Elsevier has released it's latest journal impact factors - SNIP and SJR - and announcing some changes in the way they are calculated. They are now called SNIP2 and SJR2. SNIP stands for Source Normalized Impact per Paper and takes into account that citation practices vary across fields and the possibility to get cited is higher in some fields and lower than others. However, instead assigning each article to a field the adjustments are based on the length of reference lists in articles citing the journal in question and what share of those citations the journal in question has. SNIP2 makes some adjustments to reduce the effect of outliers and now the average SNIP will be 1 as will the average SJR. SJR is a recursive indicator similar to Article Influence or PageRank. Previously SJR strongly favored natural science journals and biomedical journals in particular. In other words, it wasn't source normalized. A first glance at the new numbers shows that SNIP doesn't look a lot different before, but that SJR numbers are now in similar ranges to SNIP.
At Crawford we are planning to allocate internal research funds awarded according to publication quantity and quality using SNIP to weight journal articles. It looks like the new metric will be just as suitable. We'll need more experience to see if SJR2 is superior to SNIP.