How do teachers impact educational outcomes for students?   This page links to research that I have found helpful in exploring the very complex answer to this question. 

Most of these articles are from academic journals.  Therefore they are technical in nature, but are still accessible and understandable to non-specialists like me.  My advice:  read the introduction and the conclusion.  Skim the rest.  Don’t get scared off by the math & statistics.

1.  Teachers and student Achievement in the Chicago Public High Schools (Aaronson, Barrow, Sander 2003)
Bob’s simple summary:  Teacher quality has a material impact on student performance on tests, according to this study of Chicago high schools.  And teacher quality ratings are stable from year to year.   And the characteristics of a teacher that you would find in HR system (race, gender, years of experience, advanced degrees, certifications, and undergraduate college) explain almost none of the variability in teacher quality.  And because most teacher pay systems focus on experience and advanced degrees,  “these facts highlight the disconnect between teacher pay and productivity.”

2. The Impact of Individual Teachers on Student Achievement (Rockoff 2004)
Bob’s simple summary:   This research, now 10 years old, was one of the first to track teacher performance across multiple years in order to track their impact on multiple classes (cohorts) of students.   The result:  there are indeed large differences in teacher quality within schools.  And getting a higher quality teacher for one year results in excess learning of 5%-10% vs.being in a class taught by an average teacher. Finally, Rockoff concludes that “policies that reward teachers based on credentials may be less effective than policies that reward teachers based on performance.”

3. Teachers, Schools, and Academic Achievement (Rivkin, Hanushek, Kain, 2005)
Bob’s simple summary:   Using an enhanced statistical method, the authors examine Texas school data to show teachers have powerful effects on reading and mathematics achievement.  The results suggest that improving teacher quality is a much more cost effective intervention than reducing class size.  They also find that little of the variation in teacher quality is explained by observable characteristics such as education or experience.

4. What Does Teacher Testing Tell Us About Teacher Effectiveness (Goldhaber, 2006)
Bob’s simple summary:   Teachers in most states (including Pennsylvania) must pass a licensure exam as part of the qualification process to teach . The author studied licensure exam results in North Carolina to determine if the exam predicted teacher quality. While he does find a predictive effect, it is very small.  Goldhaber concludes that licensure test performance is not a ‘silver bullet’ credential that can be used to predict teacher effectiveness.   He writes, “credentials, like teacher licensure, provide only a weak signal of teacher quality.”

5. How and Why do Teacher Credentials Matter for Student Achievement (Clotfelter, Ladd, Vigdor, 2007)
Bob’s simple summary:   The authors examine all math and reading teachers in North Carolina over a 10 year period. They find that characteristics of higher quality teachers include: attending a competitive undergraduate institution, having more than 2 years of teaching experience, having a ‘regular’ teaching license, being a white female(!), being the same race as the students you are teaching (!) and having National Board Certification. Characteristics that do not matter: having a graduate degree, attending an elite undergraduate institution, and scoring well on the teacher licensing exam.  This is an important study because of the deep set of data analyzed — an entire state, over a long period.

6. The Distribution of Teacher Quality and Implications for Policy (Hanushek & Rivkin, 2012)
Bob’s simple summary:   The authors examine the validity of value-added estimates of teacher quality and look at policy implications, especially for teacher layoff decisions. The authors examine recent research on validity and bias problems with value-added models, including Rothstein’s widely-cited 2010 critique.  They find that value-added models are valid, and that the biases Rothstein identified are small, and can be controlled for with small adjustments to the models

7. Non-cognitive ability, test scores, and teacher quality (Jackson, 2012)
Bob’s simple summary:   Adaptability, self-control, grit, and motivation are important determinants of adult outcomes.  Jackson shows that teachers have a measurable impact on these non-cognitive abilities, and that these abilities are associated with desireable outcomes, including higher high school graduation rates, more SAT-taking, and making plans to attend college.  Therefore measuring teacher quality only through student test scores fails to identify teachers who excel in improving students non-cognitive abilities.