RCR represents the field- and time-normalized citation rate, and is benchmarked to 1.0 for a typical (median) NIH-funded paper in the corresponding year of publication. For a portfolio of papers we present mean, median and weighted RCR as alternative and complimentary ways to interpret the results.
Median RCR shows the value for the middle paper in the portfolio, after ranking on RCR, so will not be affected by extreme values (RCR values follow a log distribution). Weighted RCR is the sum of the RCRs in the portfolio. Mean RCR is Weighted RCR divided by the number of papers. Both metrics reflect the RCR values of all papers in the portfolio. A highly influential set of articles will have a higher Weighted RCR and Mean RCR than the number of Publications, while a set of articles with below-average influence will have a lower Weighted RCR than Total Pubs. For example, a portfolio with 100 papers could have a paper with an RCR of 101 and 99 papers with an RCR of 1. This portfolio would have a weighted RCR of 200, a mean RCR of 2 and a median RCR of 1.
It is also possible to look at the number of papers in the top percentage of RCR papers e.g. the proportion of papers in the top 10% of NIH funded papers would have an RCR>3.81 (see here for more information).