It is a study of teacher effectiveness using a value-added model. The sidebar link to the paper detailing the statistical and modeling aspects of the study are worth a look. The punch lines are:
1. Good teachers matter a great deal to the performance of students.
2. Teacher background, training, and ongoing education have a very small effect on teacher quality.
3. Variation in teacher quality is greater in "Math" than in "English."
As the parcel tax debate continues, you should think about why step-and-column automatic raises remain non-negotiable, when their rationale may be invalid. Is there a body of modern research showing otherwise than is claimed in this latest study?
Note that the authors found that teachers had overcome their first-year jitters by the third year, when their performance did not differ hugely from teachers with 10 or more years of experience.