Development of a General Undergraduate Estimation Skills Survey (GUESS)

Abstract

We describe the development of a ten-question diagnostic designed to characterize the estimation skills of undergraduate students in science and engineering. In order to establish a baseline and look for possible gains in skill level we have developed a multiple-choice assessment designed to probe student ability and confidence in estimating physical quantities such as mass, size, and time. The diagnostic was administered as a pre-test and post-test to a class of first-year engineers and given to a set of experts to establish its discriminatory power. Item response curves were then used to evaluate each question and multiple-choice answers. The results show that the assessment has the resolution to distinguish between student and expert scores, and that the distribution of expert confidences is qualitatively different than the students in both pre-test and post-test.