Piaget states that learning occurs when there is interaction with others, and Vygotsky added, especially when “these others” (mediators) possess slightly more knowledge than the learner (Vygotsky, 1978). Seeing the teacher as the mediator, traditional teaching methods assumed that peer interaction was of little relevance and therefore should be eliminated, or at least minimized, in the classroom. However, for some time now we have realized that when interaction between pupils is appropriately structured, it leads to peer learning.

A wide range of peer learning scenarios exist. The distinction that Damon and Phelps (1989) proposed is a classic example. This situates peer learning on a continuum with three main scenarios. Based on the participants’ characteristics, objectives, and type of interaction, the authors distinguish between peer tutoring (an asymmetric relationship between two pupils with different levels of skill or knowledge in a given area), cooperative learning (relationships based on the acquisition and/or application of knowledge in a group with heterogeneous, or near heterogeneous, skills), and collaborative learning (a relationship based on the acquisition and application of knowledge between two or more students with similar abilities).

Although the authors conceived of these three scenarios as a continuum, in practice the distinction is sometimes unclear. While the distinction between peer tutoring and cooperative learning is transparent because of the asymmetry existing between the role of the tutor and tutee, the distinction between collaborative and cooperative learning is less clear and generates controversy. The degree to which interactions between the team members are structured can distinguish both scenarios (McWhaw, Schnackenberg, Sclater, & Abrami, 2003). Cooperative learning constitutes a more structured scenario in which the teacher’s role is that of a facilitator who organizes relationships between team members to help them reach their objectives (Davidson & Major, 2014). Collaborative learning, on the other hand, gives pupils greater freedom to organize themselves as a group, as happens ad hoc in informal learning or professional contexts in which the members normally possess sufficient social skills and motivation to jointly solve a real problem (Topping, Buchs, Duran, & Van Keer, 2017). In formal education, especially within schools, pupils are beginning to learn how to cooperate but are still developing the complex social skills required to work in teams created artificially in order to complete tasks that they may sometimes perceive as unrelated to them.

In this context, you cannot group pupils together and hope that cooperation will simply occur. The teacher must provide structured frameworks that will help create learning through pupil interaction.

The success of peer learning, and especially cooperative learning (CL), is documented in hundreds of studies ranging from preschool to university level and in different curricular areas (Johnson & Johnson, 2009). Authors (e.g., Dillenbourg, Baker, Blaye, & O’Malley, 1996; Rodríguez, Fernández, Escudero, & Sabirón, 2000) have distinguished at least three generations of research.

The first generation compared how effective cooperation was in relation to competition and individual work. These studies show that cooperation is more effective than interpersonal competition or individual effort at all levels of education (Johnson, Maruyama, Johnson, Nelson, & Skon, 1981) and has powerful effects on achievement, socialization, motivation, and personal development (Johnson & Johnson, 2002). Better social relations are built because of it (Roseth, Johnson, & Johnson, 2008). Synthesizing two major contributors (Johnson & Johnson, 1989; Slavin, 1983), we can conclude that although cooperation is a powerful tool in the learning process, its benefits are not automatic.

Simply putting students into groups to work cooperatively is not enough (Baines, Blatchford, & Kutnick, 2008). Group work, especially within a school context, can be characterized by the dissipation of responsibility, with one pupil working harder to compensate for, or sometimes hinder, the work of the others (Lindauer & Petri, 1997). To overcome these disadvantages and change group work into teamwork, or cooperative work, the teacher must organize the interactions between team members to ensure that the five principles proposed by Johnson and Johnson (2009) are fulfilled. These principles are Positive Interdependence, Individual Accountability and Personal Responsibility, Promotive Interaction, Appropriate Use of Social Skills, and Group Processing. These are widely accepted by the scientific community, with a few minor variations (Kagan & Kagan, 2009; Villa, Thousand, & Nevin, 2013).

In order to help structure interaction so that groups become teams, cooperative structures have emerged. These structures can be divided into the methods and techniques or formal and informal CL (Johnson, Johnson, & Holubec, 2008). Methods are viewed as complex and sophisticated structures, which usually require time and initial training periods for the students; implementing them regularly is advised. Learning Together (Johnson & Johnson, 2002), Group Investigation (Sharan, 1994), Reciprocal Teaching (Palincsar & Brown, 1984), or Jigsaw (Aronson, 1978; Slavin, 1980) stand out among the multiple methods available. Techniques, on the other hand, are simple structures that can be implemented with a few easy steps, without the need for initial training (Kagan & Kagan, 2009).

The second generation of research on CL focuses on how effective CL methods are. Slavin (1980) reviewed studies on the use of CL in primary and secondary education. Results showed improvements in reaching goals, more positive interracial relationships, mutual interest, and greater self-esteem. Newmann and Thompson (1987) obtained similar results when comparing CL methods at the secondary school level and emphasized the necessity of rewards or group recognition, as well as individual responsibility. Slavin (1996) also stresses the effect of group rewards. Nastasi and Clements (1991) reported the importance of higher-order thinking, motivation, and social skills to explain the benefits of CL for cognitive development, academic performance, and socioemotional development.

Other meta-analysis (Lou et al., 1996) shows that learning in small groups has positive effects on performance and attitudes in both primary and secondary education. Positive results of CL methods are consistent at all levels of education (Vermette, Harper, & DiMillo, 2004). Cohen, Kulik, and Kulik (1982) were the first to provide a meta-analysis on peer tutoring, concluding that greater effectiveness is connected to the initial training the tutors receive, the distance inability between tutor and tutee, structured interaction, and longer-term projects. Bowman-Perrott et al. (2013) analyzed 26 case studies in schools and found peer tutoring to be effective. Leung (2015), after reviewing 72 papers, concludes that peer tutoring is effective for all age groups, especially those in secondary education.

With the power of cooperation and the effectiveness of CL methods proven, the third generation of research is focusing on the interaction process within teams and analyzes the exchanges and joint activities between team members, attempting to explain how and for what reasons students learn by cooperating together.

Unfortunately, there is a large number of these third-generation studies and they cannot all be covered here. Undoubtedly, more meta-analyses and reviews will be published soon. This will encourage theories to emerge, explaining the elements responsible for the effectiveness of CL.

Notwithstanding, we would like to offer some explanations that the studies on peer tutoring have provided and that can be generalized to CL. On the one hand, there is the possibility to adjust the pedagogical help that one-to-one or very small group relationships allow. In this privileged interaction, in which the teacher has deliberately structured the prototype dialogue between the teacher and the group of students—referred to as IRF—is divided into three phases: teacher initiates dialogue (I), students respond (R), and teacher offers feedback (F). In CL, this dialogue is enriched by one of five phases: IRFCE (Person & Graesser, 1999). In this interaction, both tutor and tutee follow the IRF structure but, because there is cooperation (C), responses from both pair members may improve and the quality of this joint response is what the tutor finally evaluates (E). The cooperation phase is the key to explaining the effectiveness of peer tutoring and its potential advantages. Enriching interactive mechanisms, some more tutorial in nature (led by the tutor) and others more collaborative (led by both tutor and tutee), emerge in this one-to-one scenario, when tutors display their skill and pedagogical strategies (Duran & Monereo, 2005).

This explanation allows us to understand how the pupil receiving pedagogical help from his or her teammate learns. However, we need to refer now to the pupil who offers help and the possibility of learning by teaching (Duran & Topping, 2017). Research into how pupils learn by teaching their classmates suggests that for learning by teaching to occur, as Roscoe and Chi (2007) suggested, interaction must be structured and pupils must be taught how to get away from simply “telling” knowledge and move toward a constructive process of reflection by explaining and questioning things. This is a crucial element to be built into classrooms, where pupils not only learn from their teachers but also from the mutual pedagogical help they give each other.

In the classroom, not only students, but also teachers, can collaborate. If we follow the classical definition by Friend and Cook (1992), teacher collaboration can be characterized as an interaction between at least two teachers, who act as peers, with the opinion of either having the same value regardless of their role in the school. Also, although the school should promote this interaction, the teachers decide if they want to collaborate or not. Decisions will be made in order to reach a common goal in a shared, but not necessarily identical, way. This allows for the differences in teachers’ contributions, which are affected by their knowledge and/or role in the school.

A wide consensus exists in the academic field (Brown & Poortman, 2018; Hargreaves & Fink, 2006; Hargreaves & Fullan, 2012), and among studies carried out by international education organisms (Organisation for Economic Co-operation and Development, 2009; Schleicher, 2016), on the need to develop and adjust teacher collaboration in schools and collaboration networks. Teacher collaboration has been promoted as a decisive factor that contributes to school improvement and teacher development (Fulton & Britton, 2011; James, Dunning, Connolly, & Elliot, 2007). It is widely recognized that teacher collaboration is directly linked to the improvement of practices in innovative educational situations, because of the learning processes it promotes in the participants, and that have positive effects on students’ achievement (Doppenberg, den Brok, & Bakx, 2012; Meirink, Imants, Meijer, & Verloop, 2010; Ohlsson, 2013; Ronfeldt, Farmer, McQueen, & Grissom, 2015).

Professional learning communities are a good example of such collaborative practices (Little & Horn, 2007; Stoll, Bolam, McMahon, Wallace, & Thomas, 2006), which allow innovative methods to be implemented and sustained. Examples also exist of the creation of very promising networks between these communities (Katz & Earl, 2010; Miquel & Duran, 2017).

After reviewing studies on collaboration, Kelchtermans (2006) argues that joint work by teachers must be, but sometimes is not, correctly organized; its success can depend on the school context. Vangrieken, Dochy, Raes, and Kyndt (2015) present the diversity of perspectives and ways of understanding teacher collaboration on a continuum, ranging from a group of individuals who simply work together to highly collaborative teams. They underscore that the benefits are greater the more deeply rooted collaboration is among teachers (they feel more motivated, more efficient, experience a lighter workload, their communicative skills increase, and teaching strategies become more student-centered), among pupils (improving understanding and performance), and among the school (there is a positive influence in the perception that the school is supportive of innovation, making a cultural shift to more equity, greater school-wide attention to the needs of students, and a professional culture of intellectual inquiry).

Vangrieken et al. (2015) identify an array of factors that influence teacher collaboration, including those factors referring to the characteristics of the participants (positive attitudes toward teaming, experience, and knowledge of team processes), group structure characteristics (staff continuity, meeting time, and including them in decision-making processes), characteristics of working together (strong sense of team community, relationship-building, clear direction and common goals, observation and discussion of each other’s practices, and self-management), and characteristics of the school and the support received from the school (school administration’s commitment to teaming, pre-service, and in-service training). However, they also identify teachers and schools strongly opposed to collaboration. Some factors that hinder teaching collaboration are a lack of skills or training, unwillingness to collaborate, balkanization, individualism in school culture, and ineffective leadership (Plauborg, 2009). In conclusion, Vangrieken et al. (2015) define two basic aspects of effective collaboration: the work process supported by the school principal and the outcomes obtained from effective collaboration.

These studies do not dedicate attention to the differences between divergent collaborative teaching styles. We find styles of indirect collaboration (using the classical terminology of Bauwens & Hourcade, 1995) that examine collaborative relationships between teachers in planning, preparing, and discussing aspects related to teaching-learning; each teacher then implements it into practice individually. We will now focus on cooperative teaching or co-teaching, the most genuine form of direct collaboration.

Co-teaching is defined as a situation in which teachers, normally two, share responsibility for educating the whole class, within an inclusive classroom, by designing a program adapted to everyone, implementing agreed upon procedures, and reviewing the results obtained (Friend, Cook, Hurley-Chamberlain, & Shamberger, 2010; Murawski & Dieker, 2008). The teachers who collaborate are usually general- and special-needs teachers, but any teacher, even the students in the school, can adopt the role of co-teacher (Villa, Thousand, & Nevin, 2010).

Most of the studies on co-teaching focus on describing the different roles and relationships between teachers or other organizing aspects that are established in the classroom, rather than demonstrating its impact on student achievement and other key outcomes (Friend et al., 2010). These studies insist, however, on the benefits both participating teachers and pupils gain from co-teaching.

With reference to pupils, research on the efficacy of co-teaching considers different dimensions, such as the improvement in learning, social skills, and behavior, especially among pupils with learning difficulties (Hang & Rabren, 2009; Scruggs, Mastropieri, & McDuffie, 2007; Strogilos & Stefanidis, 2015). Co-teaching is proving to be a useful strategy in the advancement of inclusive education (Pancsofar & Petroff, 2016; Villa, Thousand, & Nevin, 2010). This is for a variety of reasons. First, it supposes that support is available to those who require it constantly, as well as to those who require support only occasionally. A second reason is that support can be given in all areas and in different content, not just in specific or isolated skills, which usually occurs when support is provided outside the classroom. Additionally, co-teaching makes it easier to assess the pupil’s needs and to provide the best support at a particular moment. And finally, co-teaching makes it easier to use teaching resources, which might otherwise be impossible if teaching was done individually. These resources may include strengthening oral expression, monitoring the class by focusing on the pupils, or making video recordings to analyze afterward. All these situations, demonstrating the use of co-teaching in the classroom, are designed to favor greater learning among the pupils.

Co-teaching also benefits teachers; it allows them to share and design new materials and methodologies, share the processes of assessment and monitoring, offer mutual support in difficult situations, and favor classroom management and the working atmosphere (Nevin, Thousand, & Villa, 2009). One of the most beneficial aspects that co-teaching offers, which comes from constructive discussions together, is that it acts as a strategy for which teachers can build knowledge (McDuffie, Mastropieri, & Scruggs, 2009; Pratt, 2014). Thus, co-teaching should be seen not only as a form of support for students but also as a source of learning for the teachers involved—and therefore an essential element in promoting improvement in the school (Pancsofar & Petroff, 2013; Rytivaara & Kershner, 2012).

Recent studies on the implementation of co-teaching as a model for the teacher education practicum also illustrate these benefits for both student teachers and their mentors (Baeten & Simons, 2014; Guise, Habib, Thiessen, & Robbins, 2017), and is thus an optimal strategy in pre-service and in-service professional development programs.