This paper presents the design, implementation, and empirical evaluation of an experiential learning module for an undergraduate Quality Management course. The module employs a real-time dart-throwing simulation to demonstrate common-cause (random) variation, special-cause (non-random) variation, and the costly effects of tampering - over-adjusting a stable process in response to normal fluctuations. A mathematical model formalizes three adjustment rules: no tampering (stable baseline yielding variance σ²), last-aim adjustment (an MA(1) process that exactly doubles variance to 2σ²), and original-target adjustment (a unit-root recursive rule that inflates variance multiplicatively with run length). Probabilistic analysis of student-generated simulation data (N = 162 valid runs, 20 students) confirms these predictions: Experiment 2 yields ≈1.99× the variance of Experiment 1 at large σ, and Experiment 3 reaches 14.8× both patterns visible on a log-scale display. Pre- and post-module assessments (eight-item quiz, N ≈ 30–40 students per section) showed a statistically significant overall improvement of 9 percentage points (paired t = 2.82, p = 0.026), with the largest gain (+27%) on the foundational concept of predictable but inexorable random variation. The results validate the module as a replicable, cross-disciplinary tool for cultivating the probabilistic mindset central to quality improvement.

Keywords

Process variation, tampering, experiential learning, probabilistic thinking, simulation