CALDER Conversations: Tenure Reform

Tuesday, November 18, 2014

Tenure and tenure reform are in the news of late given the Vergara decision in CA and similar efforts challenging tenure laws in NY. This CALDER conversation covers some of the issues surrounding tenure and research that speaks to how tenure reforms might play out in practice.

Michael Hansen
Michael Hansen

Eric Hanushek
Stanford University

Helen Ladd
Helen "Sunny" Ladd
Duke University

James Wyckoff
University of Virginia

Q1: Recent legislative changes and court cases, particularly Vergara v. California, have weakened teacher job protections by lengthening time to tenure or, in some cases, eliminating tenure all together. How do you see this affecting the teaching profession?

Ladd: Superior Court Judge Treu ruled in the Vergara case that California’s job protections for teachers were unconstitutional because they interfered with the right of the student plaintiffs to have access to quality teachers.  The Court declared that by giving tenure to teachers after less than two years of teaching and then making it very expensive and time consuming to remove them, California permitted ineffective teachers to remain in front of classrooms, with particularly damaging consequences for disadvantaged students in high-need schools.  This ruling is consistent with a broader movement to weaken job protections for teachers.  

A number of arguments underlie this decision.  One is that good teaching is essential to student learning. Research by CALDER researchers clearly supports this position.  Another is that many existing teachers are ineffective and should not be permitted to remain in the classroom.  Although this view appears to be consistent with widely cited low dismissal rates for poorly performing teachers, such rates understate the extent to which weak teachers leave by ignoring ones who are counseled out or who leave voluntarily. Nonetheless, it is undoubtedly true that some of the remaining teachers are not competent to be in the classroom.  The proportion of such teachers, however, is likely to vary across districts depending on the capacity of their resource management departments.   A third and key argument is that tenure and job protections significantly contribute to the prevalence of ineffective teachers.  Here the argument seems to be that many unqualified teachers are given tenure, that once they are given tenure it is very difficult to remove them, and that the very existence of tenure means that low performing teachers have little or no incentive to improve on the job.   Finally, the argument is that the reduction of job protections will lead both to a higher quality teaching force and to fewer underperforming teachers, particularly for the students in high poverty schools. 

Although weak tenuring and dismissal processes in some states, including most notably California, should indeed be improved, it is hard to make the case that removing tenure and other job protections is the route to a more productive teaching force.  If tenure were eliminated and job protections in the form of due process rights were substantially weakened, teaching would become a much riskier career choice that, at current salary levels, would be far less attractive to potentially strong teachers. 

Instead, the path to a more productive teaching force would combine higher salaries and greater respect for teachers, along with good human resource policies.  Higher salaries and treating teachers as professionals would enable the sector to compete on a more level playing field with other sectors, such as law and business, for the best and the brightest college graduates.  The upgrading of human resource capacities would facilitate greater support for new teachers, better career counseling for all teachers, and better continuous professional development and improvement.

It is even harder to make a convincing argument that removing tenure and other job protections would improve the quality of teachers in schools serving large concentrations of high need students.   As is well documented by many CALDER researchers, teaching in high need schools is hard, and much harder than in more advantaged schools.  To attract and retain effective teachers in high need schools, we will most likely need to pay them differentially higher salaries, assure that the schools have strong leadership, and attend to the many out-of school challenges disadvantaged children bring with them to the school house door.   Eliminating job protections will do little to address their problems and could well be harmful.     

Hansen: I expect the weakening of tenure protections alone to have two direct effects on the current teacher workforce, which could work in opposite directions.

First, more teachers who score poorly (based on the evaluation system’s measures) will be terminated or discouraged from coming back in the next school year. Inasmuch as the evaluation system uses performance metrics that are valid and meaningful, then removing the weakest teachers based on these scores will increase the average productivity of the workforce. Even if these measures are noisy, as long as there is some signal, this will work to improve workforce quality (though noisy measures would make a smaller difference on the workforce, see recent work on selecting teacher quality with the evaluation system). This directly benefits students, and disadvantaged students in particular would benefit from having greater access to effective teaching. 

The second effect of weaker job security is that it will require more effort from teachers to prove they are effective enough to hold their position. Requiring more effort of teachers, however, could be met with either of two responses from teachers. Some may choose to exert effort to improve in productive ways, which would be a good thing for students. Yet, others may choose to exert effort in unproductive ways (i.e., by gaming the system or cheating) or may even be discouraged enough to exit the workforce altogether.

It is this second effect from teachers’ responses that I feel is the most uncertain element in trying to predict what might happen when job protections are weakened. I believe the potential productivity gains of requiring more effort from teachers can be meaningful, but likely will only motivate relatively few teachers  (studies from Dee and Wyckoff and Taylor and Tyler both find positive gains for some groups of teachers). However, many more could potentially be discouraged if the additional effort required to prove themselves is onerous—this could either be good for students (if this discourages the weakest teachers to leave) or bad for students (if this also discourages the strongest teachers, who are also likely to have other promising job prospects).

Beyond these effects on the current pool of teachers, this also changes the incentives for potential teachers who may be considering a career in public education. Here again, it seems lower job protections will have an uncertain effect on how this will shape the teacher workforce. On the one hand, it does make the profession slightly more risky as a whole, but on the other it sends a signal about taking quality seriously that may attract some candidates who may have otherwise been deterred by low performance expectations.

Ultimately what will happen to the teacher workforce is an empirical question, because I cannot say which of these potential effects will prove to be the largest influence overall. It will be interesting to see how this plays out in California and other states.

Wyckoff: The Vergara decision and similar pending cases in New York and elsewhere focus attention on the elimination of tenure or the substantial weakening of its protections.  While provocative, this framing creates an unproductive, polarized debate, where the real issues go unattended.  Short probationary periods coupled with perfunctory performance reviews make it likely that ineffective teachers are granted tenure to the long-term detriment of all students, but disproportionately poor, nonwhite students. 

Rather, we should focus on evidence of how to strengthen the tenure process to insure that it reduces the likelihood that ineffective teachers are given job protections that make it difficult to remove them.  There is developing evidence by CALDER researchers (Miller, Loeb and Wyckoff, 2014) that more rigorous tenure reviews induce weak teachers to voluntarily exit. There is also increasingly strong evidence that rigorous teacher evaluations have effects on the composition of teachers (Dee and Wyckoff, 2013) and on teacher improvement (Taylor and Tyler, 2012; Dee and Wyckoff, 2013).

Strengthening the tenure process requires empowering and incenting effective principals to support the development of novice teachers and providing them with more and better information about the effectiveness of probationary teachers.  This implies a longer tenure review process and more rigorous teacher evaluations than are typical in most districts. However if such a system were faithfully implemented, initially ineffective teachers would be much more likely to improve or be screened out, sending a strong signal to the vast majority of teachers that their effectiveness is recognized and valued.  In that case, I suspect there would be little discussion of whether tenure should be eliminated. 

Hanushek:  Virtually every business in the country effectively has a tenure system for its employees.  It is not in any business’s interest to arbitrarily or capriciously fire employees – because they would have to pay higher wages to attract workers and they would lose any firm-specific human capital that the employee has built up.  The firm balances the value of seniority against the interests of the firm (that are directly related to the productivity of the employee).  Tenure in education has unfortunately gone beyond any sense of reasonable balance between employee security and the interests of children.  Depending on how post-Vergara systems evolve, it is likely that individual teachers will be somewhat more responsible for their student outcomes.  This might change the mix of people interested in entering teaching – moving away from some of the highly risk averse people currently entering teaching toward a more risk neutral group that is willing to accept personal responsibility.  But it is unlikely to have a huge impact on the overall labor market for teachers.  It would be hard to argue, for example, that the potential teachers who will be most effective in the classroom are also the most risk averse.  It is also hard to argue that the entire labor market will dry up because there no longer is a promise of a lifetime job.   No matter what the replacement for the current system looks like, it will not involve large numbers of fired teachers or, put  another way, it will not involve very large changes in the risk to any individual.  For example, if the system simply reverted to civil servant due-process laws and away from the über–due process for teachers today,  one would not expect any rash of firing.  Vergara also addressed the LIFO statutes governing reductions in force.  If faced with a reduction-in-force such as seen after the 2008 recession, the specific people affected might change, but the total number would be largely unaffected.  (In California, it was illegal to consider teacher effectiveness in deciding on layoffs before Vergara; layoffs had to be strictly reverse seniority).  In sum, while some would like to paint a picture of massive uncertainty, steeply climbing wages to attract anybody into the job, and losses of the most effective teachers, these outcomes are just not realistic.  We will still have a supply of dedicated and high-quality teachers.  We may, however, have fewer misplaced people who do not meet the standards we expect of our teachers.  Indeed this could make teaching more attractive – because good teachers are no longer lumped together with the ineffective ones.

Q2: Do you think the weakening of job protections will have a net positive or net negative impact on the quality of the teacher workforce and student achievement? What empirical evidence that you would point readers to in support of your answer?

Hanushek: Most of the visible policy actions related to tenure do not focus on any specific policies that will come into force, and, whenever they do, they lack details.   As a result, the outcomes depend crucially on a series of subsequent decisions which can lead to very different institutions and incentives at the local level. 

The Vergara decision removes a set of onerous statutes that made it extraordinarily difficult if not impossible for California schools to manage their teaching force.  The legislature in California is likely to consider substitute statutes that meet the court’s constitutional objections but that at the same time provide a structure for teacher evaluations and teacher personnel decisions.  And those actions will help to shape the resulting impact on students. 

California is the only state so far with a court judgment, but a number of state legislatures have already taken their own steps to alter the evaluation of teachers and the restrictions of tenure and dismissal statutes.  It is also likely that the legislatures in a number of states, emboldened by the California decision, will contemplate pre-emptive actions before any court involvement. 

But state actions will not by themselves completely determine the results.  Local districts, even given unrestricted contracting possibilities, must in fact decide how they will proceed.  While everybody finds it convenient to blame either the rules, the unions, or the contracts for any problems, district administrations and school boards are often complicit in not actively making personnel decisions.  Take for example the response to the removal of tenure and a requirement for evaluating existing tenured teachers through legislation in North Carolina.  Two district superintendents challenged the law in court, essentially arguing that it was too difficult and divisive for them to evaluate their teachers.  And, from states that currently do not have tenure, it does not appear that the removal of tenure by itself will generally propel administrators toward better personnel policies.

The improvement of student outcomes will depend crucially on the actions of local officials, and these actions will depend on the incentives that they face and their reactions to them.  The tenure changes remove a constraint on local decision making but do not per se improve the incentives for quality.  One might hope that the focus on threats to performance that some of the more onerous restrictions imply will lead both legislatures and local districts to take actions that lead to positive gains for students.  But we do not yet have empirical evidence that we can rely on just changes in tenure policies to produce general improvements in the performance of schools.

Ladd: I agree with Rick that the outcome will depend on how state and local policy makers design policies to replace tenure or respond to them. As a replacement for tenure protections, state policy makers will presumably move teachers onto short term contracts, perhaps for two to four years, with the proviso that a contract will not be renewed if the teacher is not performing up to expectations.  The question then is how district or school administrators will evaluate their teachers and what criteria they will use to judge poor performance.  Because of their apparent objectivity, measures of teacher effectiveness that are based on gains in student test scores have recently received a lot of attention.  Research by CALDER researchers and others, however, show that such value-added measures have serious limitations for the purposes of making high stakes decisions about teachers. The general consensus is that, at a minimum, administrators should use multiple measures, including those based on classroom observations, and possibly information from student or parent surveys.   

Even if we assume that administrators implement fair and reasonably reliable measures of teacher effectiveness, there is no guarantee that the outcome will be a higher quality teaching force. As I mentioned above in question 1, the replacement of tenure with a system of this type increases the risk for individual teachers.  Hence, unless policy makers increase teacher salaries to compensate for this greater risk, the quality of the teacher pool is likely to decline. In addition, without the necessary compensating salary increases, the states or districts that move the most rapidly in this direction are likely to experience the greatest reductions in teacher quality as potential teachers seek better opportunities in other areas. 

Based on my own recent research showing that teachers continue to learn on the job well beyond the first few years of teaching, I worry that the move to short term contracts will adversely affect the quality of the teacher labor force by redefining teaching as a short term job rather than a longer term professional career.  The costs of that could be significant as schools would lose the benefits that come with more years of experience:  greater student learning, greater reductions in undesirable student behaviors such as absenteeism, and the experience-based knowledge and wisdom needed to mentor new teachers.      

Hansen: I fully agree with Rick and Sunny’s summary conclusion that the effect of weakening job protections on the quality of the teacher labor market will depend crucially upon policy implementation decisions that will be made in the future in most places. So, it’s hard to say exactly what effects will come of these changes.

However, I would also add that I see teacher perceptions as being another key variable in how these policies play out, because teachers will respond to what they perceive is a threat—or not—of unfairly losing their job. If teachers feel the use of multiple measures of performance is prone to excessive error, where every teacher is vulnerable to the threat of unfair evaluations, they may certainly respond in ways that are consistent with Sunny’s description where weak- or no-tenure jurisdictions could inadvertently drive away high quality teachers.

On the other hand, if teachers’ perceptions of the job-loss threat are more isolated to those on the very low end of performance – likely more in line with how new evaluation systems will play out – then an exodus of high-quality teachers seems unlikely.

For instance, Dee and Wyckoff’s evaluation of the new system in Washington, DC found over 80% of teachers earned summary performance ratings of “effective” or better, and thus were immune to sanctions or dismissal threats. Less than three percent were actually dismissed. If other new evaluation systems give the go-ahead to similar shares of the teacher workforce (consistent with the pattern we’ve seen thus far), no credible threat exists for the large majority of teachers.

This does not mean that the weakening of job protections are entirely innocuous – even a small rise in performance-based dismissals could have unforeseen chilling effects on the labor force if the messaging about it goes unchecked. States and districts, however, could likely mitigate much of these effects by making the evaluation process and the likely consequences for the teacher workforce comprehensible to teachers.         

Wyckoff: As discussed by Rick, Sunny and Mike, this is mainly about implementation. Will states and school districts choose to make the teacher assessment process a more rigorous, effective and meaningful experience with or without tenure? At this point that is quite uncertain. Providing principals with better tools to differentiate teacher performance, which may well require longer probationary periods, and incenting and empowering them to employ that information in the tenure process could improve teaching effectiveness through differential selection as well as teacher improvement. However, there are at least two ways in which this opportunity to reform the tenure process could fail students.

First, even in the face of judicial decisions that weaken or eliminate tenure in some districts, we could easily continue a system of teacher assessment whereby nearly all teachers are judged to be satisfactory or better after a perfunctory evaluation. Students are not well-served by a system that does not accurately differentiate teacher effectiveness, does little to identify and address teaching weaknesses, and fails to recognize truly effective teaching.

Second, this opportunity would also be squandered if the only response was a centrally mandated formula that attempted to define effective teaching. Teachers, principals and superintendents need good information about various aspects of teaching effectiveness, including components like value-added and rigorous observations. If principals merely apply the evaluation formulas, some truly ineffective teachers will be dismissed and highly effective teachers are recognized, each of which undoubtedly benefits students, but that would only realize a small portion of the needed improvement.  The real issue is creating the information, opportunities and incentives that allow most teachers to meaningfully improve their practice.

Rigorous, multi-metric evaluation systems can provide the information necessary for such improvement, but to be most effective that information needs to feed a principal led, school-wide culture of improvement that is grounded in rigorous measures of student learning and effective teaching practice. Tenure reform may be a catalyst for such a change, but that is not obvious.

Q3: How long before we might see changes to job protections translate into measurable positive or negative changes in student achievement that might be convincing to policy makers?

Wyckoff: Because there are several mechanisms by which changes to tenure policies may influence student achievement, I would expect these effects to develop over a period of years. As we found in our analysis of tenure revisions in New York City (Loeb, Miller and Wyckoff, 2014), some effects may occur within a couple of years, as policies become implemented and less effective teachers exit and are replaced by higher performing teachers. Other effects may take longer to develop.  We might expect the applicant pool of teachers to be influenced by tenure reforms and this is likely to occur over a number of years. Similarly, tenure reform may well lead teachers to develop stronger teaching skills.  This is, at least in part, dependent on the support to development that exists within schools and thus could play out over several years. For these effects to be convincing to policy makers it will require a research design that is able to identify tenure reform as the causal effect of changes in student achievement. 

Hanushek:  The clearest way to trace the effects of tenure reform to students is to follow changes in the distribution of value-added of teachers.  Are there exits of low performers?  Do those entering teaching after changes look any different from similarly situated teachers before reform?  Some of these could be observed within a few years, if there is a strong teacher evaluation system already in existence in a state.  Without an adequate evaluation system, however, it will be next to impossible to parse out rigorously any effects of tenure changes on the schools.

Ladd:  I have little to add to Jim’s statement, but take issue’s with Rick’s assertion that the clearest way to trace the effects of tenure reform is to look at changes in the distribution of teachers defined by their value added, whereby value added I assume he is referring to the contribution of teachers to student test scores.  First, one can construct plausible value added measures for only about 20 percent of the teachers, namely those teaching students in math and reading in grades 3-8 where students are subject to annual tests.  Second, even those measures fall far short of measuring the full contributions of teachers to the long run outcomes of students. Good teachers also contribute by influencing many of the harder to measure socio-emotional skills of students that may be even more important than cognitive achievement to their long term success.  Moreover the teachers who are effective along one dimension need not be as effective along another. I agree that states will need strong teacher evaluation systems, but those should rely only minimally on value added measures. Fair and comprehensive evaluation systems, along with competitive salaries, are both required to raise the quality of new entrants to the teaching profession.  It will take many years to see the effects on the talent pool, but raising the quality of that pool is the ultimate goal and is what we should be looking for.       

Hansen:  While we have addressed the possible impacts of how changing job protections might influence teacher labor markets in previous questions, the primary outcome of interest is not the teacher workforce itself but in student achievement. However, given the myriad of factors that contribute to student achievement, focusing on observable changes in the teacher labor market itself is the most practical way to estimate the extent to which changes in the teacher workforce contributes to student learning. On this point, I agree that the first place to look for evidence is on the characteristics and performance measures of exiting teachers. If policymakers are only concerned about the net change on the existing workforce, this evidence should be manifest within a few years of these changes. Isolating the overall impact, including the impact on the supply of teachers into the profession will be much more difficult, given the time horizon needed for these effects to play out. And given that tenure laws are generally not changed alone, in recent years at least, but have been accompanied by a host of other changes to teacher labor policies, I expect that it will be difficult to ever present convincing evidence of these specific changes alone.

Q4: Before we would see any changes in achievement, are there other signals that you would look to indicate that tenure reforms are having the effects you anticipate?

Hansen: These kinds of changes to tenure will likely affect the teacher workforce in two different ways: teachers leaving the profession and those who choose to enter the profession.

Of these two areas, I anticipate any effects would first be detectable among teachers leaving the profession. If school districts have been unable to remove ineffective teachers from the classroom under the old tenure policies, as some have argued, then changes to the law should result in a qualitative shift in the types of teachers that leave teaching (both voluntarily and involuntarily). In other words, the characteristics of exiting teachers—at the very least in terms of performance metrics, but presumably in terms of other credentials such as college selectivity or SAT scores—should show a decline in levels, consistent with weaker teachers leaving the workforce in larger numbers. This shift may occur over a few years, as districts’ evaluation practices and termination policies may take some time to full implementation, but I expect any effects should be evident within a short- to medium-term time horizon (roughly 2-5 years after the policy change).

On the other hand, the effects of this law on those choosing to enter the teaching profession (if there are any) will likely not be fully manifest for a much longer time period, perhaps 10 years or even more. Public attitudes towards the teaching profession and public knowledge about state-specific tenure regimes would likely evolve much more slowly, accounting for the longer expected incubation period for these effects on the workforce. Of course, incoming teachers may show noticeable changes in their pre-service characteristics compared against earlier entrants within just a few years of the policy change; however, these preliminary changes in trends would likely only capture a fraction of the full effect that will not become realized for many more years.

Wyckoff: I would be looking for changes in: the effectiveness of exiting teachers, the qualifications of the applicant pool and effectiveness of entering teachers, changes in the transfer or exit behavior of tenured teachers and changes by teachers and schools in the ways that new teachers are mentored and supported in their development.  Most of these effects are straightforward, but let me highlight the potential effects on tenured teachers. There is good evidence that teachers value the quality of their peers and to the extent that tenure reform meaningfully improves teacher quality going forward that could reduce the exit of effective teachers from schools, especially those with high rates of student poverty. For each of these potential signals I would be particularly interested in changes pre/post reform within the reforming districts but also changes in the pattern of teacher transfers and initial applications between reform and non-reform districts.

Hanushek:  The most important question is whether a high quality teacher evaluation system is in place or develops.  Additionally, to see any real gains it must be the case that a district moves toward a more differentiated personnel system that rewards highly effective teachers and considers dismissal of highly ineffective teachers.  If the system simply elevates the risk of being fired without also paying attention to good performance, it is unlikely that any significant changes will be seen – because it is no more attractive to high performers.  Thus, the impact of loosening up on dismissal statutes will depend on deeper changes because it is doubtful that entry and exit will be appropriately affected without deeper changes in personnel practices and in compensation systems.

Ladd: I would look for evidence of changes in personnel systems at both the district and school levels.  Even with the removal of tenure, teachers would, I trust, still have job protections in the form of due process requirements for dismissal or nonrenewal of contracts. For the new system to generate positive effects on student outcomes, which I would define more broadly than student achievement, districts and schools will need to have well-designed and well-managed personnel systems that lead to fair treatment of all teachers.  Such systems would include strong mentoring programs for new teachers, procedures to help teachers develop to their full potential, and clear standards for dismissal or nonrenewal. Adequate capacity at the district level is needed not only to address district level concerns, but also to assure that appropriate support is available for overburdened principals in high-need schools.   Failure to develop such systems could return schools to the historical patterns of discriminatory and arbitrary teacher hiring and firing that led to the adoption of tenure in the first place.