Bringing the Splendor of Statistics to All Students: Animations, Adaptive Problem Sets, & More
Kara Kockelman -- Department of Civil Engineering May 13, 2004
OVERVIEW: This Academic Development Funded project involves two major pieces of work: Interactive display of traffic simulations to emphasize data collection and evaluation, and adaptive on-line problem sets to hone student skills.
Work began in September 2003 on the simulations, using an ECE graduate student working with Flash MX 2004 Professional software. Signal timing and average vehicle arrival data were gathered at the intersection of Dean Keaton and Red River Road, and these were used to code a series of 100 full-cycle operations based on a Poisson arrival process. Since January 2004 FIC employees have taken over the programming, in order to tackle many issues unresolved by the ECE student and to enhance the visual displays. The Flash program animates the traffic while simultaneously charting relevant statistics, including traffic counts and inter-vehicle headways, for each direction (e.g., northbound through traffic). This program will become available online, accessible to faculty during lecture and students of statistics worldwide at any time. The simulations provide a concrete example of complex real-world processes, data collection and analysis, while demonstrating the connection between Poisson vehicle counts and exponentially distributed inter-arrival times, a subtle but important topic for stochastic operations analysis. The FIC team continues to work on the program files, but a prototype will be ready for the May 10 seminar.
Work began in January 2004 on the adaptive problem sets. A CE graduate student (who had TA’d the Fall 2003 CE311S course) and a top CE undergraduate have created a series of problem sets for a sophomore-level introduction to probability and statistics in the civil and architectural engineering disciplines. These problem sets allow students to test their skills and assess their knowledge in a dynamic and appealing manner, while offering instructors a way to prepare students and gauge student ability. The problem sets consist of six “modules,” each representing a different topic covered in the class: Probability, Discrete Distributions, Continuous Distributions, Confidence Intervals, Hypothesis Tests, and basic Linear Regression. Each module is a skill-adaptive quiz in which students must pass through a series of increasingly more difficult questions relating to the module topic. The questions relate to random events in construction, water resources, traffic, materials and other civil and architectural engineering topics.
The problem files were created using Brownstone Enterprise Diploma software and presently are available via a class web page. The program allows randomization of all numbers used in the questions so that equivalent problems rarely repeat, reducing the potential for cheating. Almost every one of the over 200 questions uses randomized numbers. A student taking a quiz from a module will start with a relatively low-difficulty question. If the student answers that question correctly, her/his next question will be from the next (harder) level of difficulty. A student must answer one question correctly from each difficulty level in order to pass the module. The modules each have five difficulty levels, except for the Simple Linear Regression module, which has three. Problems are being beta-tested, but are largely complete.