Poor children show little year-to-year movement in achievement because the system does not systematically give them several good teachers in a row—a good year is typically followed by being cycled through a poor teacher.

The most important school-based determinant of student achievement is the teacher, but not the commonly measured attributes like having a master's degree or more experience—rather some people are simply better at teaching than others.

In a poor inner-city school in Gary, Indiana, some teachers produced a year and a half of learning in an academic year while others produced only half a year, meaning a child's knowledge could differ by a full year depending on which teacher they were assigned.

A good teacher three or four years in a row can overcome the achievement gap associated with coming from a poor family versus a better-off family, refuting the view that disadvantaged students' outcomes are hopeless due to family background.

A great teacher placed in a school with poor overall culture (e.g., students roaming halls) would struggle to make a large difference unless the broader school culture improved and the teacher was complemented by other good teachers.

Reading and math are the focus of value-added teacher research not because they are uniquely important but because student gains can be tracked year-to-year; the same distribution of teacher quality likely exists in history and science but is harder to identify analytically.

Despite the move toward accountability under No Child Left Behind since 2001, the tests themselves have not improved as much as hoped, and there has been little solid effort to develop tests that capture the deeper understanding educators care about.

Political pressures following the imposition of testing—exemplified by the Atlanta scandal where teachers changed pencil marks after exams—are an incentive problem, and the same forces push teachers to advocate for easier tests nationally and resist testing improvements.

The Los Angeles Times in roughly 2010 hired a researcher to compute value-added scores for about 3,000 LA public school teachers and published them by name, which Hanushek did not consider a sound personnel approach but which prompted the district and unions to develop a better evaluation system.

Value-added scores can be confounded because a great teacher may be assigned challenging students and show only modest improvement, analogous to hospitals with low survival rates because they take the hardest surgical cases—so good evaluation must separate the teacher's effect from the composition of the class.

Earning a master's degree gives teachers about a 20% salary bump but has no average impact on their teaching performance, and additional experience beyond the first couple of years also has no average impact on student outcomes—yet these are exactly the attributes school pay schedules reward.

Experience can reduce teacher effectiveness after a point because teachers burn out and stop improving, with the metaphor that 'those old pages get pretty yellow' for some teachers.

Because hiring teachers involves irreducible uncertainty even with interviews, demo lessons, and lesson plans, the better policy lever is to make better decisions about who stays in the classroom rather than relying on screening at hiring.

Tenure rules (e.g., after two years in California, three in many states) protect ineffective teachers indefinitely, and the system uses devices like the 'dance of the lemons' (shuffling bad teachers between schools) and 'rubber rooms' (paying incompetent teachers to sit idle) because due-process clauses make removal nearly impossible.

A teacher at the 75th percentile adds about $400,000 per year to the present value of her students' lifetime earnings relative to an average teacher, and symmetrically a 25th percentile teacher subtracts about $400,000 per year—derived by combining teacher achievement-gain estimates, the earnings return to achievement, and class size.

Because teacher value is so large, the logic implies you should give your best teachers the largest classes (and small classes to poor teachers to limit their damage), but the current system does this exactly backward since pay is disconnected from value created.

Roughly 70% of what a child gains in fourth grade carries through to graduation (after allowing for depreciation each year), so a student who gets 75th-percentile teachers every year from kindergarten through 12th grade could accumulate around $250,000 in extra lifetime present-value earnings relative to one with average teachers.

The skeptic's view—that gifted students will eventually learn what they need regardless of a bad teacher and that family/genetic effects dominate—is contradicted by the policy consensus that schools do make a difference; in practice a student who falls behind in algebra is often shuffled onto a different path and never catches up.

The variation in teacher effectiveness is larger within any given school than between schools, meaning it is not the case that suburban schools get all the good teachers and city schools the poor ones.

Nations whose students score higher on the same math tests grow economically faster, other things being equal, suggesting that learning multiplies and has real macroeconomic impact.

Teacher unions dislike the finding that a small percentage of teachers are doing real harm because it bears directly on whether tenure protects all teachers or whether the system moves toward better evaluation, and some critics misinterpret any claim of teacher differences as being 'anti-teacher.'

Over the past ~20 years, US schools saw roughly a 20% increase in students, a 34% increase in teachers (driving smaller class sizes), and a 41% increase in total staff—showing that the non-teaching workforce grew even faster than teachers, both far outpacing student growth.

If the worst 5-8% of teachers were eliminated and class sizes allowed to be slightly larger, international estimates suggest the US would jump from below the developed-country average in math and science to near the top—so a fiscal crisis could be turned into a large gain for schools.

Last-in-first-out layoff rules force districts to fire the youngest teachers regardless of effectiveness while protecting ineffective veterans, so reductions hurt schools more than necessary—an effect a recent LA court order tried to limit for disadvantaged schools.

Stimulus spending was marketed by transferring money to states to prevent teachers, firefighters, and police from being laid off, which is an effective political marketing strategy because the public fears schools without teachers or unsafe neighborhoods.

Saying a child will learn 'half a standard deviation more' is meaningless to most parents, so achievement gains must be translated into clearer measures like dollars of lifetime earnings to be policy-relevant.

Good information about teacher quality can come from normal in-class evaluations by trained observers (as in Washington DC, where part to half of evaluation comes from observations rather than test scores), not just from testing—just as businesses combine measures rather than relying solely on counting output.

The current US education system is an emergent 'invisible hand' arrangement (not designed by anyone), and it is at least imaginable that, absent legal and governmental constraints, the value of great K-12 instruction would be higher relative to university instruction, since young students benefit more from a great teacher while adults can largely teach themselves.

Hanushek favors a middle path on reform: broader choice and charter schools (e.g., a quarter of schools being non-unionized charters) put beneficial pressure on the system, but because politics will keep ~80% of students in public schools, strong accountability and performance measurement are also needed, and marginal changes within the current system can yield substantial improvement.

Looking only at salaries understates high school teacher compensation because the job includes non-monetary benefits like summers off and generous pensions.

Most schools hand out pink slips to a vast majority of teachers every spring because they don't yet know next year's budget and contracts require notice, but historically almost none of those slips were executed—though more layoffs have actually begun to occur recently.

Good teachers are better than bad ones.