Psychometric construction of the Rey-Osterrieth Complex Figure: methodological considerations and interrater reliability.

Formal scoring criteria for the standard 18 elements of the Rey-Osterrieth Complex Figure were developed and applied to reproductions of 87 healthy young adults. In addition, the effects of immediate recall on delayed performance was investigated. Results indicate (a) the formal scoring criteria can be reliably applied by independent raters; (b) 30-minute recall is significantly higher if immediate reproduction is included compared to 30-minute delay performance without immediate recall; (c) significantly fewer qualitative scoring errors are present at the 30-minute delay if immediate memory is previously assessed; and (d) in addition to an absence of decay over a 30-minute delay, performance significantly improves over this interval. These findings underscore the caution that must be exercised when making specific inferences from test scores alone based upon measures that were obtained in a variety of contexts employing unspecified scoring criteria.