The Great Recession probably played a role in the story. If you started work in 2007, the options at the end of your first and second school years may have seemed bleaker than for any class in a long time, perhaps leading the teachers in the national sample to hold on tight. In contrast, Kentucky teachers beginning in 2009 likely saw improving job markets as they made their choices about whether to continue: moving to another field may have seemed much more possible.
The data gathering methods may also play a role. The federal study tracked a sampling of teachers, surveying the same individuals each year except for those who could not be reached. An obvious risk there is that the people who could not be found again were heavily people who left teaching –maybe yielding too low a report of leavers. The Kentucky numbers reflect "3,452 traditional school teachers with teacher job codes" using data from the Kentucky Center for Education and Workforce Statistics. A risk there is not knowing about people who moved to other states but are still teaching –maybe producing too high a report of those no longer teaching.
On the other hand, the difference between 12% leaving (calculated one way) and 30% leaving (calculated another way) may partly be a reflection of other differences in new teacher experiences. The data may partly be a signal that Kentucky has room to improve our new teacher retention rates.
The new federal data and the state numbers discussed last week do not offer answers about what kinds of improvements would be helpful to teachers or to students, but they do clarify some questions that may be deserve systematic attention.