age, while the relationship between confidence and peak force was only found in errors and only in younger adults. Overall, these findings jointly provide novel insights deepening our understanding of the observed decline in metacognitive performance with older age. A similar modulation of the Pe/c by confidence across the lifespan suggests that the post-decisional process of accumulating evidence about the correctness of a prior decision might generally be intact until old age. Instead, the age-related decline in metacognitive accuracy appears to be related to a multitude of cognitive and neural changes, which might reflect increased noise and hence higher uncertainty in older adults’ computation of confidence. Moreover, I discuss how a metacognitive decline could manifest in real life and how recent findings offer a promising view regarding the effect of training on metacognitive performance.
-
ItemCognitive Mechanisms of Information-Seeking in Humans( 2023-02)To inform decision-making, we are often required to actively sample information from our environment. In an information-rich environment, this necessarily involves selectively sampling some information signals while ignoring others. How we choose which signals to sample and which to ignore, as well as the downstream consequences these choices have on our beliefs and actions, are key areas of interest in the field of information-seeking. The present thesis aimed to examine how the characteristics of information signals and the contexts they inhabit affect information-seeking behaviour in humans. Study 1 combined a novel information-seeking paradigm with hierarchical Bayesian computational modelling to examine how the expected accuracy and valence of an information source affect our information-seeking behaviour. In Experiment 1, participants earned money from random, binary lotteries. They were not shown the outcomes of the lotteries, but instead could view a prediction of the outcome made by one of two available information sources. Before choosing which source to view the prediction from, they were first shown a series of example predictions made by each. The sources varied systematically in how accurate and how positive their predictions were. It was found that accuracy was the most important determinant of participants’ choices, but a substantial proportion of our participants also chose to view predictions from positively-biased information sources, even when they were objectively less accurate. The study further showed that participants who sampled positively-biased information tended to form positively-biased beliefs, and that these beliefs were held in high confidence. Experiment 2 of this study built upon these findings to show that the decision to sample biased information was not attributable to a desire to resolve uncertainty and may instead be due to hedonic motives. Importantly, these findings show that the accuracy of an information source does not always take precedence over the anticipated valence of the information it reveals. Study 2 aimed to examine how the act of freely choosing a prospect affects the value we place on information pertaining to its outcome. In this task, participants were presented with three equivalent lotteries and asked to make an arbitrary choice amongst them. Their choice was either accepted or vetoed, in which case one of the two remaining lotteries was automatically selected instead. They then rated their belief in the likelihood of winning and probed their desire to know the outcome of the lottery. Results showed that participants rated their confidence in achieving a winning outcome higher for freely-chosen lotteries compared to externally imposed alternatives. Further, participants also showed greater curiosity in the lottery outcome, as they were willing to pay more to reveal the outcomes of freely-chosen lotteries. Computational modelling revealed that increases in curiosity were not attributable to the effect of choice on the perceived likelihood of lottery outcomes. Instead, freely-chosen lotteries were subject to an enhancement in the value of resolving uncertainty about their outcomes. These results demonstrate that choice enhances curiosity, with potential implications for motivation and learning. Finally, Study 3 investigated how the presence of explicitly unreliable and misleading information affects curiosity and beliefs. In a first set of three experiments, participants were presented with a series of information sources that predicted which of two colours would win an upcoming lottery. Each source was comprised of a predicted colour and a percentage reliability score. After viewing the predictions of these information sources, participants were offered the opportunity to pay a small amount of money to reveal the prediction of an additional information source. They then selected which colour they believed would win the lottery, with 15 cents awarded for each correct prediction. In the baseline experiment, all sources made reliably correct predictions (above 50\% reliability). In the second experiment, an additional source was added to each sequence that was entirely unreliable (50\% reliable). In the third experiment, one information source was inverted such that it now predicted the opposite colour but was reliably wrong (below 50\% reliable). Despite the sources in each experiment containing the same amount of information, participants' choices were less accurate when the information environment included either an unreliable or reliably wrong information source. More interestingly, despite having lower choice accuracy, participants in those experimental conditions were also less willing to pay to view the prediction of an additional information source. In a series of three follow-up experiments, participants were subject to the same experimental conditions, with two key changes. Instead of being offered the chance to sample additional information, participants were required to immediately choose which colour they believed would win. Following this, they rated their confidence in their choice. Confidence was significantly reduced when unreliable or reliably wrong information sources were present. These results suggest that the decrease in curiosity observed in the first three experiments was not due to an enhanced perception of their ability to correctly choose the winning colour. Instead, explicitly unreliable and misleading information might not only negatively affect belief formation but also reduce curiosity. Overall, these findings revealed substantial and consequential differences in the motives underlying information-seeking behaviour. These motives were similarly diverse, including both inherent characteristics of information signals and the contexts in which they were situated. Finally, a novel effect of misinformation was elucidated, finding that exposure to unreliable information not only reduced the accuracy of held beliefs, but also the desire to seek information.
-
ItemMetacognitive Decisions on Decision Accuracy: Confidence Judgment and Changes of Mind( 2022-12)Even in the absence of external feedback, humans are capable of subjectively estimating the accuracy of their own decisions, resulting in a sense of confidence that a decision is correct. While decision confidence has been proposed to be closely related to other metacognitive judgments, including error awareness (i.e., awareness that a decision error has occurred) and changes of mind (i.e., reversal of previously made decisions), their relationships so far remain unclear. The current project aimed to investigate how confidence could be related to metacognitive judgments from two perspectives. First, Studies 1 and 2 investigated how confidence and changes of mind were affected by changes in different stimulus properties, particularly absolute evidence strength. In a brightness judgment task, participants were presented with two flickering, grayscale squares and required to select the square that appeared brighter. After each trial, participants reported their subjective accuracy on a rating scale ranging from “surely incorrect” to “surely correct”. Results showed that with stronger absolute evidence (i.e., increased overall luminance across both squares), confidence was increased and the proportion of changes of mind trials was reduced. These consistent changes support the hypothesis that higher confidence could contribute to less frequent changes of mind. Second, Study 3 investigated the relationships between confidence and the event-related potential (ERP) components of the centro-parietal potential (CPP) and the error positivity (Pe), which have been respectively proposed to be indexes of pre- and post-decisional evidence accumulation processes. In the same brightness judgment task, it was found that the relationships between confidence and these two ERP components depended on decision accuracy: Confidence was positively related to CPP amplitudes in correct trials, but negatively related to Pe amplitudes in error trials. These findings suggest that confidence in correct and error decisions involves different pre- and post-decisional processes. Overall, the current findings suggest that (a) confidence could serve as a basis of changes of mind, and (b) confidence in correct and erroneous decisions was differentially related to pre- and post-decisional ERP indexes of evidence accumulation. Taken together, they suggest that confidence might emerge during decision formation and could, with the contribution from post-decisional processes, serve as a basis of changes of mind.