This paper illustrates how graphical analyses can enhance the interpre
tation and understanding of multidimensional item response theory (IRT
) analyses. Many unidimensional IRT concepts, such as item response fu
nctions and information functions, can be extended to multiple dimensi
ons; however, as dimensionality increases, new problems and issues ari
se, most notably how to represent these features within a multidimensi
onal framework. Examples are provided of several different graphical r
epresentations, including item response surfaces, information vectors,
and centroid plots of conditional two-dimensional trait distributions
. All graphs are intended to supplement quantitative and substantive a
nalyses and thereby assist the test practitioner in determining more p
recisely such information as the construct validity of a test, the deg
ree of measurement precision, and the consistency of interpretation of
the number-correct score scale.