Thesis (Ph.D.) - Indiana University, School of Education, 2014Competent assessment practices allow psychologists to answer specific questions and formulate recommendations for intervention. One test, the Woodcock-Johnson Tests of Cognitive Abilities, 3rd Edition (WJ-III COG), has grown in popularity recently as a theory-driven test of cognitive skills. However, like all standardized tests, the WJ-III COG is only useful insofar as it is administered correctly. Minimizing error is crucial because results on standardized, norm-referenced tests only have meaning to the extent that they accurately measure the examinee's performance relative to the normative sample. Studies of other tests show that graduate students frequently commit errors, but previous research has only examined written products (test protocols) rather than assessing error on observed administrations. This research analyzes data gathered as part of the normal training process for graduate students in school psychology. First-year students in a school psychology program were required to conduct practice administrations of cognitive tests with children and submit videotapes of these administrations to their course instructor. These materials were obtained for research purposes and analyzed for errors. This study represents the first systematic examination of administration errors on the WJ-III COG (or any other standardized intelligence test) employing analysis of videos. In total, 34 videos were analyzed from 15 examiners. An average of 34.5 errors were committed per video (SD = 21.9). All examiners committed errors, most frequently failing to read test directions verbatim and improper administration of corrective feedback procedures. These errors have more potential to decrease examinee scores than the reverse. The vast majority of errors were not detectable by analysis of protocols alone.