Students Use Statistics to Justify Senior Project Selection

Engineering and technology curriculums typically include senior projects as the culmination of a student’s degree program. Students encounter difficulty during the selection process of the project due to uninformed decisions, lack of a structure, method or model, or an insufficient understanding of the resources needed to complete the project. The authors address this issue by introducing the utilization of a statistical model in a senior project course. In the course, students work in teams and propose, design, test and build their senior project over the course of two semesters. Previously, teams proposed a set of projects and then settled on a final selection after consultation among the members and the professor. In order to make the selection process more structured, the use of metrics was implemented to evaluate alternate project concepts. Students were given a spreadsheet template model for the metrics and instructed to adapt the model for their measurements. The metrics were based on parameters that measure the adherence to the principles of engineering design, man hours estimated, team skills compatibility, resource availability, team interest correlation and team mission statement and public need alignment. For example, the resource metric measured the availability and the relative cost of all major components, hardware, software, and tools. For each metric, the student team specified how it was to be measured. This process included the identification of what data was available that could be measured. Data sets utilized included relative cost of the proposed components, sales figures of similar products, and consumer satisfaction with similar products. From the data sets, a standardized, normalized score was computed for each metric. Using Bayes theorem, the prior probability was estimated and using the metric scores, the likelihood probability of project success and the posterior probability of project success was calculated using R language programs in the literature. The results of the utilization of the statistical model indicate an increase in student success, preparedness, and overall achievement of the outcomes of their degree program. Introduction Metrics are used to make measurements about performance in order to evaluate and compare. 1 They are widely used in sports to compare the performance of athletes in a game (e.g. batting averages and slugging average). 2 Likewise, Metrics are used to compare the performance of a task. 3 Software metrics are applied to measure the efficiency of the software/algorithm by measuring parameters such as speed and storage use. 4, 5 A simple metric can measure how long it takes to perform a task in actual time or man-hours (quantity), the number of resources required (quantity) and the quality of the outcome. A metric therefore usually measures quantity and/or quality. Since engineering design is performed using a series of tasks and has an outcome, it can be similarly measured using a metric. 6 The benefit of the metric is that it can be used to evaluate and compare outcomes such as alternate designs. The metric can be used to measure actual performance or projected performance. 7 Let us look at the metric for measuring the time of a task. In order to estimate the time of each task, the students first time a base task of each type: conceptualize, design, implement and test. An example of a base conceptual task is researching similar products on the market. A base design task is drawing the schematic for a circuit board using Multisim. A base implementation task is wiring the circuit board. A base test task is running the motor and recording the specification data. The time for each of these tasks is measured several times. Each task to complete the project is then measured in relative base times. In the authors’ course in senior project, the students start the course by brain-storming and conceiving three alternative and completely different concepts for their senior project. Project teams typically consist of two to three students. The course eventually will go through four phases: proposal, design, construction and testing. The project typically consists of a control system powered by microcontrollers and sensors and actuators. There are requirements and expectations for the project. The project must be unique and not commercially available. The project should further the interests and goals of the team members. It should contribute to society; be a potential product that the society will identify with and improve the standard of living or quality of life (make life easier). The potential product must have some part that is unique compared to anything that is in the marketplace. The product can be a totally new technology or a modification and improvement in current technology. It can be similar but cheaper to manufacture than those found in the marketplace. It can be constructed so it is easier to manufacture. Students must research thoroughly similar products in the marketplace. They must document this research and state what difference(s) their potential product has to the similar products they found through their research. To evaluate each of the three concepts and choose the best one of the three, the authors derived an evaluation system of the seven metrics described above. The evaluation method is shown below. The student teams were given a spreadsheet template for the seven metrics. They were encouraged to modify the metrics as appropriate for their concepts. Each metric included a fuzzy-weighting factor which imposed an increasing or decreasing emphasis on that particular metric. 9 The same seven metrics (Figure 1) were applied to all three senior project concepts.