During the Spring 2011 semester, an engineering graphics literacy assessment was developed by a visiting Fulbright Scholar at North Carolina State University (NCSU) in USA. The assessment was developed for constraint-based modeling course and tested at NCSU. The faculty from two other universities joined the project to test this methodology. The students from Latvia University of Agriculture (LUA) in Latvia in Spring 2011 and later during the Fall 2011 semester the students from two sections at Riga Technical University (RTU) in Latvia participated in the test using SolidWorks 3D modeling software. All the 75 students from three universities were asked to create 3D models for seven parts given in an assembly drawing of a mechanical device within two hours time period. The parts in the assembly ranged in complexity from a simple ball to a complex valve body. Students were given a ruler to measure parts on the B-size third quadrant (A3 size first quadrant for Europe) drawing and determine sizes of geometric elements based on the given scale (2:1). There was a positive relationship between the scores on the activity and the pace at which each student completed the parts. It was difficult to compare the test scores on the modelling assessment and other measures in the course (final project, final exam, and final course average) because universities have different grading system. This paper summarizes how students performed on the assessment (number of parts correctly modeled, scores, total time, etc.), reports analyses of relationships between their scores on the assessment and other measures in the course, and also presents ideas for future studies. An average test scores in two universities were significantly lower than in LUA where the students had much extensive descriptive geometry and engineering graphics background.