Katie Paine, one of the measurement industry's pioneers, brightest minds and arguably most vocal supporters of the role that research and evaluation (or business intelligence) should play in communications best practices--who had this to say about measuring blogs back in 2005--recently published a paper for the Institute for Public Relations called:  How to Measure Social Media Relations: The More Things Change, the More They Stay the Same. 

My comment here is an observation on how times have indeed changed, not at all a critique of Katie's perspective in 2005 vs. 2007.  In 2005 the perspective of many in the measurement industry was essentially consisitent with Katie's: don't bother measuring blogs.   I suspect many would have attached the word 'yet' to that back then. 

In 2007, Katie's paper appropriately recognizes that that has changed calling for blogs to be measured using a variety of tools and tactics (a blend of some pre-existing off-line / traditional methods, some new on-line methods).  

While the paper is a good thought provoking discussion piece and a healthy contribution to a necessary dialogue, what's missing (though it might make for a whole other paper down the road, in fairness) is more depth on the notion of measuring and analysing the inter-relationships that make up the social network (and even between on and off-line media).  A sort of applied social networking theory used for some time now in sociology.  The paper does touch on it briefly but there is a great deal of value in taking this approach to measuring social media.  

There are a variety of commercial tools/services/software popping up or that have been around for some time intent on getting at this and that could provide another important layer of context and depth:

Hill & Knowlton's Influencer Network Analysis / Communications Mapping

Other commercial networking applications ( hat tip to Glenn O'Neil )

Bottom line: if we're talking about social media, which are really a series of nodes or communities of interest or social networks, let's look at using social networking theory (and related tools) to measure and analyse them.     

Moreover, there's some great blog discussion to with whether or not we need entirely new models to measure on-line PR.