The normative foundation (axioms) of Bayesian belief updating has long been established in the literature of decision science. However, psychology and experiments suggest that while rational decision making is ideal, it is rarely achievable for ordinary people. Therefore, it is important to explore the foundations and consequences of rational decision making within the field of economics. This thesis involves three papers on this. In the first paper, I explore the consequences of wishful thinking on mechanism design. It suggests that wishful thinking bias could be profit-generating for mechanism designers. In the second paper, I investigate conservative updating and provide a foundation for it. The main behavioral axiom, ``conservative consistency," suggests that decision-makers may partially incorporate information, particularly when it requires them to revise their previous preferences (the preferences order according to their prior belief). In the third paper, I reframe the model selection problem as a rational decision-making problem. The decision-maker is restricted to choosing an advisor to delegate their choices. I explore the conditions under which a rational decision-maker selects models (or advisors) according to Bayes factor criteria. / Doctor of Philosophy / Most of us do not always make decisions completely rational. This thesis digs into how irrational decision-making fits into economics, with three papers to break it down. The first paper looks at wishful thinking and how it affects our decisions. It suggests that if we understand our biases, we can design better mechanism to generate profit. The second paper talks about conservative updating, which is all about how we pick and choose what information matters, especially when it clashes with our existing belief. Lastly, the third paper explores how we choose advisors to help us make decisions. It looks at when it is smart to pick based on Bayes factor criteria. Through these papers, this thesis helps us understand how rational decision-making plays out in real-life economics.

Die Dissertation befasst sich mit Computationalen Mechanismen von Belief-Updating und Neuromelanin als Dopamin-Proxy bei Psychosen. Die Entstehung psychotischer Erfahrungen beruht möglicherweise auf einer veränderten Aktualisierung von Glaubenssätzen (Belief-Updating) durch die unausgewogene Gewichtung der Aktualisierungs-Signale anhand von Unsicherheit. Ziel von Studie 1 war, die verhaltensbezogenen und neuronalen Mechanismen von Belief-Updating bei Schizophrenie und speziell von Wahnvorstellungen zu untersuchen. Menschen mit Wahnvorstellungen zeigten stärkere neuronale Enkodierung der Aktualisierungs-Signale, während diagnostizierte Schizophrenie mit höherer Einschätzung der Umweltvolatilität und geringerer neuronalen Repräsentation derselben einherging. Studie 2 zeigte, dass Menschen mit subklinischen psychoseähnlichen Erfahrungen dazu tendierten, Veränderungspunkten der Umwelt weniger zu berücksichtigen. Ihre Beliefs waren weniger präzise, da sie eine geringere Anzahl voriger Beobachtungen berücksichtigten. Dies spricht für verändertes Belief-Updating als relevanten Mechanismus bei psychotischen Erfahrungen, speziell Wahnvorstellungen, über das Psychosespektrum hinweg ist. Belief-Updating Signale werden über dopaminerge Transmission vermittelt. Quantitative Dopamin-Proxys könnten zukünftig Behandlungsentscheidungen informieren. In Studie 3 führten wir eine Meta-Analyse durch, um die Evidenz für Neuromelanin-sensitive MRT als nicht-invasiven Dopamin-Proxy bei Schizophrenie zu untersuchen. Die Resultate weisen auf eine erhöhte Neuromelaninkonzentration in der Substantia nigra der Schizophreniegruppe hin. Studie 4 untersuchte experimentell die effektive transversale Relaxationsrate (R2*-Kontrast) als Neuromelanin-sensitives MRT-Maß. Es zeigte sich ein erhöhter R2*-Kontrast in der Substantia nigra bei Patient:innen mit Psychose-Spektrum Störungen. Beide Studien liefern Belege für Neuromelanin-sensitive MRT als nicht-invasiven Dopamin-Proxy bei Psychosen. / This dissertation is concerned with computational mechanisms of belief updating and neuromelanin as dopamine proxy in psychosis. The formation of psychotic experiences may be rooted in altered belief updating due to misbalanced weighting of belief updating signals according to uncertainty. Study 1 aims to delineate behavioral and neural computational mechanisms of belief updating underlying specifically delusions versus diagnosed schizophrenia. People with delusions showed increased neural updating signals in the caudate and anterior cingulate cortex, while people diagnosed with schizophrenia showed higher estimation of environmental volatility and lower neural representation thereof. Study 2 focuses on behavioral alterations of belief updating in people with subclinical psychotic-like experiences. People with psychotic-like experiences disregarded the probability of environmental change points. Their beliefs were less precise because they took fewer previous observations into account. Both studies support altered belief updating as a relevant mechanism in the formation of psychotic experiences, particularly delusions, across the psychosis spectrum. Belief updating signals are conveyed via dopaminergic transmission, which is a main treatment target of antipsychotic drugs. Quantitative dopamine proxies may inform treatment decisions in clinical settings. In study 3, we conducted a meta-analysis to examine the current evidence for neuromelanin-sensitive MRI as non-invasive dopamine proxy. The results indicate increased neuromelanin concentration in the substantia nigra of patients with schizophrenia. Study 4 experimentally investigated the effective transverse relaxation rate (R2*-contrast) as neuromelanin-sensitive MRI-measure to delineate people with and without psychosis spectrum disorder. The study showed higher R2*-contrast in the substantia nigra in patients. Both studies provide evidence for neuromelanin-sensitive MRI as a non-invasive dopamine proxy in psychosis.

abstract: Prior research suggests that people ignore evidence that is inconsistent with what they want to believe. However, this research on motivated reasoning has focused on how people reason about familiar topics and in situations where the evidence presented interacts with strongly-held prior beliefs (e.g., the effectiveness of the death penalty as a crime deterrent). This makes it difficult to objectively assess how biased people are in motivated-reasoning contexts. Indeed, recent work by Jern and colleagues (2014) suggests that apparent instances of motivated reasoning may actually be instances of rational belief-updating. Inspired by this new account, the current studies reexamined motivated reasoning using a controlled categorization task and tested whether people assimilate evidence differently when they are motivated to maintain a certain belief versus when they are not. Contrary to earlier research on motivated reasoning, six studies with children and adults (N = 1295) suggest that participants’ motivations did not affect their information search and their beliefs were driven primarily by the evidence, even when the evidence was incongruent with their motivations. This work provides initial evidence for the account proposed by Jern and colleagues. / Dissertation/Thesis / Masters Thesis Psychology 2019

Diese Dissertation erforscht, auf welchen unterschiedlichen Wegen Informationsverarbeitung Investitionsentscheidungen beeinflusst. Auf der Basis kontrollierter Laborexperimente wird untersucht, wie Entscheidungen mit der Art der Information sowie mit dem Entscheidungskontext variieren. Im ersten Kapitel legt ein Experiment die Schwierigkeit mit hypothetischem Denken bzw. mit dem Lernen aus hypothetischen Ereignissen offen. Im Kapitel Zwei untersucht ein anderes Experiment, wie Informationsverarbeitung die Reaktionen der Investoren auf Ambiguität verändert, denn ein eindeutiges, optimales Lernverhalten gibt es unter Ambiguität nicht. Das letzte Kapitel stellt anhand desselben Experiments die Unabhängigkeit zwischen dem Lernprozess und den Risikopräferenzen in Frage. / This dissertation investigates various channels through which information processing affects investment decisions. Controlled laboratory experiments allow for studying how subjects’ decisions vary with the type of information and the decision-context. The experiment in the first chapter discloses the difficulty with contingent reasoning, i.e. learning from hypothetical events. A different experiment in Chapter Two analyzes how information processing changes investors’ reactions to ambiguity—an environment with multiple rational learning rules. Using the same experiment, the last chapter questions the independence between belief updating and risk preferences.