This just in: The EAC has released their biennial UOCAVA report, along with the raw data. Here is the link to the report page.

Key to any type of management-by-the-data is simply knowing what your work flow is like.  From a quick look at the tables in the back of the report, reporting of basic statistics got better in 2010, compared to 2008.  The percentage of jurisdictions actually reporting the number of UOCAVA ballots transmitted was at roughly the same level as in 2008, 98%.  The fraction of jurisdictions reporting how many ballots were returned rose, from 90% to 94%.

The details of transmittals and returns remains a bit of a black box, however.  For instance, when asked to account for the reasons why some returned ballots were rejected once received back for counting, only 75% of jurisdictions (up from 72% in 2008) reported how many ballots were rejected because they arrived too late to be counted.  Because a central issue in the passage of the MOVE Act was making it more likely that overseas ballots would return in time to be counted, it is disappointing that we do not have a more complete picture of where tardy returned ballots are a problem.

Of all the sections of the Election Administration and Voting Survey (EAVS), the UOCAVA data tends to be the most complete.  However, as with past years, the uneven appearance of missing data in the dataset underlying the report makes interpreting the statistics contained in the tables at the back a bit of an apples-and-oranges exercise.  With my colleagues at this blog, Pew, and MIT, I hope we can dig a bit deeper into the raw data in the weeks to come.