Co-Teaching Models for Inclusion

Co-teaching is a service delivery model in which two professionals, typically a general education teacher and a special education teacher, share responsibility for planning, delivering, and assessing instruction for a diverse group of students in a single classroom. The definition most widely used in research comes from Friend and Cook (1996): co-teaching requires that both professionals have joint instructional responsibility for a heterogeneous group of students, including those with disabilities, in a shared physical space. That definition rules out a lot of what schools informally call "co-teaching. For more on this topic, see What is inclusion." A teacher and a paraprofessional are not a co-teaching pair. A special educator who dips in and out of a general education class is not co-teaching. The two-professional, shared-accountability standard is the starting point, and everything else follows from it.

‍

‍

IDEA made co-teaching popular in the US. This law, Individuals with Disabilities Education Act, states learners with needs learn in the least restrictive setting. For most learners, this means the general classroom. Co-teaching, like what Friend and Cook (2013) describe, delivers support inside that class. It avoids removing learners, as Murawski and Dieker (2013) found.

‍

What Co-Teaching Actually Is

Friend and Cook (1996) say co-teaching has three key parts. First, it needs two qualified teachers, not an aide. Special educators bring vital skills, like disability knowledge. They also know individual instruction and assessment under IDEA. Aides cannot give specialised instruction legally. Using them as a second teacher hurts learning.

Second, co-teaching involves joint delivery of substantive instruction. Both professionals must be teaching, not one teaching while the other manages behaviour or copies notes onto the board. If the special educator spends most of the lesson circulating quietly and correcting work, that is not co-teaching by any research definition, and it will not produce the outcomes associated with co-teaching in the literature.

Third, co-teaching serves a heterogeneous group that includes students with disabilities. It is not a model for general ability grouping or team teaching among general educators. Its purpose is to bring the expertise of special educational needs provision into the general education environment so that students with IEPs receive their services in the LRE.

Co-teaching is also distinct from two related but different models. Push-in refers to a specialist coming into the general education classroom to work with specific students, typically a small cluster of students with IEPs, while the general education teacher continues with the whole class. This is not co-teaching because there is no joint instructional responsibility for the whole group. Pull-out refers to removing students from the general education setting to receive services in a resource room or separate space. Pull-out remains an appropriate model for students whose IEPs require it, but it is at a more restrictive point on the LRE continuum. Co-teaching sits between these extremes: it keeps students in the general education classroom while ensuring that the specialised expertise of the special educator is present for everyone.

‍

The Six Co-Teaching Models

Friend and Cook identified six distinct models of co-teaching, each with different structures, different demands on teachers, and different benefits for students. Understanding all six and knowing when to use each is the practical core of effective co-teaching.

‍

One Teach, One Observe

In this model, one teacher delivers instruction to the whole class while the other systematically observes and collects data. This is not passive monitoring. The observing teacher should have a defined observation purpose, a data collection tool, and a debrief plan. What are they watching for? It might be how many students respond to open questions, which students never volunteer, how long a student with ADHD sustains attention before going off task, or whether a student with a reading IEP goal is tracking text during shared reading.

One Teach, One Observe is most useful at the start of a co-teaching relationship, when both teachers are still building a shared understanding of which students need what. It is also useful at specific diagnostic moments during the year, when both teachers want an accurate picture of classroom functioning before making decisions about grouping or intervention.

A concrete example: a Year 8 science class is learning about states of matter. The general education teacher leads the lesson while the special educator uses a structured observation form to track which students answer questions correctly on the first attempt, which students answer after a prompt, and which never respond at all. That data goes directly into the planning conversation for the following lesson, where the special educator can suggest which students need additional pre-teaching before the next topic.

One Teach, One Observe is not a sustainable everyday model and should not become the default. Its value is in the quality of data it generates, not in its instructional power.

‍

One Teach, One Assist

In this model, one teacher delivers instruction to the whole class while the other circulates, providing quiet support to individual students as needed. This is the most commonly observed model in co-taught classrooms, and it is also the most frequently misused. Scruggs, Mastropieri and McDuffie (2007) conducted a qualitative meta-synthesis of 32 co-teaching studies and found that One Teach, One Assist was overwhelmingly the dominant model in observed classrooms, with the special educator almost always in the assistant role.

The problem is not with the model itself. It is appropriate when the general education teacher is introducing new content for the first time, when a student with a specific support need requires brief in-the-moment help, or when the lesson structure does not lend itself to splitting the class. The problem arises when it becomes the only model used, because it positions the special educator as a glorified aide, makes students with IEPs visible as the recipients of that assistant's attention, and removes any instructional contribution from the co-teacher's expertise.

Used correctly, One Teach, One Assist looks like this: a maths class is working through a new multi-step problem. The general education teacher explains the procedure at the front. The special educator circulates, watching for students who have stalled at step two, quietly rereading the prompt to a student with a processing difficulty, or redirecting a student with executive function challenges back to their work. The circulating teacher is not interrupting the flow of instruction. They are providing responsive, just-in-time support that keeps students on track without public identification.

The risk to monitor is whether students with IEPs learn to wait for the special educator rather than attempting work independently. If the same students receive assistance every lesson, the model is reinforcing dependence rather than building capability.

‍

Station Teaching

In station teaching, the class is divided into three or more groups, each rotating through learning stations at intervals. Both teachers lead one station each, and at least one station involves independent or collaborative work without direct teacher involvement. This model reduces the teacher-to-student ratio substantially and gives both professionals genuine instructional roles with equal weight.

Station teaching works when content splits easily (Good & Brophy, 2000). Learners gain from varied activities in one lesson. It aids classes with mixed readiness (Tomlinson, 2014) where whole-class teaching is hard.

A concrete example: a Year 6 English lesson on figurative language. Station one is run by the general education teacher and focuses on identifying figurative language in a poem. Station two is run by the special educator and focuses on students writing their own examples of two figurative language types. Station three is an independent station with a card-sorting activity matching figurative language terms to definitions. Each group of eight students spends fifteen minutes at each station. The special educator's station can be calibrated for additional scaffolding without identifying specific students as "the group that needs help," because all groups cycle through all stations.

Station teaching requires more planning than One Teach, One Assist. Both teachers need to design their stations in advance, agree on the content focus of each, and establish clear transition routines. The payoff is a lesson where both professionals are genuinely teaching and where student groupings can be adjusted lesson by lesson based on observed need.

‍

Parallel Teaching

In parallel teaching, the class is divided into two roughly equal groups, and each teacher delivers the same instructional content simultaneously to their group. The content is the same. The approaches may differ. The benefit is a halved student-to-teacher ratio, which increases the frequency of student response and the teacher's ability to monitor understanding.

Parallel teaching helps discussion lessons where thirty learners struggle to speak. Review and practice sessions benefit too, letting learners get feedback often. General and special educators can vary teaching, keeping content the same (Cook & Friend, 1995).

The key design principle is that both teachers must teach the same learning objective. The differentiation is in the mode of delivery, the examples used, or the level of scaffolding provided, not in what students are expected to learn. A student with an IEP who receives parallel teaching with the special educator should be working toward the same grade-level standard as their peers, with access to whatever accommodations and modifications their IEP specifies.

A concrete example: a Year 9 history class is reviewing causes of the First World War. The general education teacher takes half the class through a discussion-based analysis of the alliance system. The special educator takes the other half through the same analysis using a graphic organiser framework that provides more structure for students who need it. Both groups are working on the same historical thinking skill. Fifteen minutes in, both groups have produced analysis they can share with the other group.

Scheduling, tight timetables and classroom space can make co-teaching hard. Parallel teaching, say researchers (Friend, 2008), needs a big room. Two groups must learn without too much noise. Both teachers should feel happy leading on their own (Cook & Friend, 1995).

‍

Alternative Teaching

In alternative teaching, one teacher works with a small group of students for a targeted purpose while the other teacher works with the remaining, larger group. The small group might be receiving pre-teaching, re-teaching, enrichment, or assessment. This model addresses the reality that in any class, some students are ready to move forward, some need more time with current content, and some have gaps from previous learning that are blocking progress.

Alternative teaching is the model most directly connected to tiered support within an MTSS framework. The small group pulled by the special educator is not necessarily always students with IEPs. It might be students who struggled on last week's exit ticket, or students who are ready for an extension task, or a rotating group receiving strategic intervention on a specific skill.

A concrete example: a Year 7 maths lesson on fractions. Exit data from the previous lesson showed that six students cannot yet reliably convert between improper fractions and mixed numbers. At the start of the lesson, the special educator takes those six students to one side of the room for a fifteen-minute re-teaching session using manipulatives, while the general education teacher leads the rest of the class through the day's new content: adding and subtracting fractions with different denominators. The six students then rejoin the class for the independent practice phase. No student has missed the new content. Six students have received targeted intervention that would otherwise require a pull-out session.

The risk is that the same students are always in the small alternative group, which re-creates the stigma of pull-out inside the general education classroom. The composition of the alternative group should rotate, and the purpose of the small group should vary across the week.

‍

Team Teaching

In team teaching, both professionals teach the whole class simultaneously, sharing delivery, discussion facilitation, and instructional decision-making in real time. One teacher might be explaining a concept while the other models a worked example on the board. One might ask a higher-order question while the other listens to student responses and follows up. One might notice a widespread misconception mid-lesson and signal to the other to pause and address it.

Team teaching requires the highest level of trust, communication, and planning of all six models. Both teachers must be genuinely comfortable with the content, comfortable with each other, and comfortable departing from a script when the lesson calls for it. Structured questioning strategies work especially well in team teaching because one teacher can ask while the other listens and probes student responses. It takes time to develop. New co-teaching pairs rarely manage genuine team teaching in their first semester together.

When it works, team teaching is the most powerful model because students benefit from two expert perspectives integrated seamlessly. The special educator is not positioned as the helper. Both teachers are positioned as co-authorities on the content. Students with IEPs receive the implicit message that their teacher knows the subject matter, not just "the special education stuff."

A concrete example: a Year 10 biology lesson on natural selection. The general education teacher explains the mechanism of natural selection while the special educator draws a visual timeline of generations on the board. Mid-explanation, the special educator asks a clarifying question: "So when you say 'survival of the fittest,' do you mean the physically strongest individual?" The general education teacher unpacks the misconception that the question surfaces. The special educator follows up with a student who shrugged: "What do you think? Strongest or best adapted?" The exchange feels like a genuine intellectual dialogue. Students see two professionals thinking through content together, which models the kind of academic talk they are expected to produce.

‍

Choosing the Right Model

The most common mistake co-teaching pairs make is treating the choice of model as fixed, either defaulting permanently to One Teach, One Assist or rotating models by day of the week regardless of what the lesson needs. Model selection should be driven by three questions: What does this lesson require structurally? What do these students need right now? What are both teachers comfortable delivering?

Model selection should also account for where you are in the unit sequence. New content introductions often suit One Teach, One Observe or One Teach, One Assist. Mid-unit practice suits station teaching or parallel teaching. Pre-assessment or re-teaching suits alternative teaching. End-of-unit review suits parallel teaching or team teaching. This is not a rigid formula, but it gives both teachers a starting framework.

‍

Co-Planning: The Non-Negotiable

Every piece of research on co-teaching effectiveness points to the same finding: co-teaching without co-planning produces little or no benefit over single-teacher instruction. Murawski and Swanson's (2001) meta-analysis of seventeen co-teaching studies found effect sizes that ranged from negligible to substantial, and the variable most consistently associated with strong outcomes was the presence of structured, joint planning time. Schools that give co-teachers a weekly protected planning period consistently outperform schools that tell co-teachers to "plan when you can."

Planning time has a specific content. It is not social catch-up, and it is not one teacher briefing the other on what the lesson will cover. Effective co-planning addresses at least five questions for each lesson or lesson sequence.

What is the learning objective, and what does mastery look like for students at different readiness levels? This question requires the general education teacher to articulate the content standard and the special educator to identify where students with IEPs or other learning needs are relative to that standard, and what accommodations or modifications will allow them to access it.

What is the lesson structure, and which co-teaching model fits it? Model selection happens in planning, not at the classroom door. Both teachers need to know their role before the lesson begins.

What does each teacher specifically do at each phase of the lesson? "We'll use station teaching" is not a plan. "I lead station one on paragraph structure using the graphic organiser we agreed; you lead station two on revision using the mentor text; station three is independent writing" is a plan.

Which students need what, and how will we know? This is where the special educator brings IEP knowledge to bear. Which students have accommodations for extended time, reduced visual complexity, preferential seating, or verbal rather than written response? How will the co-teachers coordinate delivery of those accommodations in a way that does not single students out? Understanding how working memory capacity affects a student's ability to process instruction simultaneously with completing a task is relevant here, since many accommodations, such as chunked instructions and reduced written output, exist precisely to manage working memory demand.

How will we assess understanding at the end of the lesson, and what do we do with that information? Formative assessment strategies in a co-taught class can be more powerful than in a single-teacher class because two teachers can each observe half the room simultaneously, or because the special educator can circulate and gather verbal responses from students who do not respond in writing.

A co-planning template that both teachers complete in advance of the meeting is more efficient than open-ended discussion. Dieker (2001) developed a lesson plan format specifically for co-taught classrooms, which uses a two-column structure showing what each teacher does at each phase of the lesson. Using a shared template keeps planning focussed and ensures both teachers leave with the same understanding of who does what.

‍

Making the Co-Teaching Relationship Work

The instructional quality of a co-taught class is inseparable from the quality of the relationship between co-teachers. A skilled pair of educators who have built genuine trust and mutual respect can make even an imperfect model work well. A pair with unresolved tensions about roles, expertise, and authority will underperform regardless of which model they use.

The most persistent structural problem in co-teaching relationships is the power imbalance between the general education teacher and the special educator. General education teachers typically control the content, the pace, and the classroom. They are often more senior in building seniority or more familiar to the students. Special educators are often perceived, and sometimes treated, as guests in the classroom. This perception undermines the model at its foundation.

Friend (2008) describes "parity" as the observable indicator of a genuine co-teaching relationship. Parity means both teachers have an equal presence in the classroom. Both names appear on the door, the whiteboard, and parent communications. Both teachers have a desk or working space in the room. Both address all students and are addressed by all students. Both teachers participate visibly in routines like homework collection, behaviour management, and transitions. When parity is absent, students quickly identify the "real teacher" and the "helper teacher," and the special educator's authority and effectiveness drop accordingly.

Handling disagreements professionally is part of co-teaching competency. Both teachers will have different views on pace, on how much scaffolding is too much, on whether a student needs more challenge or more support. Disagreements that are worked out in planning conversations are healthy and productive. Disagreements that play out in front of students, or that are suppressed until they become resentments, are destructive. Establishing a norm early in the partnership, such as signalling a need to talk with a mutually agreed phrase and then following up outside the classroom, prevents public conflict while keeping both teachers honest.

Collaboration requires mutual respect. General teachers should trust special educators on disabilities (Cook & Friend, 1995). Special educators trust general teachers on curriculum (Idol, 2006). This model succeeds by merging different expertise (Murawski & Dieker, 2013). No single teacher knows it all.

‍

What the Research Says

The research base on co-teaching is substantial, but it carries important qualifications. Murawski and Swanson's (2001) meta-analysis remains the most cited quantitative synthesis. They analysed seventeen studies that met their inclusion criteria, which included a control group, a quantitative outcome measure, and a description of the co-teaching model used. The mean effect size across studies was 0.40, which is educationally meaningful, equivalent to roughly four to five months of additional learning. However, the effect sizes varied dramatically across studies, from near zero to above 1.0, and the authors noted that the variability was largely explained by implementation quality. Co-teaching with planning time, training, and administrative support produced large effects. Co-teaching without those conditions produced effects near zero.

Scruggs, Mastropieri, and McDuffie (2007) reviewed 32 co-teaching studies. They found "One Teach, One Assist" was common, with special educators assisting. Teachers wanted co-teaching but faced barriers; joint planning was rare. Leaders thought simply placing teachers together sufficed. The researchers concluded co-teaching's reality often fails to meet its promise.

Conderman and Hedin (2012) found headteacher backing is key for successful co-teaching. Support includes scheduled joint planning, training on all six models, and classroom observations for feedback. Headteachers should tell staff both teachers share responsibility for learner results. Like Rosenshine's Principles, effective co-teaching needs careful planning, not improvisation.

Dieker and Murawski (2003) say co-teaching quality is both a process and an outcome. Poor partnerships make teachers quit, hurting learners. Schools should match co-teachers, train them in communication, and provide coaching. This improves co-teacher retention and learner results.

‍

Co-Teaching and IEP Implementation

Co-teaching is a service delivery model, not a placement. Students with IEPs who are educated in co-taught classrooms are still entitled to all of the services, accommodations, and modifications specified in their IEPs. The co-teaching structure is how those entitlements are delivered in the general education setting. This distinction matters because it is easy for co-taught classrooms to drift toward whole-class instruction that does not specifically address the individual needs of students with IEPs.

Special educators must monitor IEP implementation. This includes consistent extended time (Wiggins and McTighe, 2005). Check learners get reduced assignments if needed. Ensure reading accommodations, like text-to-speech, happen in lessons (Vaughn and Fuchs, 2003). Measure progress towards IEP goals at the correct frequency (Friend and Cook, 2010).

Specialised instruction (SDI) needs careful planning in lessons. SDI isn't just help for struggling learners. It's a planned method, based on evidence, matching each learner's needs (Cook & Friend, 1995). In co-teaching, SDI can be small group teaching, pre-teaching vocabulary, or systematic phonics (Murawski & Swanson, 2001; Vaughn et al., 2003). Its form changes, but it must always be systematic and intentional.

Co-teaching allows frequent, detailed progress checks. The special educator can gather data during independent work (Stecker et al., 2005). They can collect writing or observe behaviour (Deno, 2003). Use this data for planning, IEP reports, and placement decisions (McCook, 2006).

Teachers should know 504 plans and IEPs, as co-taught classes include learners with both. Obligations for each plan differ. 504 plans provide accommodations, but not special instruction. The 504 coordinator, not a special educator, monitors this plan.

‍

Common Pitfalls

This creates what Cook and Friend (1995) call the "glorified aide" trap. When special educators act as assistants, co-teaching fails. Learners with IEPs are present but receive a poor education. Walther-Thomas (1997) found neither teacher is truly engaged with their learning.

The structural default to One Teach, One Assist is the mechanism by which the glorified aide problem perpetuates itself. It is low-planning, low-conflict, and low-disruption, which makes it comfortable for both teachers. It also produces the weakest outcomes of any model when used as the default across the school year.

Lack of planning time is the most frequently cited barrier to effective co-teaching in both quantitative and qualitative research. When schools schedule two teachers in the same room without providing protected joint planning time, they are running co-habitation rather than co-teaching. Teachers in this situation typically describe their arrangement as superficial, stressful, and unsatisfying, and they are correct in their assessment. The fix is structural: it requires administrative commitment to scheduling.

Mismatched co-teaching partners cause issues. General teachers resisting inclusion paired with new special educators struggle, even with training (Friend & Cook, 2013). Schools should consider relationships when assigning partners, not just schedules (Idol, 2006). Poor pairings lead to attrition and lower quality co-teaching (Sweigart & Land, 2017).

No training is a systemic failure. Friend (2008) consistently emphasises that co-teaching is a professional skill that must be learned and practised. Placing two qualified teachers in a room together is not sufficient preparation. Both teachers need training in the six models, practice with co-planning tools, and ongoing coaching. Without it, teachers default to what is familiar, which for the general education teacher is solo whole-class instruction, and for the special educator is circulating support.

Learners choosing "the real teacher" flags an issue. Parity suffers when learners only ask the class teacher questions. If learners seek permission or react differently to instructions, discuss teacher roles (Friend & Cook, 2020).

‍

Making Your Next Lesson Better

Before your next co-taught lesson, both teachers should answer three questions in writing. Which of the six models fits this lesson's structure and this group's current needs? What is each teacher specifically doing at each phase of the lesson? Which IEP accommodations need to be implemented during this lesson, and who is responsible for each?

If you cannot answer all three questions before the lesson starts, the planning has not happened yet. Start there.

‍

‍

Frequently Asked Questions

‍

‍

‍

Further Reading