Multidimensional Scoring

Historical Perspectives

World War II produced an abundance of brain-injured patients, which, in turn, stimulated attempts at the development of more formalized psychological assessment for localizing where in the brain the lesions had occurred. Brain scans had not yet been developed, and so the localization of the various brain functions that could be discovered through aphasia tests was of considerable interest.

In the 1960s, Norman Geschwind resurrected and formalized some of the early models proposed by Paul Broca and Carl Wernicke and others, and he developed a nomenclature for classifying aphasia. This led to the development of several tests to facilitate the localization process, although few of these were subjected to close psychometric scrutiny. The existing scoring methods in use included descriptive methods, plus-minus scoring, ratings of goodness of responses, and category scales that were subjective and lacking in capturing the specific nature of the response or not replicable across scorers and over time. In answer to this clinical need, Bruce Porch developed the concept [Page 1198]of multidimensional scoring and, in 1966, published the Porch Index of Communicative Ability.

Binary Choice Multidimensional Scoring

After observing that a wide range of responses to a given test item could be considered accurate, as there were a variety of possible attempts at each item that were inaccurate responses, Porch concluded from a detailed analysis of innumerable responses that the primary dimensions were the following:

• Accuracy: The degree of rightness and wrongness.
• Responsiveness: Indicates whether additional information is needed from the tester or from the patient’s first inaccurate attempt at the response, before successfully producing an accurate response. It includes self-corrections or getting a repeat of the item or a cue from the tester.
• Completeness: An accurate response that is incomplete grammatically, syntactically, or pragmatically in terms of the test requirements.
• Promptness: An accurate response that requires more than normal processing time to initiate or complete the response.
• Efficiency: The motoric skill with which the patient is able to respond or the degree of distortion or slowness in the response.

These dimensions were applied to the quantification of all accurate responses, but for inaccurate responses, it was not clinically important to [Page 1199]note whether the error was delayed or incomplete. That is, if a patient in trying to name a test object said “do-do-do,” it wasn’t significant if the patient took extra time to make a minimal response. However, it was important to document whether the patient could respond intelligibly or not, could attempt the task being tested, made an error that approached being correct, or was clearly wrong. Using this strategy, Porch was able to develop a binary choice multidimensional score system (see Figure 1).

...