Crème de la Crème : The Teach For All Experience and Its Lessons for Policy-Making in Latin America

This paper analyzes the experience of Teach For All partners in different countries around the globe in three areas: (a) Recruitment and Selection; (b) Training and Development; and (c) Placement. Using information from semi-structured interviews with Teach For All staff members and national partners, we analyze the key elements of the Teach For All approach in these areas. Based on that analysis we argue that Teach For All’s active recruitment and rigorous selection of candidates, emphasis of quality over length in training, the continued support offered to teachers and the heavy emphasis on data and evaluation constitute practices that, if pursued at the systems level, could enhance the quality of teaching and, consequently, learning outcomes in Latin America.


Introduction
Teach For All partner organizations enlist their nation's most promising future leaders in the effort to expand educational opportunity by teaching for two years in high need schools and becoming lifelong leaders for educational excellence and quality.(Teach For All mission statement, personal communication, April 14, 2014) Crème de la Crème: The best people in a group or the best type of a particular thing (Cambridge Dictionaries Online, n.d.)In this paper we review the international experience of the Teach For All network and ask what can policy-makers in Latin America learn from that experience that is of relevance when crafting teacher policies at the system level.
The Teach For All network has developed an innovative model that recruits, prepares and places college-graduates as teachers in schools serving vulnerable students.At the core of its model is the presumption that bringing the most motivated and qualified -the Crème de la Crème -to education and further developing their strengths provides the basis for a transformation of the education system.
The central hypothesis of the paper is that there are some practices in the Teach For All approach that, with the proper adaptation, could serve well to improve some of the constraints experienced by the teaching profession in Latin America -and we seek to identify them.Their active recruitment and rigorous selection of candidates, emphasis of quality over length in training, the continued support offered to teachers and the heavy emphasis on data and evaluation constitute practices that, if pursued at the systems level, could enhance the quality of teaching and, consequently, learning outcomes in Latin America.The Teach For All experience also suggests that traditional routes to teaching are not the only way to bring effective teachers to the profession.
Through semi-structured interviews with Teach For All staff members and CEOs of country partners and a review of reports and documents we analyze the network's experience in three areas: (1) recruiting and selecting candidates, (2) training and supporting them, and (3) the placement of program participants in schools serving vulnerable students -all critical areas in which teacher policies in Latin America are lacking.
Acknowledging the difficulty of extrapolating lessons from the experience of a small organization for national level policy-making, we explore how Teach For All pursues these three core aspects of teacher policies that all education systems must contend with, and consider the extent to which their experience offers insights for policy-making.We do not seek to evaluate Teach For All, its approach or the performance of its national partners, but to learn from its experience.
The first section of the paper provide the justification for our approach, briefly reviewing the state of teacher policies in Latin America, identifying the key policy areas that are the focus of debates and introducing the Teach For All approach.Next we describe the methodology we followed and briefly review other studies, before presenting our analysis in the three core areas of teacher policy.This is followed by a presentation of policy lessons and a short conclusion.

Background and Framework for the Study Teacher Policies in Latin America
In the last 10 years, Latin America has made significant progress in terms of access, coverage, and investment in education.However, the quality of education continues to be very low and, consequently, learning outcomes tend to be poor in all international comparisons.For example, the latest PISA results (2012) -which evaluate 15 year olds in 65 countries, including eight in Latin America -revealed that in all subjects, Latin American countries ranked among the 20 countries with the worst results.Latin American countries perform worse than what their GDP per capita would predict.For example, Vietnam, a country with a lower GDP per capita than many Latin American countries, had much better results on the PISA tests (Vegas, Ganimian, & Bos, 2013).
Recent studies show major quality gaps in the teaching profession that, arguably, play a significant role in explaining poor student performance.A recent World Bank study by Bruns and Luque (2014) shows that, at present, no country in Latin America can be considered to have a high quality teaching labor force.The study argues that poor management of the academic content and ineffective classroom practices compromise teacher performance.For example, the study found that, on average, teachers use only 65% of the time in class instruction tasks, and use traditional teaching methods, which involve little student participation.
Country studies confirm this bleak picture.A recent study by Fundación Compartir for Colombia shows that, in the mandatory test that all university graduates must take, graduates from teacher training programs receive significantly lower results in national tests than the average graduates (García, Maldonado, Perry, Rodríguez, & Saavedra, 2014).In Guatemala, applicants to teaching positions that opened for competition in 2014 on average responded well to only half the questions related to language and teaching strategies, and to only a third of those related to math (Centro de Investigaciones Económicas Nacionales, [CIEN] 2014).The results of the Teaching and Learning International Study (TALIS ) of teachers conducted by the Organization for Economic Cooperation and Development (OECD) shows, for example, that only about a third of the teachers in Brazil and Chile (and slightly more than half in México) had any kind of induction in their current jobs (Cumsille R., 2014).
It is, thus, not surprising that governments in Latin America are, increasingly, recognizing the importance of improving the quality of teaching and undertaking policy changes to that effect.Chile has been the pioneer in this area.In the early 2000s, the government implemented a number of reforms, including an incentive-based teachers performance assessment system -accepted by teachers and their professional association -that evaluates teachers using a wide array of methods, including the preparation of a portfolio (with a videotaped lesson and a written component), a colleague's interview, a supervisors' evaluation and a self-assessment (Taut & Sun, 2014).Wellevaluated teachers are rewarded with 15% to 20% of their salary while poor performers are at risk of being fired (Mizala & Schneider, 2014).
Several countries are introducing reforms to their teacher training programs, establishing stronger admissions requirements and oversight of training institutions.For example, in 2006 Peru established a national bar for admissions to non-university teacher education programs.Only those applicants who demonstrated minimum competencies in a cognitive skills test, a writing test and a personal interview were able to enter the training program (Bruns & Luque, 2014).By 2010, the enrollment in non-university teacher training programs had decreased sufficiently and, in 2012, the Ministry of Education replaced the bar by yearly enrollment caps.That same year, Ecuador closed 14 teacher training institutions due to quality problems and is currently developing a pedagogic university run by the Ministry of Education to train new teachers (Bruns & Luque, 2014).
Teacher policy reforms are complex both in terms of the number of inter-linked aspects they involve and the political and institutional ramifications they have.Within that complexity, the following four questions define key entry points through which policies can influence the quality of teachers: • Who decides to become a teacher and how do they enter the teaching profession?
• How are candidates trained to become teachers?
• How are teachers assigned to schools and students?
• How are teachers supported and rewarded?Most teachers in Latin America are women (68.5%), and have been certified: teacher certification at the primary level is 75% (2008) and at the secondary level 64% (2009) (Oficina Regional de Educación para América Latina y el Caribe [OREALC] & United Nations Educational, Scientific and Cultural Organization [UNESCO], 2012).Even though the level of formal education among Latin-American teachers has increased over the years, there are clear indicators that those entering the profession do not have strong academic backgrounds.Bruns and Luque (2014) find that 15-year old students who are interested in teaching as a career have lower math scores in PISA than the average in all the Latin American countries participating in PISA -with the exception of Uruguay.They also report that pedagogy students in Chile and at the University of Sao Paulo in Brazil score significantly lower in the admissions tests than students applying to medicine or engineering schools.Moreover, admission requirements for training programs tend to be low, reinforcing the sense that the teaching profession is not very demanding.
There is substantive variation in the way teacher-training programs are organized in the region.According to OREALC and UNESCO (2012), there are four types of institutions that train teachers in Latin America: secondary-level institutions known as Normal Schools, non-university tertiary institutions called Teacher-Training Institutes, Pedagogical Universities (only devoted to teachers) and Universities with their Faculties of Education.Only Guatemala uses secondary-level institutions to train their initial and primary teachers (a practice that will end starting in 2015), and the rest of the countries use a combination of the other three types of training institutions.
There is also significant variation in the length of training programs.A pre-school teacher in Trinidad y Tobago will go through one or two years of training, while one in Argentina through four to five years.An upper secondary teacher in Guatemala is required to complete three to five years of training while one in Perú has to complete five to six years.
A recent assessment of teacher training programs in the region found a range of problems such as disconnect with school curricula, insufficient training in specific subjects for primary teachers, and an overspecialization in content mastery for secondary teachers at the expense of pedagogical aspects (OREALC & UNESCO, 2012).Moreover, the interaction of education students with real school environments tends to occur towards the end of the training programs, contributing to the perceived disconnection between training institutions and schools.
On the other hand, alternative routes of teacher certification -legal pathways to recruit professionals from other disciplines to become teachers -are not widespread in the region.Studying in education schools is almost the only option for those who want to teach in most Latin American countries.Alternative certification is only allowed in Colombia, and it is under consideration in Mexico and Chile (Bruns & Luque, 2014).Most of the Teach for All partners in Latin America have used legal loopholes to place their teachers in schools.
Very few countries have introduced teacher certification or competency tests for recent graduates.According to OREALC & UNESCO (2012), Mexico, Colombia and El Salvador require recent education graduates to pass mandatory competency exams to be hired in the public sector, while other countries like Chile and Brazil have voluntary examinations.
Once novice teachers are hired, the responsibility of training and supporting them falls mainly on individual schools.Even though most of the existing evidence suggests that the first three years of teaching are the most critical in terms of learning, induction programs are not a common practice in the Latin America and Caribbean region (Bruns & Luque, 2014).The exceptions are five Caribbean countries: Anguilla, Antigua and Barbuda, Belize, St. Kitts and Nevis, and St. Vincent and the Grenadines.However, Bruns and Luque (2014) find that these induction programs are extremely short -ranging from one week to four months.
During the 1990s, governmental institutions in several countries developed continuing training programs for teachers, including courses, workshops, and virtual training programs (OREALC & UNESCO, 2012).These programs attempted to unify training standards within countries, but research has shown that these initiatives lack consistency and resources in their implementation (OREALC & UNESCO, 2012).
Schools with high concentrations of poor students or located in rural areas and low-income neighborhoods in urban areas tend to have less qualified teachers.For example, Falus and Goldberg (2011) found that a larger proportion of teachers in schools located in high and middle class neighborhoods have a higher education training, know how to operate a computer and are more experienced than in schools located in low income neighborhoods and in rural schools.This pattern is found in all countries and is confirmed in studies of countries as diverse as Mexico and Uruguay (Luschei, 2012;Luschei & Carnoy, 2010).
Low academic standards for admission into teacher training programs, questionable quality of pre-service training, scarcity of induction opportunities for novice teachers, and weak professional development programs, combined with the low social status of the teaching profession further complicate the scenario of the teaching profession and represent some of the key constraints affecting the quality of the educational system in Latin America.

Teach for All
Teach For All is a not-for-profit organization founded in 2007 that promotes education equity, by supporting partners' programs to train college graduates and professionals and place them as teachers in vulnerable schools for a two-year period.The program includes a leadership component developed to drive long-term change in the education field through their alumni.
Teach For All has supported entrepreneurs to adapt the Teach for America (US) and Teach First (UK) models in different countries around the world.Teach for America, created in 1990, was the first organization that implemented the program and aimed to reduce education inequality in the United States.That year, 500 college graduates joined the program and began teaching in high-need schools.In 2002, the same model was adapted in the United Kingdom through Teach First.In 2007, Teach For America and Teach First co-founded the Teach For All network.The model has since been employed to reduce education inequality in 32 other countries, including partners in Africa, Asia and Latin America (see Table 1).More than 43,000 teachers have worked for a Teach For All partner, and the organization has reached more than five million students around the globe.The first Latin American partner -Enseña Chile -began its operations in 2009 and the region currently has six partners: Argentina, Chile, Colombia, Ecuador, Mexico and Peru.The region has had more than 300 Teach For All teachers and has reached around 20,000 students.Prospective partner organizations are also being developed in El Salvador, Haiti, Jamaica, and Uruguay.The Teach For All model promotes a non-conventional route to teaching by placing talented college graduates and young professionals from all academic disciplines in high-need schools (Teach For America, n.d.).The program provides an alternative route to the educational system of different countries, with the aim of improving equity in education within a nation.The model applied throughout the network includes (Teach For All, 2013): 1. Recruiting and selecting high-achieving college graduates and young professionals.
2. Training participants with the skills and knowledge to perform as effective teachers.
3. Placing participants as teachers for two years in disadvantaged schools.
4. Providing leadership experience to alumni by fostering connections among participants and offering training in leadership skills.5. Driving measurable impact in terms of student attainment and igniting long-term change in the education systems, by the leadership of their alumni across different sectors.Each partner has the autonomy to adapt the model to its local educational landscape, adjusting for the different challenges faced in diverse educational contexts and the organization has developed a network that offers access to global resources to assist partners in their work.

Methodological and Analytical Approach
Non-conventional routes to teaching are very scarce in Latin America, and as mentioned above, only Colombia has an alternative certification program in place.Non-traditional routes for teacher certification can help circumvent some of the limitations of traditional training programs to provide sufficient content mastery and student-centered pedagogy, as well as adequate practical exposure to work in schools (OREALC & UNESCO 2012), diversifying who enters the teaching profession.Non traditional routes played a significant role in raising teaching quality steeply in New York City and helped attracting teachers to disadvantaged areas in other cities (Bruns & Luque, 2014).
Teach For All, through its programs across the world, provides a unique experience with alternative paths to teaching.Their approach, as described above, includes innovations in some of the critical areas in which teacher policies in Latin America appear to be lacking: bringing strong candidates into teaching, training them and providing them with close supervision and support, among others.
Our goal in this paper is to review Teach For All's experience in three areas: (1) recruiting and selecting candidates, (2) training and supporting them, and (3) the placement of program participants in schools serving vulnerable students.In doing so, we seek to identify practices that could conceivably be applied at the level of a school system.We do not aim to evaluate Teach for All, nor to suggest that countries should adopt the Teach For All approach 'lock-stock-and-barrel'.The central hypothesis of the paper is that there are some practices in the Teach For All approach that, with the proper adaptation, could serve well to improve some of the constraints experienced by the teaching profession in Latin America -and we seek to identify them.As far as we know, there has been no previous attempt to do this.
First, we briefly examine the evidence from impact evaluations of the Teach For All approach to assess the effectiveness of their participants at the school level.Before we look at the implementation of their recruitment and selection, training and development, and placement practices, we want to know how effective the program has been in improving students' results.
Next we examined the Teach For All partners' experience implementing the network's approach, concentrating on the three areas mentioned above.For that purpose, we conducted semistructured interviews with three Teach For All staff members and CEOs of eight country partner organizations (see Table 2).The interviews were conducted using an interview guideline (included in the Appendix), with questions for each area of analysis: recruitment and selection, training and placement.We also reviewed reports and documents facilitated by the Teach for All staff and partners that had official information about each program.The interviews were arranged through the Teach For All network.At our request, Teach For All staff proposed a list of country partner organizations that reflected the diversity in implementation experiences in the three areas of focus in our analysis, following a maximal variation sampling strategy (Flick, 2009).Teach for All staff facilitated their names and email addresses and we interviewed all those included in the list with the exception of partners in Chile and the United States (who did not respond to our invitation to participate in the interview), and in Israel and India (who were not contacted due to time limitations).
The interviews were conducted by at least one of the authors using the interview guideline in the Appendix.The interviews lasted between 30 and 60 minutes approximately, and they were taped for the analysis.The interviews were selectively transcribed in grids, according to the three areas of interest for the analysis: recruitment and selection, training and development, and placement.The analysis was performed looking at the commonalities and differences between the partners' experience in these three areas based on the grids.

Teach For All's Impact
The literature evaluating the performance of members of the Teach For All network is uneven in coverage, with many more studies for Teach for America (TFA) than for the rest of the countries.As is often the case, studies vary in their methodologies.A meta-analysis -including a review of methodologies -is beyond the scope of this paper.Instead, in the case of the United States, we briefly reference the results from the most cited studies.Outside the United States, given the lack of any published studies using random assignment, we review the evidence from quasiexperimental studies using propensity score matching methods.
Overall, the research finds mixed results of TFA teachers' effectiveness, and the results are highly dependent on the comparison groups used to evaluate TFA fellows (Vasquez Heilig & Jez, 2014).In the United States, results from random assignment studies (Antecol, Eren, & Ozbeklik, 2013;Clark et al., 2013;Decker, Mayer, & Glazerman, 2004) have found that TFA teachers are more effective than traditional teachers, measured by student achievement.Experimental research has consistently found a positive impact on students' math outcomes, but these studies have not been published in peer-reviewed journals.On the other hand, in a paper published in this journal, Darling-Hammond, Holtzman, Gatlin, and Vasquez Heilig (2005) use non-experimental methods and find that TFA uncertified fellows are less effective than traditionally certified teachers.In a couple of reports, also not published in peer-reviewed journals, Vasquez Heilig andJez (2010, 2014) have questioned the positive results in Clark's et al. (2013), due to the lack of representativeness of their TFA teacher sample and the small magnitude of their results.
The international evidence available has supported positive impacts of Teach for All partners in the United Kingdom and Chile.In the United Kingdom Allen and Allnutt (2013) find that having a Teach First educator in the school improves schools' exam performance in the second and third years after the program is implemented.However, results do not compare the effectiveness of Teach First educators vis-à-vis traditional teachers; instead, the study assesses the effects on school performance of having a Teach First instructor in a school.In Chile, Alfonso, Santiago, and Bassi (2010) estimate the impact of having an Enseña Chile teacher on student test scores and other nonacademic indicators such as self-esteem and self-efficacy.Their results show that having an Enseña Chile teacher improves students' result in math and Spanish, even when the teachers do not necessarily teach those subjects (Alfonso et al., 2010).The same study also reveals that Enseña Chile also has a positive impact on students' self-efficacy and self-esteem.Overall, the evidence from studies using experimental or quasi-experimental approaches tends to show some positive results but is too limited (particularly outside the United States) in scope to make definitive statements about the impact of the Teach for All model.

An Analysis of the Teach For All Experience
The core aspect of Teach For All's approach is attracting talented college graduates to teach in vulnerable schools.This section analyzes the strategies followed by Teach For All partners in three areas, (1) Recruitment and Selection, (2) Training and Development, and (3) Placement of Participants.Our analysis is based on the information collected through interviews with Teach For All staff and partners and selected documents and information facilitated by the Teach For All staff and partners.

Recruitment and Selection
Recruiting a large pool of qualified candidates and following a systematic process to assess them are key aspects of the Teach For All model.Application requirements vary from country to country, but in all cases include having a college degree or its equivalent by the time participants start their training.Teach For All partners follow an active and targeted recruitment strategy.Most of them visit prestigious universities in their home countries to recruit students into the program, organizing informational sessions to encourage students who meet the requirements to apply.They also actively target young leaders and set up personal meetings with recruiters to encourage them to apply to the program.In some cases, such as in Teach For Malaysia, national recruitment is supplemented by international efforts.Indeed, given the large number of college-age Malaysians studying abroad, the recruitment strategy includes visits to universities in Australia, the United Kingdom, and the United States, as well as follow-up calls to students who have started an application to persuade them to join the program.
Recruitment efforts also involve media campaigns seeking to create interest in the program.The messages in those campaigns are often adjusted in response to reactions by potential participants and emphasize both altruistic and personal goals.For example, by stressing the message that participants could have a positive impact at the school level, EnseñaPerú was only succeeding in recruiting candidates interested in voluntary work but not other potential candidates more inclined to work in private companies.They rethought their communication strategy and developed a new message focusing on the opportunities the program offers for personal and professional development.Teach For Malaysia for example, stresses the idea of teaching as contributing to 'nation-building' while Enseña por México has chosen to focus on personal opportunities for the fellows with the slogan "Do it for yourself" (Hazlo por ti).
The experience of Teach For All partners suggests that the desire to perform a meaningful job and the opportunity to learn and develop skills that will be valuable in future jobs including ones not related to the education sector-serve as strong motivators for applicants.Furthermore, the perception that the program is highly selective in its admissions policy helps attracting competitive applicants to the program.
According to most of the partners we interviewed, salary levels are not the main factor in the attracting talented candidates.Some fellows are paid at the level of the average teacher in their country (Teach For Malaysia), others are paid less (Empieza por Educar from Spain, Enseña por México) and a last group receives more than the average teacher (EnseñaPerú, Teach For Pakistan).Nevertheless, they are all paid below the market salary for a young professional.This does not appear to limit the ability to recruit talented candidates into the program.Teach For All partners in Spain and Malaysia noted that as long as the fellows can sustain themselves, low salaries are not a problem.In Pakistan, where teachers' salaries are particularly low, the program pays twice the average teacher salary, but still 35% below the market salary for a young professional working in other sectors.
It appears that the experience gained through the participation in the program is valued sufficiently to compensate for the opportunity cost of being a teacher for two years.However, this might not be true for the most competitive and best-prepared applicants.An experimental study by Ganimian, Alfonso, and Santiago (2013) that examines the experience of Enseñá por Argentina, reveals that candidates at the very top on the quality distribution had a higher probability of dropping out, and receiving information on their pay did not change their decision.
Moreover, the evidence is insufficient to determine whether salary levels would act as a disincentive for remaining in the teaching profession after completing the program.Recruitment messages that describe the program as an opportunity to both perform a meaningful job and develop skills that will be valuable in the job market may be sufficient to counterbalance the opportunity cost for two years but it is unclear to what extent those effects can be sustained over time.
After the partners have received the applications, they engage in an exhaustive selection process that includes several steps and evaluates skills and attitudes.Teach For All partners have developed a competitive selection process for their fellows, in which only a small number of applicants are selected to participate in the program.The acceptance rate varies from around 5% (Spain) to approximately 21% of applicants (United Kingdom) according to the figures provided by Teach For All staff.The process is long and intensive to prove candidates' willingness and motivation to participate in the program, and includes an online form, group activities and personal interviews.
Candidates are appraised not only in terms of their academic and professional achievement but also in terms of the mindsets and competencies needed to be an effective teacher.Each country defines the specific mindsets and competencies according to the challenges teachers faced in the schools they serve.Subject knowledge is not the only requirement for admission-partners also seek participants demonstrating the soft skills associated with effective teachers.The most mentioned personality traits are leadership capacity, resilience, and self-evaluation.Interviewees also cited others such as humility, empathy, communication skills, perseverance and perception of self-efficacy.
Overall, the program emphasizes interpersonal skills as fundamental for successful classroom experiences.
Evaluators look for demonstrated evidence of those abilities in fellows' past activities and observing the way they behave in group activities and personal interviews.Empieza por Educar's CEO describes the selection process as follows: This is the philosophy of our selection process.We base [our decisions] on evidence.
We must obtain that evidence.If we do not have that evidence, it does not mean it does not exist.It means either the candidates have not demonstrated them or that we have not been able to see them.And well, part of the challenge is to minimize those risks.But we do not base decisions on intuitions.Teacher training institutions and school systems more broadly could conceivably apply both the intensive recruitment efforts and the systematic evaluation of candidates' abilities, including those related to inter-personal skills.This approach increases the cost of recruitment and selection processes -especially in the case of high turnover programs like TFA (Vasquez Heilig & Jez, 2010) -but their adoption by education schools, which prepare teachers for long-term careers in education, could have a major impact on the school system.To the extent that they increase the quality of those entering the teaching profession, the returns to that investment may still be high.Clearly, Teach For All considers this is a valuable, worthwhile long-term investment for their program.
Critical to the approach is the definition of the 'right' evaluation criteria used for admission into the program.We did not review those criteria nor are we aware of studies that show their predictive power.

Training and Development
Teach For All's approach to training includes both pre-classroom preparation and support during program participation.Here we review both aspects.
Once fellows are selected, but before school placement takes place, they participate in a training program.In most countries, pre-placement training involves participation in a Summer Institute that combines theoretical and practical training.Some countries also include practical training activities before the Summer Institute.Teach For Malaysia and Teach First, for instance, place participants in schools for a week to observe and get familiarized with a school environment and context before the Summer Institute begins.Teach First also has online preparation activities before the start of in-person training.Most programs have alumni or experienced teachers involved in the training.A few, such as Teach For Malaysia and Teach First, also work closely with universities for the training.
Summer Institutes prepares fellows for their placement in schools.Training includes the basics of pedagogy, classroom management skills, leadership abilities, and, when needed, content mastery.Institutes last between four to eight weeks.Some partners describe Institutes as the final step in the selection process -even through dropout rates are low.Indeed, based on the information gathered during the interviews, drop out rates once the Institutes start are between 1% and 2%, suggesting that the rigorous selection process is effective in identifying qualified candidates for the program and training does not discourage participation.
Practical training is one of the main pillars of Summer Institutes.Most of the programs include a summer school with real students as part of the training to encourage learning by doing.Enseña Perú's CEO compares the classroom with a soccer playing field: "…[Fellows] learn by doing themselves."Fellows usually teach in the mornings and spend the afternoons and evenings receiving feedback and analyzing their classroom practices.Summer Institutes also emphasize evaluation.Instructors and tutors assess participants' progress in technical competencies using defined rubrics.Each country defines certain skills to be developed during the Institute and a minimum threshold of mastery for their participants.In the case of Teach For Pakistan, for example, before they are placed in the classroom participants must perform at least 3 on a scale from 1 (pre-novice) to 5 (exemplary) in a set of teacher actions such as being able to set a big goal for their class, planning a lesson in which all elements align, or engaging students in class activities, among others.Most partners also place strong emphasis on participants' values and behaviors in their evaluations.Teach For Malaysia's CEO mentions that when evaluating a participant, …we assess more potential and willingness to develop, so if someone is not technically competent but is putting in a lot of effort, and they can believe they can change, if they have a positive mindset, we'll continue to support them.The ongoing evaluation and supervision of fellows during the training process allows partners to identify candidates in need of extra help and to design targeted improvement plans for them.Enseña por México and Enseña Perú, for example, offer two-week improvement plans to support fellows that require extra help.
Partners do not expect to have their fellows completely ready for their first day in the schools.The idea of "first day readiness" -i.e.fellows being completely prepared for their first day in the schools -is a controversial idea, usually discussed among the partners and sometimes criticized by other stakeholders like teachers unions or teacher training institutions.Some of those we interviewed admitted that even an intensive Summer Institute cannot prepare fellows 100% to confront all the challenges they will face in the classroom.However, these same people argue that the Summer Institute seeks to develop and strengthen the attitudes and skills that will allow fellows to improve their performance in a more independent way during their placements.
More generally, the "first day ready" challenge is addressed through support and professional development after placement.All programs have a permanent support mechanism for fellows during the two years.Every fellow has a mentor or tutor, and in some programs one participant can have up to three different mentors, as in the case of Teach First in the United Kingdom.Mentors can be teachers from their same schools, professors from a program's partner university or alumni from the Teach For All network.Mentors usually serve 10 to 20 fellows, visiting them at school, observing their classes and meeting with them regularly to provide feedback.A useful distinction made by Teach First is between tutors that offer professional and subject support.In their experience, professional tutors (usually an experienced teacher in that school) provide guidance on administrative matters or class management practices.Subject tutors, on the other hand, are faculty in teacher training programs and offer support on content related matters.Support activities also include periodic fellow meetings during the weekends.Teach For Malaysia, for example, organizes targeted seminars on weekends that bring together groups of fellows based on the needs reported by their tutors.
Beyond the role of mentors, there are a number of ways in which fellows are evaluated as a means of defining the kind of support they need.The programs emphasize peer-feedback as one such mechanism.Fellows' performance is also appraised through students' achievements, providing information on areas in need for improvement.Empieza por Educar in Spain has also incorporated an interesting evaluation mechanism: they assess their fellows using student surveys twice a year.Using survey information they evaluate their teachers' strengths and weaknesses and identify areas for improvement or support.
Mentorship and ongoing support, particularly to novice teachers tends to be a weak aspect in most education systems in Latin America.While there is certainly not just one way of providing those services, the strong emphasis they receive in the Teach For All model serves as a reminder that a lot more could be done at the education system level in this critical area.
One controversial aspect of the Teach For All training model is its length.It is hard to argue, based on the existing evidence, that shorter training programs ought to become the rule for teacher preparation.Another critique points to the fact that the Teach for All training model focuses mostly on the professional exercise of teaching based on available theory and models, while schools of education contribute to the production of theory and knowledge (Friedrich, 2014).In terms of system-level policy formulation, intensity, a focus on practical (as opposed to theoretical) training and the emphasis on mentorship and ongoing support are potentially more relevant aspects of the Teach for All training model than its length and its contribution to research and development of methods in education.

Placement of Participants
Teach For All partners do not run schools but place their fellows in already functioning schools.This is, perhaps, the most challenging of the three aspects we review in this paper because it requires the cooperation of other actors that are not always aligned with their approach and programmatic goals.
The first challenge involves identifying schools that willingly accept incorporating Teach For All fellows and agree giving them a space in front of the classroom.There is significant variation in how this happens across countries, reflecting differences in the institutional context.Some educational systems are highly centralized and most of the decisions are made by the central government (e.g.Malaysia); others are decentralized and give individual schools more rights (e.g. the United Kingdom); while yet others have certain levels of regional or state autonomy (e.g.Mexico).
Teach For All partners have been more successful in placing participants when they obtain the full support of the relevant education authorities.A case in point is Malaysia, where the national government fully endorses the approach and even pays for the fellows' training.In the other extreme of the decentralization spectrum, schools in England have complete autonomy to choose who they hire and thus individual schools have significant flexibility to employ Teach For All fellows as teachers, creating a fertile ground for the program.
Another critical aspect influencing placement is the extent to which the existing legal framework allows graduates of disciplines other than education to teach in schools.For example, both in England and Malaysia, the legal framework makes it possible for someone that has not majored in education to obtain a certification while teaching -thus facilitating the placement of Teach For All fellows.In many other countries, however, legal frameworks are more rigid making it very difficult for fellows to be hired as teachers in public schools.This had led many Teach For All partners to either place fellows in private schools or seek creative ways of working around these rigid norms, for example by placing them as additional resource teachers or aides, as in the case of Mexico or Spain.
The second challenge is often handling the potential resistance from school staff to what could be perceived as an intrusion by 'outsiders'.Indeed, the interviews identified that it is not unusual for fellows to encounter resistance by other teachers (or even the principals) when they first join the school.However, in most cases, they argue that after a couple of months this resistance weakens or even disappears.
Partners have adopted different strategies to cope with this type of problem at the school level.Some have changed the narrative of the program, presenting it as a leadership program instead of a teacher support program, as in the case of Enseña por México.Others have decided to hold informational sessions in the schools to introduce fellows to other teachers, presenting their fellows as additional resources for schools -instead of teachers -as in the case of Empieza por Educar in Spain.In some countries Teach For All has conducted surveys of principals and teacher at the end of the school year to evaluate the placement of the fellows in the schools, further contributing to create a more welcoming environment for future placements.
Not all partners, however, have been negatively received in schools: in some locations where there are teacher shortages, fellows have been welcomed.EnseñaPerú, for example, serves multigrade schools in rural areas.For the sole teacher in school, receiving an EnseñaPerú fellow is a welcome development.
Even though the difficulties Teach For All experiences at the placement stage are specific to their modus operandi, they are also illustrative of both the complexity of policies that seek to affect the distribution of teachers across schools and the strategies that may be needed to implement such policies in an effective way.
Education authorities have the formal powers that Teach For All lacks but often experience the same limitations to use the placement of teachers in a strategic way whether due to rigid personnel rules, resistance from individual teachers to be assigned to less desirable positions, or difficulties in group dynamics at the school level when placement is felt to be unfair or nontransparent.Moreover, schools might find the placement of TFA fellows for only two years detrimental, given the associated administrative costs of incorporating new teachers, and the negative effect this turnover may have on student's learning and tracking (Darling-Hammond & Sykes, 2003;Donaldson & Moore Johnson, 2011).It is also important to note that the placement of well-prepared teachers is insufficient to generate lasting systemic improvements in education if other compensatory measures are not undertaken as well (Darling-Hammond & Sykes, 2003).

Lessons Learned
In this paper we set up to learn lessons from the Teach For All experience that could serve informing the design of teacher policies at the system level, either by education authorities, by schools and/or by teacher training institutions.Teach For All partners have accumulated years of experience in recruiting, selecting, training, and placing talented college graduates and young professionals as teachers in high-need schools.Our analysis of their experience reveals a number of good practices and innovations that we find valuable contributions to ongoing efforts seeking to improve the teaching profession in Latin America.We review these lessons not to suggest that the practices of a particular program ought to be scaled up and replicated but, rather, to argue that such practices could help inspire and inform the design of reforms at the systems level.

Active Recruitment and Rigorous Selection at Teacher Training Programs
As discussed above, the status of the teaching profession in Latin America is low and teaching is not considered to be a prestigious job career.A majority of those that study to become teachers are among the lowest performing students in their cohorts (Bruns & Luque, 2014).In contrast to what happens in Latin America, the best performing education systems in the world make special efforts to recruit top talent to become teachers (Barber & Mourshed, 2007;Mourshed, Chijioke, & Barber, 2010).
The Teach For All experience suggests that attracting stronger candidates to teacher training programs may require a combination of a more active approach to recruitment and more rigorous selection methods.It also suggests that 'rigor' should not be interpreted solely as holding candidates to high academic standards but also having demonstrated the kind of interpersonal skills required to become an effective teacher.
Setting a higher bar for entry and pro-actively seeking good candidates will typically require actions both by education authorities -that regulate and oversee teacher training institutions and thus can influence admission policies -and by the institutes themselves -that may need to develop outreach and evaluation capacities most of them currently do not have.

Quality Over Length in Teacher Training Programs
There is increasing recognition that teacher education in Latin America is in need of a shakeup to achieve higher levels of quality (Bruns & Luque, 2014;OREAL & UNESCO 2012).The Teach For All experience provides hints of some key elements that such a shake-up should consider.The experience of the Summer Institutes -short, intense, and personalized; with a strong emphasis on practice and on evaluation and feedback -could serve as inspiration for the design of teacher training programs.This is not to say that they provide a blue print.Rather, that many of their practices follow principles that teacher training institutions could apply.
Intensity appears to be more important than length.Quality involves learning pedagogical and technical content as well as the intensive practice of teaching from the early stages of the training process.Practice must be accompanied with evaluation, feedback and analysis for teachers to identify and understand their strengths and weaknesses.These are principles that can be applied in different ways by teacher training institutions, even without formal changes in regulations and policies -even though such changes could provide the necessary encouragement for those institutions to apply them.

Support, Support, Support
Perhaps one of the most innovative aspects of Teach For All's approach is the strong emphasis on mentorship for participants both during the training phase and throughout the duration of the two-year teaching experience.In all our interviews this was identified as, perhaps, the most fundamental aspect of their approach to delivering quality teaching.Classroom observations from, and regular meetings with, experienced teachers are core components of Teach For All's model.Building opportunities to reflect about the teaching experience -starting from day one -is also an integral part of their approach.
As the TALIS results indicate, mentorship and support are clearly weak areas for education systems in Latin America.Establishing a formal support mechanism for teachers in schools is crucial for their performance in the classrooms, particularly in the case of novice teachers.Institutionalizing a coaching or tutor system can be important for teachers' professional development.While individual schools must be willing participants in such efforts, this is a function that demands an active role for education authorities.

More Evaluation and Data
From recruitment, to training and support, the Teach For All approach puts an emphasis on measurement and evaluation as a crucial element for ensuring good teaching practices.The decision of who enters the program is based on the analysis of data collected at the application stage.Fellows are monitored intensively throughout the Summer Institutes and that information is used to devise tailored strategies to help them improve.And data -including from students' outcomes and perceptions -plays an important role in the support function when fellows are already teaching.
The Teach For All experience highlights how the systematic collection and use of data to evaluate teachers' performance can be an asset for teachers seeking to improve their practices, schools providing targeted support to teachers that need it, and teacher training institutions adapting and improving their curricula.The development of a strong 'data culture' and the capacity to 'consume' data is another area in which there are important lessons from education systems emerging from the Teach For All approach.

Diversify Paths to Teaching
The Teach For All experience suggests that traditional routes to teaching are not the only way to bring effective teachers to the profession.At a minimum, their experience serves as a demonstration effect of one alternative: attracting skilled young professionals that did not go through a standard teacher-training program.Reviewing the Teach For All experience it is hard to avoid thinking about the severe losses of systems that close the doors to potentially good teachers only because they did not decide to study in a teacher training institution.This does not mean opening the door to unqualified candidates.Instead, it implies that certification rules can be made more flexible to allow candidates to demonstrate they are qualified without having to go back to school and obtain a new degree.
The Teach For All experience also raises the possibility that schools could benefit from opening the door to effective teachers that may not be interested in making teaching a life-time career.In most countries, however, regulations (particularly on paths to teacher certification) make such diversification difficult -if not impossible.This appears to be one key area in which policy innovation can have significant payoffs.

Conclusion
In this paper we reviewed the international experience of the Teach For All network and asked what can policy-makers in Latin America learn from that experience that is of relevance when crafting teacher policies at the system level.
Based on that analysis we argue that Teach For All's active recruitment and rigorous selection of candidates, emphasis of quality over length in training, the continued support offered to teachers and the heavy emphasis on data and evaluation constitute practices that, if pursued at the systems level, could enhance the quality of teaching and, consequently, learning outcomes in Latin America.

III. Placing Participants As Teachers
• What are the main institutional barriers to place participants as teachers?Has the current legislation, professional associations (teachers) or schools' organizational cultures been an obstacle for participants' placement?Touch on teacher certification.• How do you encourage schools to participate and contribute to the program and with the new teachers?• The public sector has supported or hindered the program?
• What has been the program's action to overcome the problems?

Table 1
Teach For All Partners