Psychometric analysis of the selective reminding procedure in a sample from the general elderly population

The selective reminding procedure (SRP) has been proposed for the assessment of distinct aspects of episodic memory, i.e. storage to and retrieval from short-term and long-term memory, item learning and list learning, and as dementia screening tool. In the present study SRP results were analysed in 256 probands from the general elderly population. SRP scores were highly intercorrelated, and principal component analysis yielded only one single factor. The SRP scores were moderately and not differentially correlated with immediate and delayed free recall and recognition and with verbal fluency. All SRP scores discriminated nondemented probands with episodic long-term memory impairment from those without. The MMSE performed significantly better than any SRP score in detecting dementia. The theory-based assumption that the SRP allows assessment of different, independent aspects of memory could not be validated. It is suggested that the SRP is a mixed measure of semantic memory, episodic long-term and short-term memory, and working memory, and that the different SRP scores do not allow to assess different memory functions. Thus, the SRP may neither be recommended for assessment of different subfunctions of memory nor for dementia screening.