Transforming STEM assessment methodologies: Research on cyber-enabled measurement of cognitive models of natural selection

The investigators propose to develop a form of assessment in natural selection -- and a computer based system to implement such -- that would use Latent Semantic Analysis (LSA) and hot term extraction as a means of transforming rich, open ended student responses into scores that are usable by faculty yet retain the complexities of the learning assessed.

The LSA approach will be compared to interview and multiple choice methods of assessment. Some comparative cost analysis will be performed. The project will be performed at OSU with undergraduates (both white students and students from underrepresented racial and ethnic groups).

Should this research be successful, it will create a new and more informative way to assess students? knowledge -- one that takes into account all the knowledge he/she brings to bear on the study of STEM education. This could provide instructors, and the students themselves, with more helpful and powerful information to improve teaching methods and learning approaches, respectively.