World Development Report 2018: Learning to Realize Education’s Promise

After prepared and motivated learners, equipped and motivated teachers are the most fundamental ingredient of learning. Teachers are also the largest budget item, with their salaries accounting for over three-quarters of the education budget at the primary level in low- and middle-income countries. Yet many education systems put in classrooms teachers who have little mastery of the subjects they are to teach— especially in classrooms serving poor children. Once in place, most teachers take part in some professional development, but much of it is inconsistent and overly theoretical. Meanwhile, education systems often lack effective mechanisms to mentor and motivate teachers. Such failures can be illuminated through models of human behavior—which also point to solutions (table 6.1). A synthesis of the evidence in these areas reveals three principles that are key to achieving learning success through teachers:

• To be effective, teacher training needs to be individually targeted and repeated, with follow-up coaching, often around a specific pedagogical technique. • To avoid learners falling behind to the point where they cannot catch up, teaching needs to be pitched to the level of the student. • Increasing teacher motivation with incentives can increase learning if the incentivized actions are within teachers’ capacity and if the failure to perform those actions has impeded learning.

Most teacher training is ineffective, but some approaches work

In-service professional development requires significant time and resources. A survey of 38 developed and developing countries found that 91 percent of teachers had participated in professional development in the previous 12 months. Two-thirds of World Bank projects with an education component in the last decade incorporated teacher professional development. Developing countries spend many millions a year to strengthen teachers.

But a lot of teacher professional development goes unevaluated—and much of it may be ineffective. One team of teacher training experts in the United States characterized professional development in the country as “episodic, myopic, and often meaningless.” Teacher training in low- and middle-income countries is often short and of low quality. Countries often have many training programs under way at the same time—in some cases dozens—with little to show for them (box 6.1).

Though preservice teacher training is important in providing basic skills (box 6.2), evidence on teacher training credentials is mixed. Much of the limited evidence on teacher credentials, generally from high-income countries, indicates they have no or extremely small effects on student learning. Simple statistical associations across francophone Africa suggest a positive relationship between teacher preparation and student performance, but that relationship could be driven by other factors, such as strategic placement of good teachers in desirable areas (where students would perform well in any case). Preparing teachers better is crucial, but the political economy challenges to doing so may be greater than for in-service training, and the evidence is more limited. The same principles that lead to effective in-service training serve as useful starting points for improving preservice training.

Is there hope for in-service training or professional development? Decidedly yes. Experience from high-income countries shows that practicality, specificity, and continuity are key to effective teacher professional development. Practicality means teachers are trained using concrete methods as opposed to theoretical constructs, and the training is classroom-based.

Specificity means teacher training programs are most effective when they teach pedagogy specific to a subject area (say, how to effectively teach a mathematics class). Continuity means teachers receive significant continual support—not one-offworkshops.

In teacher training programs, the inclusion of follow-up visits in school leads to higher learning gains. To bridge the gap between learning new methods in training and implementing them in practice, developing countries should make more use of follow-up visits in which trainers observe and support teachers in the classroom. In Africa, a range of programs with long-term teacher mentoring and coaching has shown sizable learning effects. In India, a program that provided little initial training to teachers but then provided support throughout the year significantly increased both math and language ability, with the largest gains for those students who were performing poorly at the outset. Teachers in Shanghai, China—where performance is high by global standards—participate in ongoing Teaching-Research Groups, which provide development, mentoring, and peer evaluation based on classroom observation.

Likewise, training associated with a specific pedagogical technique tends to be more effective. Across educational interventions in the United States, programs teaching a specific pedagogical method have more than twice the impact of programs focused on general pedagogy. Globally, specific guidance is crucial for low-skilled teachers, who may lack the ability to be effective even when motivated. At times, in settings where teachers have limited skills, this involves providing lesson plans that are highly scripted, outlining concrete steps for teachers. Many countries will protest that high-quality in-service professional development—repeated, with follow-up visits in school, often around a specific technique—is beyond their budget to deliver at scale. But teachers will not learn without receiving high-quality teaching themselves. A country facing this conundrum may be better served by delivering high-quality training in stages rather than ineffective training to all in the short run.

Helping teachers teach to the level of the student has proven effective

In many countries facing the learning crisis, it may be that only students who start at the highest levels of learning are able to keep learning. This is in part because teachers tend to teach to the most advanced students in a class. These students are the easiest to teach, and when teachers solicit answers from students, the high performers are the most likely teaching the students in the lower-performing group, and higher-quality teachers may be assigned to the higher-performing classes because these students may be easier to teach and so the assignment appears to be a reward.

In school systems with very low learning levels, ability grouping has had positive impacts on both lower- and higher-performing students. In Kenya, grouping students into classes by ability led to improved outcomes across the board, with the highest impacts among learners with more motivated teachers. In India, schools reorganized classes by group for just an hour a day and observed major gains in learning. Much of the rest of the evidence comes from the United States. Studies that relied on a credible counterfactual found that grouping students by ability either helps some students or at least has no adverse impact. In low-performing education systems, the lowest-performing students learn little to nothing (figure 6.1), so allowing teachers to target pedagogy may have a positive net effect.

Another way to help teachers teach to the level of to volunteer them. That leaves behind the students the student is to help them conduct better diagnoswho entered the class with less knowledge. Indeed, tics. In Liberia, an intervention that taught teachers to

Kenyan school dropouts identified that problem as a better evaluate their students was effective, especially primary reason for leaving school. Another reason when combined with training and additional materithat many students fall behind is that in many counals. So was a similar program in Malawi. In Singatries the curriculum may simply be too ambitious.

pore, students take screening tests at the beginning of

Teachers feel constrained to teach to the curriculum grade 1, and those who are behind in reading receive even when students have trouble keeping up.

additional support daily. By contrast, an intervenA key principle in leaving no learner behind is to tion in India that merely provided formative evaluhelp teachers teach to the level of their students. This ation was not effective; nor was another program in technique has been successful in different formats

India that provided diagnostic reports and written across a range of scenarios, whether by using comsuggestions on how to use the reports to strengthen munity teachers to provide remedial lessons to the teaching. Clearly, helping teachers to better underlowest performers, reorganizing classes by ability, or stand their students’ ability levels is worthwhile, but using technology to adapt lessons. In many cases, it if teachers lack the tools to respond effectively or the does not require a significantly greater teacher effort, incentives to do so—given that teaching students at but rather relies on restructuring classes or providing multiple levels is challenging—then it may not be remedial lessons for the lowest performers. A related sufficient. The diagnostics can work where a system principle of effective instruction is to reach students is in place to follow up, as well as where teacher motiby teaching them in their mother tongue (box 6.3).

vation is less of a binding constraint (box 6.4).

Grouping students by ability may allow teachers

New technologies offer promising ways to help with to more effectively target teaching to the levels of stuteaching to the level of the student. Computer-assisted dents in their classes. The theoretical effects of such learning programs can permit students to go at their grouping are mixed. The positive effects of betterown pace or adjust the level of instruction based on an targeted teaching have a potential downside: the initial screening test. More advanced software can adverse effect for lower-performing students of no not only screen students initially but also dynamically longer learning from their higher-performing peers.

adjust questions based on ongoing performance.

Furthermore, in early grades in particular, student

Although the overall evidence on computer-assisted ability is not always easy to measure, so separating instruction is markedly mixed, such a dynamic learnstudents by ability can put students on the wrong ing program among secondary school students in track. Teachers may also reduce their efforts when

Delhi, India, led to striking gains in both mathematics and language. Teaching at the level of the student is not a novel idea, but a range of new evidence is showing how it can be implemented—even at scale—in developing countries.

Teacher motivation and incentives make a difference, even with few inputs

No amount of training or inputs can substitute for teacher motivation. Because of high teacher absenteeism in many countries, fostering effort is a serious challenge. Moreover, even when they are in school, teachers are often not in class teaching. Yet education systems in many countries neither reward teachers for performing well nor penalize them for performing poorly. Teachers need to be treated as professionals—and good professionals receive support and respect, but are also held to high expectations. A system that does not pay attention to what its teachers are doing does not afford teachers the respect they deserve (box 6.5).

Over the long run, the best way to strengthen teacher ability and motivation may be to attract capable, intrinsically motivated people into the profession. In many countries and economies, the youth who plan to go into teaching are not among the highest academic performers (figure 6.2). In Finland, teaching is a coveted profession, largely because teachers receive great respect, are well trained, are reasonably paid, and have autonomy to implement teaching standards. Across many countries, average teacher pay has fallen relative

Better selection and retention policies will result to that of other professions. At the same time, the wage in better teachers. More meritocratic hiring—say, distribution in teaching has narrowed. High-ability based on a test instead of patronage—could improve candidates may be less attracted by a narrow pay student learning. One proposal would be to introstructure because it gives them little opportunity to duce a teaching apprenticeship of three to five years, reap professional rewards from high performance.

allowing systems to identify effective teachers.

Restructuring teacher pay both to remunerate competThe least effective teachers could then be transiitively and to provide returns to good performance— tioned out of the teaching force. In the United States, whether directly through pay or indirectly through proposals to phase out the least effective teachers promotion or retention—may improve the quality of suggest that the gains to learners over time would candidates entering the teaching profession. But this is be sub stantial: replacing the least effective 7–12 a long-term solution, not a quick fix, and even the best percent of teachers could bridge the gap between candidates need a supportive system to maintain their

U.S. student performance and that of Finland. Estiskills and effort over time.

mates of teacher value added in other countries are comparable, suggesting similarly large gains around the world to improved teacher selection.

Education systems need to build accountability to align incentives between teachers and others. Teachers have incentives and information that are distinct from those of students, parents, and administrators, and mental models and social expectations affect the decisions of all actors. In the absence of accountability to provide motivation, teachers may minimize their efforts even as learners and parents wish for them to exert more. In Argentina and Uganda, more than one-third of teachers surveyed do not see themselves as responsible for their students’ learning; in Senegal, the share is more than half.

Teacher motivation works through various behavioral mechanisms and comes in multiple forms. The fact that another person may observe their performance offers a form of professional motivation. So do evaluations, where teachers expect their performance to be assessed, with the associated consequences. Financial incentives for successful teachers and firing of neglectful ones are just two important parts of a broader spectrum of accountability interventions. At the same time, teachers in many environments face multiple demands beyond teaching, as well as risks such as late payment of salaries and even physical danger (box 6.6). It can be tempting, in light of data on high teacher absenteeism and low teacher skills, to blame teachers for many of the faults of education systems. But these systems often ask far more of teachers than teaching—and at times offer relatively little in return.

Financial and nonfinancial incentives are one possible mechanism for teacher motivation. In India, students performed better in primary schools that provided teachers with financial incentives for higher reading and mathematics scores. Students also scored higher in science and social studies, despite no financial incentives being offered in those areas. Other financial incentive programs were successful in two districts of Kenya and elsewhere in India. In the United States, by contrast, teacher financial incentives did not improve test scores in several states. However, large financial incentives for teachers did increase student learning in the District of Columbia, United States. In Mexico and Tanzania, teacher financial incentives were effective only in conjunction with another intervention. One interpretation of this scattered evidence is that financial incentives are most likely to be effective when teachers can take straightforward actions to improve learning. In environments with high teacher absenteeism from school or from the classroom while at school, it is likely to be clear to teachers that they can improve learning by simply coming to school and spending more time teaching. Alternatively, in environments like that found in the United States where teacher absenteeism is minimal, the specific actions that teachers should take to improve learning may be less obvious and less easy to implement. Nonfinancial incentives may include providing successful teachers with special recognition. Evidence of the effectiveness of these incentives in education is limited, although there is suggestive evidence in other sectors—for example among health workers in Zambia, where public recognition of worker achievement markedly improved performance.

Financial incentives can also create challenges. In Kenya, responding to a student incentive program, teachers taught specifically to the test, potentially neglecting more holistic learning. In a teacher incentive program in Mexican secondary schools, a significant portion of the identified increase in student learning was attributed to student cheating. In the United States, teacher cheating rose strongly when incentives were increased. And when teacher incentive programs are removed, the results can also be adverse.

With financial incentives, the devil is in the details. Incentives can be based on teacher inputs such as attendance or on outputs such as student learning. They can be based on reaching an absolute level of achievement or on gains. They can be available to all who reach a goal, or they can be competitive across schools. They will vary in size relative to teacher salaries. The evidence on these design elements is still limited, but they merit careful consideration, taking into account local institutions.

Likewise, the precise shape of a system’s overall incentive structure will vary by context. In some places, financial incentives may be worth piloting. In others, increased community accountability may be effective. The mixed evidence on these interventions suggests a need to examine carefully the context and to test programs locally. But while details will vary, no education system will be successful unless it provides incentives—whether implicit or explicit—for teacher effort.

Over time, education systems perform best when their teachers are respected, prepared, selected based on merit, and supported in their work. Countries should work toward these objectives. But in the short run, countries can take actions to strengthen the performance of teachers. They can improve the quality of professional development, shifting resources to the kinds of professional development that will change teacher performance in the classroom. They can support teachers in teaching to the level of the student. They can provide a professional structure so that teachers feel motivated to apply what they know. Teachers are key to learners’ education. Making them more effective in both the short and the long run is an excellent investment.