Research

In consumer research, I focus on heuristics and biases, specifically overconfidence. In the past, I explored and defined shared decision-making as a confounder of investor overconfidence. I reassessed the connection between excessive confidence and gender within the financial realm. In addition, I shed light on the impact of overconfidence on preschoolers and kindergartners by using a novel experimental setting with an innovative video intervention.

In data science, I am interested in Bayesian statistics, quasi-experimental methods using electronic health records, and other applications of causal inference.

Below is a list of selected past and current research:

Peer-Reviewed Journal Article:

Objective:  This paper uses quasi-experimental models to assess the impact of the pandemic on mental health among those who lived alone. The project combines tens of millions of electronic health records and data from cohort studies to provide a wide-ranging account of how the prevalence of mental illnesses among those living alone vs not living alone has developed over time.

Objective:  This paper uses quasi-experimental models to assess the impact of the pandemic on health outcomes in different social groups over time. Millions of electronic health records from several countries are analyzed to shed light on healthcare disruptions and inequalities.

Objective:  In three studies, we assess the effect of overconfidence on fear of COVID-19 and attitudes toward the virus. In Study 1, more overconfident participants had a laxer attitude toward COVID-19. While knowledge had an increasing effect on worry, confidence in said knowledge significantly decreased worry about COVID-19. In Study 2, participants who were more worried about COVID-19 were more likely to engage in protective behaviors (e.g., wearing masks). In Study 3, we show that when overconfidence was experimentally diminished, fear of COVID-19 increased. The results support our claim that the effect of overconfidence on attitudes toward COVID-19 is causal in nature. Moreover, the results show that people with higher fear of COVID-19 are more likely to wear masks, use hand sanitizers, avoid crowded places or social gatherings, and get vaccinated.

Objective: This study examines the impact of shared decision-making on investor overconfidence. It analyzes nationally representative data from 2,000 US investors, approximately 6,400 US consumers, and 239 experimental subjects to answer the question of whether investors who share the decision-making responsibility are less affected by the overconfidence bias than those who decide on their own.

Objective: The study sheds light on the motivation of innovators and early adopters to utilize a novel financial service to invest. The results of a series of robust generalized linear and structural models suggest that these consumers are excessively confident in their financial knowledge. Overconfidence increases the propensity to use automated financial advice and outperforms any income or risk aversion effect in this process. Causality is established with an instrumental approach and supported with a nonparametric matching function.

Objective: This exploratory study utilizes primary data from 60 participants aged 4 - 6 and their caregivers. The experiment involves a game theoretical gambling task and a video intervention. The aim is to examine the presence of excessive confidence and its potential impact on young decision-makers.

Objective: This study uses secondary data from 1,371 married investors and primary data from 320 married panelists to test whether sharing with someone in the household decreases overconfidence. It is argued that the perception of shared ownership of money partially explains the decrease in overconfidence among investors who share financial decisions.

Books and Chapters:

Objective: This chapter explores missing data in customer loyalty research in order to proactively assess and handle incomplete observations. Three types of missingness are defined and differentiated. Ad hoc, likelihood, and chained equation approaches are discussed and theoretically as well as empirically compared. Lastly, the chapter provides hands-on techniques to solve missing data problems in customer loyalty research.

Objective: Semi- or fully automated response tools, also called bots, decrease data quality and reliability in online studies that rely on crowdsourced data (e.g., MTurk). This publication describes how two online studies were conducted on a crowdsourcing platform in anticipation of bot responses. Specifically, it offers insights into the study design process, the selection of appropriate survey questions and bot traps, as well as the ex-post analysis and filtering of bot responses. Best practices are identified and potential pitfalls are explained. The description should aid readers in designing anticipatory online studies and experiments to increase their data quality, validity, and reliability.

Working Papers:

Objective:  This paper revisits the issue of a possible gender effect on investor overconfidence. Information from more than 30,000 respondents was used to assess whether excessive confidence in one’s financial knowledge can be associated with a specific gender after controlling for primary vs secondary decision-making.

Objective:  This paper assesses whether investor overconfidence helps crypto assets (NFTs and cryptocurrencies) to cross the chasm. Data from government survey responses, millions of NFT transactions, and an experiment shed light on the prevalence, impact, and causality of overconfidence in crypto markets.

Objective:  This paper assesses how the unavailability of products affects anxiety in retailing. Large-scale secondary and experimental data are used to establish validity and causality.

Objective: This longitudinal study uses an experimental design to test the effectiveness of traditional financial education to increase financial skills and dominant intertemporal behavior among college students.