Statistical Reasoning with the Data Reasoning, Dual-process, and Metacognition Framework

Abstract

The current thesis explores the distinct nature of statistical reasoning in today's data-driven world, advocating for its separation from mathematical ability and emphasizing its cognitive aspects. Using three theoretical framework - the data reasoning framework, dual-process theory, and metacognition - the research provides insight into statistical reasoning, challenges traditional beliefs about intuitive responses, and examines the role of confidence and self-efficacy. The thesis consists of a correlational study identifying the relationships between various reasoning abilities, statistical reasoning, and confidence; an experimental study involving feedback on performance and mathematical perception of statistics; a systematic review of statistical reasoning ability measurements; and the development of the Statistical Reasoning and Intuition (SRI) scale. Furthermore, a conceptual replication extends the study, revealing insights into the differences between student and general populations in statistical intuition. The thesis reveals the underlying components of statistical reasoning that are distinct from mathematics, and the statistical reasoning explored in this thesis is primarily deductive in nature. It also highlights the importance of intuition in responding to statistical questions and demonstrates the use of positive, even if bogus, feedback as a robust strategy for enhancing self-efficacy, while encouraging further exploration of alternative methods to improve statistical reasoning's performance and confidence.