Psychometric properties of the revised PSVT:R for measuring first year engineering students’ spatial ability

While various spatial tests are available, the Purdue Spatial Visualization Tests: Visualization of Rotations (PSVT:R) has beencommonly used to predict students’ success in the engineering field. While many studies that used the PSVT:R exist, little attentionhad been given to its psychometric properties in measuring spatial ability and relationships to other academic indices. The purposesof this study were (a) to characterize the item- and test-level functions of the Revised PSVT:R for the use of incoming First YearEngineering (FYE) students, and (b) to investigate its relationship to academic-related variables to provide validity evidence.Approximately 2400 FYE students enrolled in the fall of 2010 and 2011 in a large Midwestern public university completed theRevised PSVT:R. Students’ academic-related variables were also retrieved from the university archive. A variety of statisticalanalyses, including exploratory and confirmatory factor analyses as well as item analyses, were conducted on the Revised PSVT:Rscores. Pearson’s product–moment correlation coefficients between the Revised PSVT:R and other academic variables were alsoobtained. The Revised PSVT:R measures a unidimensional subcomponent of spatial ability. Cronbach’swas 0.84. Items wererelatively easy and the test provides the most precise estimate for students whose ability level is at or below average. Weak tomoderate correlations were found between the Revised PSVT:R scores and the aptitude test scores. The Revised PSVT:R is apsychometrically sound instrument. However, items are relatively easy, but it is still appropriate to measure spatial visualizationability of the FYE students.