The predictive aspects of an upper division certification examination for electrical engineering students

This work is an experimental study designed to define a method of identifying students who are at risk of performing below established standards in first semester, junior level electrical engineering courses. The procedure involves the administration and evaluation of a certification test designed expressly for this research.;The design of the test is discussed in detail including the choice of using a criterion-referenced format for the examination which is intended to measure mastery of subject matter and analysis capabilities. Included in the discussion is the need to involve several faculty members in generating test questions for inclusion in an inventory of potential test items. The need for more than one test, the number of questions in each test and the decision to use a multiple choice scheme is examined, as well as the selection of an appropriate mixture of simple and complex test items.;Analysis of 273 tests completed by 189 students is presented detailing calculations of test item difficulty and discriminating power and their use in the selection of test questions. The analysis includes a section that discusses the reliability of the test and concludes by comparing first semester junior electrical engineering course grades between those students who pass the test on their first attempt to those who require multiple tries.;The results of the analysis show a significant difference in junior course grades with those passing the test on the first try averaging 3.159 and those failing the examination averaging 1.230. This result is supported by a one-way analysis of variance (ANOVA) which shows significance at the one percent level.;The research shows that a certification examination is a valuable tool in identifying potential at-risk students. It should be used in conjunction with other grades and an overall department plan for advising students on future course work or remedial action.