Anecdotes, data and the USSC's big challenge

I am hearing a range of post-Booker anecdotal reports from the federal sentencing front-lines.  I have heard stories of some district judges following the guidelines closely, apparently taking an approach similar to Judge Cassell in Wilson (basics here, commentary here and here).  I have also heard of a few judges focusing more on punishment purposes in particular cases, apparently taking an approach similar to Judge Adelman in Ranum (basics here, commentary here and here).  And, as detailed here, I have even heard of some talk of Blakely-izing after Booker.

Notably, the anecdotal reports of on-the-ground developments are now appearing in the papers.  This article details a federal sentencing from Hawai'i in which Chief US District Judge David Ezra focused on punishment purposes to justify a shorter prison term for a first offender who has undergone extensive sex-offender treatment after a conviction for downloading child pornography.  Similarly, this story from New York details a sentencing decision by US District Judge Richard J. Arcara which seemed to put the focus, in another a non-violent first-offense case, on rehabilitation concerns.

Because newspapers will always focus on cases where defendants were "cut a break," an enormous challenge for the US Sentencing Commission and others will be to keep Congress from making policy by anecdote.  In a huge federal system with thousands of sentencings every month, there will inevitably be stories of seemingly extreme decisions (both too harsh and too soft).  But no definitive judgments should be made (or responsive legislation drafted) based on a few anecdotes.  This is why I keep urging  everyone, especially Congress, to take a deep breath to allow us all to see what the US Sentencing Commission data shows for the range of post-Booker cases.

I hope, as suggested in this discussion, that the USSC's data collection will particularly focus on the distinction between violent crimes committed by persons with a long criminal history and non-violent crimes committed by first offenders.  My instinct is we will see a lot of guideline following in the former cases, and more variations in the latter cases.  But only the data, and not a few anecdotes, will tell.