Data are everywhere. Data collected from samples are often reported in the form of polls, medical studies, and advertisement information and an understanding of sampling distributions and statistical inference is important for evaluating data-based claims (Bargagliotti et al., 2020). Despite the importance of understanding statistical inference and sampling distributions, research shows that these ideas are challenging for students (Sotos et al., 2007). I argue that one source of difficulty is that the forms of reasoning that are expected and highlighted in statistics differ from those that are standard in mathematics. Mathematical reasoning is often deterministic, emphasizing proof and deduction. In contrast, statistical reasoning is about reasoning probabilistically, not deterministically. Thus, statistics involves a different type of reasoning from what is required in mathematics. Reasoning about data is about reasoning under uncertainty, a feature of both inductive and abductive reasoning. The purpose of this dissertation study was to understand how novice statistics students reason about and with sampling distributions, particularly from the perspective of Peirce’s (1878) three classic forms of inferential reasoning—deduction, induction, and abduction. Understanding how far sample outcomes vary from a population parameter is critical in reasoning about sample data. I found that students reasoned inductively when they generalized patterns that they observed in sample outcomes and, as a result, developing a better understanding of sampling variability. Statistical inference is often equated with proof by contradiction, a type of deductive proof. However, the conclusions drawn from sample data are never certain; thus, inference under uncertainty is not deductive logic. Instead, I found that reasoning abductively was particularly powerful for making inferences from sample data to a larger population. Students who reasoned abductively to repeatedly hypothesize multiple explanations for the sample data they observed provided a reasonable estimate for an unknown population parameter and coordinated multiple levels of distribution—a critical component in understanding sampling distributions. The findings of this study highlight the importance of developing research-based instructional practices that promote critical forms of reasoning, such as inductive and abductive reasoning, to support students in constructing important meanings and understandings of statistical concepts.