Assessment of a Weather Forecasting Contest in Multi-Leveled Meteorology Classes

Abstract

This study assesses an extra-credit weather forecasting contest offered in three different introductory meteorology courses, each consisting of a different student population: general education (Gen. Ed.) students, geology majors, and graduate students. Five semesters of data were collected, amounting to 15 Gen. Ed. sections (423 students), three majors sections (66 students), and one graduate class (16 students).Student participation in this optional contest is dependent on class level, ranging from 40% for Gen. Ed. students to 90% at the graduate level. Student responses indicate the primary reason for lack of participation is forgetfulness, with discouragement second.Scatterplots of final course grade versus final contest ranking indicate — to a statistically significant (> 99.5%) degree — those Gen. Ed. and majors students who do well academically will also do well in the contest. No statistical significance exists between course grade and contest ranking for graduate students.A quantitative assessment of student learning reveals that student forecasts improve relative to computer model output as the semester progresses. When averaged over all courses and semesters, the improvement is statistically significant at the 95% level. As a result of the increased experience that is acquired through contest participation, student forecasts — which initially have greater error when compared to corresponding model forecasts — have comparatively less error by semester's end.