ABSTRACT Failure in introductory statistics courses is commonly evaluated using overall course grades, potentially obscuring important differences across content areas. This study examines factors associated with failure across four instructional units of a General Statistics course taught at a private university in Colombia: Exploratory Data Analysis, Probability and Random Variables, Statistical Inference, and Linear Regression. Using a complete‐case analytical sample of 186 undergraduate students, we estimated unit‐specific logistic regression models combining exam performance and pre‐exam self‐perception measures. Failure rates varied substantially across units, ranging from 19% to 46%. Students from programs outside the School of Applied Sciences and Engineering consistently exhibited higher odds of failure, while academic stage, prior programming experience, and self‐perceived competence showed unit‐specific effects. Robustness analyses supported the stability of these findings. The results demonstrate that academic vulnerability in statistics varies across content areas and that unit‐level analyses reveal patterns hidden by aggregate course outcomes, informing more targeted pedagogical interventions.
ABSTRACT A game with 11 players, 2 dice, and simple rules. However, it turns out that the best strategy is not only absent from the literature, but that misconceptions about winning strategies also prevail among surveyed students. This makes it all the more surprising that it is not so easy to get to the bottom of the game mathematically. We present a possible simplification to help understand the phenomenon in stochastics classes.
ABSTRACT The introduction of the bootstrap resampling technique constitutes an excellent pedagogical tool for illustrating sample variation. In this instance, the sample conceptually “becomes” (or is replaced by) the population; (re)samples are repeatedly taken from the “original” sample (which is now treated as the population). We provide a pedagogical framework in which this technique can be introduced to students for the first time. We propose a broad explanatory framework, favoring the use of specific examples. An understanding of the bootstrap will cement the ideas related to parameters, statistics, and sample variation in the minds of students.
ABSTRACT The ability to read and interpret statistical results is an important goal of an introductory statistics course. The ability to critically understand empirical papers is also crucial for economics majors as economics has become increasingly empirical in recent decades. Perusall is a social annotation tool that allows students to annotate readings. While instructors in some fields have found Perusall to be a valuable tool, the literature on its use in the teaching of statistics remains limited. In my Statistics for Business and Economics I course, I use Perusall to assign both the reading of relevant chapters of a free statistics textbook and the reading of academic papers. This article describes the main features of Perusall and how I use it in my course. I also discuss the specific papers I have assigned and the pedagogical reasons for my choices. The supplemental material contains the prompts I use for the readings.
ABSTRACT This study examined how worked examples should be designed to support statistics learning in online and in‐person contexts. About 114 undergraduate students completed surveys measuring comfort with quantitative subjects, mathematics anxiety, cognitive load, example preferences, and help‐seeking frequency. Students in both formats preferred structured step‐by‐step examples with progressive complexity. However, online students sought fewer additional examples, suggesting online environments create help‐seeking barriers. Comfort with quantitative subjects predicted greater help‐seeking and lower cognitive load, while mathematics anxiety predicted higher cognitive load with weaker effects on help‐seeking. Additionally, four student profiles emerged: (1) Struggling & Anxious, (2) Confident & Calm, (3) Moderately Anxious, and (4) Comfortable/Low Anxiety. Findings indicate effective statistics instruction requires both modality‐specific adaptations and universal design principles addressing students' affective preparation. Online environments benefit from proactive examples and structured help‐seeking mechanisms; in‐person environments can leverage real‐time interaction. Regardless of format, addressing student comfort and anxiety through scaffolded practice remains essential.
