.svg)
Every staff meeting ends with another strategy you must try, and your feed offers ten more before you reach the car park. Some of it moves learning and some of it is noise, with little time to tell which is which.
This guide covers the strategies with the strongest evidence, what each really costs in prep, and one concrete move for your next lesson. It is written for teachers and school leaders, and it leans on the EEF Teaching and Learning Toolkit, John Hattie's synthesis and AERO rather than guesswork.
The quick answer
Evidence-based teaching strategies are classroom methods backed by high-quality research on how learning works. The strongest are feedback, metacognition and self-regulation, retrieval practice, spaced practice and explicit instruction. Across the EEF Teaching and Learning Toolkit, John Hattie's synthesis and cognitive science, these five reliably lift student progress, and most cost very little once you plan them.
Two things stand out. The highest-evidence strategies are not the most expensive to run, and the cheapest wins (retrieval and spacing) are the ones teachers skip most, because the prep recurs every single lesson. The rest of this article takes the top strategies one at a time.
What is retrieval practice and why does testing yourself improve learning?
Retrieval practice means asking students to bring information back from memory instead of re-reading or re-highlighting it. The act of recalling strengthens the memory and shows both of you what has actually stuck. It is one of only two techniques Dunlosky and colleagues rated high utility in their review of ten study methods, and Hattie's synthesis puts practice testing at around d = 0.54.
Its biggest advantage for teachers is cost. A five-question recall quiz needs no marking and adds almost nothing to your workload. Keep it low stakes so students answer honestly rather than guessing to protect a grade. Retrieval works best when it is frequent, spaced out and tied to prior content, not just the lesson you taught five minutes ago.
Practical ways to run it with zero marking:
- Open with a "do now" of five short-answer questions: three from last lesson, two from earlier in the term.
- Use mini-whiteboards or hands-up cold calling so every student retrieves, not just the volunteer.
- Have students close their books and brain-dump everything they remember on a topic, then compare to their notes.
- Turn old exam or quiz questions into starters instead of writing new content.
AERO's evidence synthesis on spacing and retrieval practice rates both against its Standards of Evidence and frames them for Australian classrooms, which makes it the cleanest AU reference to hand a faculty.
What is spaced practice and how do I schedule it across a term?
Spaced practice, or spacing, means spreading contact with a topic across days and weeks instead of massing it into one block. The same total study time produces stronger long-term retention when it is distributed, because each return forces a slightly harder retrieval. It is the second technique Dunlosky rated high utility, and Hattie puts spaced versus massed practice at d = 0.60.
Spacing costs almost nothing to run. It is a sequencing decision, not a resourcing one. The trap is that blocking a unit and moving on feels more efficient in the moment and tests better in the short term, which is exactly why teachers default to it. Planned properly, spacing turns your existing content into a review schedule rather than extra work.
A simple term structure that builds spacing in without new materials:
- Weekly review: every Friday starter revisits the current week plus one topic from two to three weeks ago.
- Monthly review: once a month, a low-stakes quiz sweeps across everything taught that term.
- Interleave related topics: alternate question types (for example fractions and decimals) so students choose the method, not just apply the one you named.
Spacing and retrieval are partners. You space the practice, and each spaced session is a retrieval session. Run them together and you get compounding returns from content you have already made.
How much does feedback actually improve student progress?
Feedback is one of the highest-leverage things you can do, and one of the most misused. In the EEF Teaching and Learning Toolkit it carries roughly +6 months of additional progress at high evidence security, and Hattie's synthesis puts it at d = 0.70, well above his 0.40 hinge point.
But the effect depends entirely on what the feedback does. Task-focused comments that tell a student exactly what to fix next lift progress. Grades and praise, on their own, do not, and can even reduce effort.
The short version: feedback only works when it names the next fix.
The workload risk is real: written feedback for every student every task is unsustainable. The evidence does not require it. Fewer, sharper, action-naming comments beat a marked pile that students glance at and file.
To get the impact without the marking marathon:
- Give whole-class feedback from a sample of books rather than marking every one.
- Name the two or three fixes that would move most students.
- Build in lesson time for students to act on comments before you mark again.
NSW CESE's What works best names effective feedback as one of its eight evidence-based practices, alongside explicit teaching, and its guidance is written for Australian schools.
What is explicit instruction and how is it different from discovery learning?
Explicit instruction means teaching new content in small, sequenced steps, showing students how to do something before asking them to do it alone, and checking understanding at each step. It is the opposite of discovery learning, where students are expected to work concepts out for themselves with minimal guidance.
For novices learning new material, the evidence favours explicit teaching clearly. AERO's explicit instruction practice guide reports meta-analytic effect sizes from 0.40 to 1.22, and Hattie puts direct instruction at d = 0.60.
The reason is cognitive load. Beginners have limited working memory, so unguided problem-solving overloads them before they have secure knowledge to reason with. Explicit instruction manages that load by breaking content into steps and building fluency before independence.
The practical shape is I-do, we-do, you-do. You model the skill, work through examples together, then release students to independent practice only once responses show they are ready. Check every student's response along the way, through mini-whiteboards, cold calling or quick quizzes, rather than reading the room from a few raised hands.
This is not lecturing and it is not rote learning. It is heavily interactive, with constant checks that tell you when to move on and when to re-teach.
What is metacognition and self-regulated learning, and how do I teach it?
Metacognition is thinking about your own thinking: planning how to approach a task, monitoring progress while you work, and evaluating the result. Self-regulated learning is the wider skill of managing your own effort and strategy.
In the EEF Toolkit, metacognition and self-regulation approaches carry about +8 months of additional progress at high evidence security, drawn from 355 studies. That makes it one of the best-evidenced strategies available and one of the cheapest, since it needs staff training rather than new resources.
A recent Toolkit refresh nudged the figure to +7 months as new studies were added, a reminder these estimates are periodically re-checked. Hattie puts metacognitive strategies at d = 0.60. The catch is that it has to be taught explicitly, not assumed.
You teach it by making your own thinking visible and then handing the routine to students:
- Model a plan-monitor-evaluate routine out loud as you tackle a hard problem, including the false starts.
- Give students a success checklist and require a self-check against it before they hand work in.
- Ask "what will you do if you get stuck?" before a task, so the strategy is chosen in advance.
- Prompt reflection after: what worked, what would you change next time?
What are worked examples and when should I use them instead of problem-solving?
A worked example is a fully solved problem, shown step by step, that students study before attempting similar problems themselves. Worked examples are most powerful when students are new to a topic, because studying a clear solution costs far less working memory than solving from scratch. Hattie puts worked examples at d = 0.37, near his hinge point, and they are a core part of Rosenshine's Principles of Instruction.
The key is not to leave students on worked examples forever. You use them to introduce a method, then gradually fade the support, removing one step at a time so students complete more of the solution themselves until they are solving unaided. Switch to independent problem-solving once they have the fluency to handle it without overload.
A simple fading sequence:
- Full worked example.
- A completion problem with the last step missing.
- Then one with the last two steps missing.
- Finally a problem to solve alone.
Alternate a worked example with a similar problem to solve, so students study then apply in quick cycles. The upfront cost is building clean, uncluttered solutions, which is the one moderate-prep strategy in the ranking below and a natural place to lean on automation.
How can teachers use these high-impact strategies without adding hours of prep?
The honest answer is that the science is not the hard part. The prep tax is.
Every high-impact move carries recurring prep:
- Retrieval practice needs fresh recall questions every lesson.
- Spacing needs old content re-sequenced into weekly and monthly reviews.
- Worked examples need clean, faded solutions built by hand.
Every one of those is recurring work, and OECD's TALIS 2024 survey again found time pressure among teachers' biggest constraints. That is precisely why teachers know these strategies work but cannot run them consistently.
The short version: the strategies work; the prep is what stops them.
The fix is not more willpower. It is removing the prep so the highest-impact moves happen by default, not on the days you have a spare hour.
The table below ranks the highest-impact strategies three ways at once: their evidence rating, their typical prep and marking cost, and one concrete move you can run in your next lesson. Impact and effort rarely line up the way teachers expect, and that gap is the whole point.
| Strategy | Evidence rating (EEF months of progress / Hattie effect size d) | Typical prep / workload cost | One concrete classroom move |
|---|---|---|---|
| Feedback | EEF +6 months (high security); Hattie d = 0.70 | Very low to set up, but time-heavy if written per student | Give one task-focused comment that names the next fix ("add a unit to your answer"), not a grade or praise |
| Metacognition & self-regulation | EEF +8 months (high security, 355 studies); Hattie metacognitive strategies d = 0.60 | Very low (staff training only) | Model and script a plan-monitor-evaluate routine, then have students self-check work against a success checklist |
| Retrieval practice | Dunlosky: 1 of only 2 techniques rated HIGH utility; Hattie practice testing d approx 0.54 | Low (a few recall questions, no marking needed) | Open every lesson with a low-stakes 5-question quiz on last lesson or last week, ungraded |
| Spaced practice (spacing) | Dunlosky: HIGH utility; Hattie spaced vs mass practice d = 0.60 | Low (mostly planning and sequencing) | Deliberately re-ask older content in weekly and monthly reviews instead of blocking it in one unit |
| Explicit / direct instruction | AERO meta-analyses effect sizes 0.40-1.22; Hattie direct instruction d = 0.60; a NSW CESE theme | Low to moderate (structuring lessons into small steps) | Teach in small steps with I-do / we-do / you-do and check every student's response before moving on |
| Worked examples | Hattie worked examples d = 0.37; a core Rosenshine principle | Moderate (building clean worked solutions up front) | Show a fully worked solution, then fade steps one at a time until students solve unaided |
| Mastery learning (contrast: lower evidence security) | EEF +5 months but LOW security (80 studies, few RCTs) | Very low to moderate | Require a set proficiency on a unit before moving on, with re-teaching loops for those not yet there |
This is the job tutero.ai is built for. The AI teaching platform bakes the evidence into the resources it creates: retrieval quizzes on prior content, spaced review sets that pull older topics back in, and worked examples with faded steps ready to go. You already know retrieval, spacing and worked examples work. The platform builds them for you, so the science shows up in your classroom without the extra hours.
Want the highest-impact strategies to run themselves? Create your first retrieval quiz or faded worked example free at tutero.ai. Start creating, it's free.
You already know retrieval practice, spacing and worked examples work. The problem is the prep tax.
You already know retrieval practice, spacing and worked examples work. The problem is the prep tax.
Every staff meeting ends with another strategy you must try, and your feed offers ten more before you reach the car park. Some of it moves learning and some of it is noise, with little time to tell which is which.
This guide covers the strategies with the strongest evidence, what each really costs in prep, and one concrete move for your next lesson. It is written for teachers and school leaders, and it leans on the EEF Teaching and Learning Toolkit, John Hattie's synthesis and AERO rather than guesswork.
The quick answer
Evidence-based teaching strategies are classroom methods backed by high-quality research on how learning works. The strongest are feedback, metacognition and self-regulation, retrieval practice, spaced practice and explicit instruction. Across the EEF Teaching and Learning Toolkit, John Hattie's synthesis and cognitive science, these five reliably lift student progress, and most cost very little once you plan them.
Two things stand out. The highest-evidence strategies are not the most expensive to run, and the cheapest wins (retrieval and spacing) are the ones teachers skip most, because the prep recurs every single lesson. The rest of this article takes the top strategies one at a time.
What is retrieval practice and why does testing yourself improve learning?
Retrieval practice means asking students to bring information back from memory instead of re-reading or re-highlighting it. The act of recalling strengthens the memory and shows both of you what has actually stuck. It is one of only two techniques Dunlosky and colleagues rated high utility in their review of ten study methods, and Hattie's synthesis puts practice testing at around d = 0.54.
Its biggest advantage for teachers is cost. A five-question recall quiz needs no marking and adds almost nothing to your workload. Keep it low stakes so students answer honestly rather than guessing to protect a grade. Retrieval works best when it is frequent, spaced out and tied to prior content, not just the lesson you taught five minutes ago.
Practical ways to run it with zero marking:
- Open with a "do now" of five short-answer questions: three from last lesson, two from earlier in the term.
- Use mini-whiteboards or hands-up cold calling so every student retrieves, not just the volunteer.
- Have students close their books and brain-dump everything they remember on a topic, then compare to their notes.
- Turn old exam or quiz questions into starters instead of writing new content.
AERO's evidence synthesis on spacing and retrieval practice rates both against its Standards of Evidence and frames them for Australian classrooms, which makes it the cleanest AU reference to hand a faculty.
What is spaced practice and how do I schedule it across a term?
Spaced practice, or spacing, means spreading contact with a topic across days and weeks instead of massing it into one block. The same total study time produces stronger long-term retention when it is distributed, because each return forces a slightly harder retrieval. It is the second technique Dunlosky rated high utility, and Hattie puts spaced versus massed practice at d = 0.60.
Spacing costs almost nothing to run. It is a sequencing decision, not a resourcing one. The trap is that blocking a unit and moving on feels more efficient in the moment and tests better in the short term, which is exactly why teachers default to it. Planned properly, spacing turns your existing content into a review schedule rather than extra work.
A simple term structure that builds spacing in without new materials:
- Weekly review: every Friday starter revisits the current week plus one topic from two to three weeks ago.
- Monthly review: once a month, a low-stakes quiz sweeps across everything taught that term.
- Interleave related topics: alternate question types (for example fractions and decimals) so students choose the method, not just apply the one you named.
Spacing and retrieval are partners. You space the practice, and each spaced session is a retrieval session. Run them together and you get compounding returns from content you have already made.
How much does feedback actually improve student progress?
Feedback is one of the highest-leverage things you can do, and one of the most misused. In the EEF Teaching and Learning Toolkit it carries roughly +6 months of additional progress at high evidence security, and Hattie's synthesis puts it at d = 0.70, well above his 0.40 hinge point.
But the effect depends entirely on what the feedback does. Task-focused comments that tell a student exactly what to fix next lift progress. Grades and praise, on their own, do not, and can even reduce effort.
The short version: feedback only works when it names the next fix.
The workload risk is real: written feedback for every student every task is unsustainable. The evidence does not require it. Fewer, sharper, action-naming comments beat a marked pile that students glance at and file.
To get the impact without the marking marathon:
- Give whole-class feedback from a sample of books rather than marking every one.
- Name the two or three fixes that would move most students.
- Build in lesson time for students to act on comments before you mark again.
NSW CESE's What works best names effective feedback as one of its eight evidence-based practices, alongside explicit teaching, and its guidance is written for Australian schools.
What is explicit instruction and how is it different from discovery learning?
Explicit instruction means teaching new content in small, sequenced steps, showing students how to do something before asking them to do it alone, and checking understanding at each step. It is the opposite of discovery learning, where students are expected to work concepts out for themselves with minimal guidance.
For novices learning new material, the evidence favours explicit teaching clearly. AERO's explicit instruction practice guide reports meta-analytic effect sizes from 0.40 to 1.22, and Hattie puts direct instruction at d = 0.60.
The reason is cognitive load. Beginners have limited working memory, so unguided problem-solving overloads them before they have secure knowledge to reason with. Explicit instruction manages that load by breaking content into steps and building fluency before independence.
The practical shape is I-do, we-do, you-do. You model the skill, work through examples together, then release students to independent practice only once responses show they are ready. Check every student's response along the way, through mini-whiteboards, cold calling or quick quizzes, rather than reading the room from a few raised hands.
This is not lecturing and it is not rote learning. It is heavily interactive, with constant checks that tell you when to move on and when to re-teach.
What is metacognition and self-regulated learning, and how do I teach it?
Metacognition is thinking about your own thinking: planning how to approach a task, monitoring progress while you work, and evaluating the result. Self-regulated learning is the wider skill of managing your own effort and strategy.
In the EEF Toolkit, metacognition and self-regulation approaches carry about +8 months of additional progress at high evidence security, drawn from 355 studies. That makes it one of the best-evidenced strategies available and one of the cheapest, since it needs staff training rather than new resources.
A recent Toolkit refresh nudged the figure to +7 months as new studies were added, a reminder these estimates are periodically re-checked. Hattie puts metacognitive strategies at d = 0.60. The catch is that it has to be taught explicitly, not assumed.
You teach it by making your own thinking visible and then handing the routine to students:
- Model a plan-monitor-evaluate routine out loud as you tackle a hard problem, including the false starts.
- Give students a success checklist and require a self-check against it before they hand work in.
- Ask "what will you do if you get stuck?" before a task, so the strategy is chosen in advance.
- Prompt reflection after: what worked, what would you change next time?
What are worked examples and when should I use them instead of problem-solving?
A worked example is a fully solved problem, shown step by step, that students study before attempting similar problems themselves. Worked examples are most powerful when students are new to a topic, because studying a clear solution costs far less working memory than solving from scratch. Hattie puts worked examples at d = 0.37, near his hinge point, and they are a core part of Rosenshine's Principles of Instruction.
The key is not to leave students on worked examples forever. You use them to introduce a method, then gradually fade the support, removing one step at a time so students complete more of the solution themselves until they are solving unaided. Switch to independent problem-solving once they have the fluency to handle it without overload.
A simple fading sequence:
- Full worked example.
- A completion problem with the last step missing.
- Then one with the last two steps missing.
- Finally a problem to solve alone.
Alternate a worked example with a similar problem to solve, so students study then apply in quick cycles. The upfront cost is building clean, uncluttered solutions, which is the one moderate-prep strategy in the ranking below and a natural place to lean on automation.
How can teachers use these high-impact strategies without adding hours of prep?
The honest answer is that the science is not the hard part. The prep tax is.
Every high-impact move carries recurring prep:
- Retrieval practice needs fresh recall questions every lesson.
- Spacing needs old content re-sequenced into weekly and monthly reviews.
- Worked examples need clean, faded solutions built by hand.
Every one of those is recurring work, and OECD's TALIS 2024 survey again found time pressure among teachers' biggest constraints. That is precisely why teachers know these strategies work but cannot run them consistently.
The short version: the strategies work; the prep is what stops them.
The fix is not more willpower. It is removing the prep so the highest-impact moves happen by default, not on the days you have a spare hour.
The table below ranks the highest-impact strategies three ways at once: their evidence rating, their typical prep and marking cost, and one concrete move you can run in your next lesson. Impact and effort rarely line up the way teachers expect, and that gap is the whole point.
| Strategy | Evidence rating (EEF months of progress / Hattie effect size d) | Typical prep / workload cost | One concrete classroom move |
|---|---|---|---|
| Feedback | EEF +6 months (high security); Hattie d = 0.70 | Very low to set up, but time-heavy if written per student | Give one task-focused comment that names the next fix ("add a unit to your answer"), not a grade or praise |
| Metacognition & self-regulation | EEF +8 months (high security, 355 studies); Hattie metacognitive strategies d = 0.60 | Very low (staff training only) | Model and script a plan-monitor-evaluate routine, then have students self-check work against a success checklist |
| Retrieval practice | Dunlosky: 1 of only 2 techniques rated HIGH utility; Hattie practice testing d approx 0.54 | Low (a few recall questions, no marking needed) | Open every lesson with a low-stakes 5-question quiz on last lesson or last week, ungraded |
| Spaced practice (spacing) | Dunlosky: HIGH utility; Hattie spaced vs mass practice d = 0.60 | Low (mostly planning and sequencing) | Deliberately re-ask older content in weekly and monthly reviews instead of blocking it in one unit |
| Explicit / direct instruction | AERO meta-analyses effect sizes 0.40-1.22; Hattie direct instruction d = 0.60; a NSW CESE theme | Low to moderate (structuring lessons into small steps) | Teach in small steps with I-do / we-do / you-do and check every student's response before moving on |
| Worked examples | Hattie worked examples d = 0.37; a core Rosenshine principle | Moderate (building clean worked solutions up front) | Show a fully worked solution, then fade steps one at a time until students solve unaided |
| Mastery learning (contrast: lower evidence security) | EEF +5 months but LOW security (80 studies, few RCTs) | Very low to moderate | Require a set proficiency on a unit before moving on, with re-teaching loops for those not yet there |
This is the job tutero.ai is built for. The AI teaching platform bakes the evidence into the resources it creates: retrieval quizzes on prior content, spaced review sets that pull older topics back in, and worked examples with faded steps ready to go. You already know retrieval, spacing and worked examples work. The platform builds them for you, so the science shows up in your classroom without the extra hours.
Want the highest-impact strategies to run themselves? Create your first retrieval quiz or faded worked example free at tutero.ai. Start creating, it's free.
FAQ
.svg)
.svg)
Online maths tutoring at Tutero is catering to students of all year levels. We offer programs tailored to the unique learning curves of each age group.
We also have expert NAPLAN and ATAR subject tutors, ensuring students are well-equipped for these pivotal assessments.
We recommend at least two to three session per week for consistent progress. However, this can vary based on your child's needs and goals.
Our platform uses advanced security protocols to ensure the safety and privacy of all our online sessions.
Parents are welcome to observe sessions. We believe in a collaborative approach to education.
We provide regular progress reports and assessments to track your child’s academic development.
Yes, we prioritise the student-tutor relationship and can arrange a change if the need arises.
Yes, we offer a range of resources and materials, including interactive exercises and practice worksheets.
You already know retrieval practice, spacing and worked examples work. The problem is the prep tax.
You already know retrieval practice, spacing and worked examples work. The problem is the prep tax.
You already know retrieval practice, spacing and worked examples work. The problem is the prep tax.
A five-question recall quiz needs no marking and adds almost nothing to your workload.
Every staff meeting ends with another strategy you must try, and your feed offers ten more before you reach the car park. Some of it moves learning and some of it is noise, with little time to tell which is which.
This guide covers the strategies with the strongest evidence, what each really costs in prep, and one concrete move for your next lesson. It is written for teachers and school leaders, and it leans on the EEF Teaching and Learning Toolkit, John Hattie's synthesis and AERO rather than guesswork.
The quick answer
Evidence-based teaching strategies are classroom methods backed by high-quality research on how learning works. The strongest are feedback, metacognition and self-regulation, retrieval practice, spaced practice and explicit instruction. Across the EEF Teaching and Learning Toolkit, John Hattie's synthesis and cognitive science, these five reliably lift student progress, and most cost very little once you plan them.
Two things stand out. The highest-evidence strategies are not the most expensive to run, and the cheapest wins (retrieval and spacing) are the ones teachers skip most, because the prep recurs every single lesson. The rest of this article takes the top strategies one at a time.
What is retrieval practice and why does testing yourself improve learning?
Retrieval practice means asking students to bring information back from memory instead of re-reading or re-highlighting it. The act of recalling strengthens the memory and shows both of you what has actually stuck. It is one of only two techniques Dunlosky and colleagues rated high utility in their review of ten study methods, and Hattie's synthesis puts practice testing at around d = 0.54.
Its biggest advantage for teachers is cost. A five-question recall quiz needs no marking and adds almost nothing to your workload. Keep it low stakes so students answer honestly rather than guessing to protect a grade. Retrieval works best when it is frequent, spaced out and tied to prior content, not just the lesson you taught five minutes ago.
Practical ways to run it with zero marking:
- Open with a "do now" of five short-answer questions: three from last lesson, two from earlier in the term.
- Use mini-whiteboards or hands-up cold calling so every student retrieves, not just the volunteer.
- Have students close their books and brain-dump everything they remember on a topic, then compare to their notes.
- Turn old exam or quiz questions into starters instead of writing new content.
AERO's evidence synthesis on spacing and retrieval practice rates both against its Standards of Evidence and frames them for Australian classrooms, which makes it the cleanest AU reference to hand a faculty.
What is spaced practice and how do I schedule it across a term?
Spaced practice, or spacing, means spreading contact with a topic across days and weeks instead of massing it into one block. The same total study time produces stronger long-term retention when it is distributed, because each return forces a slightly harder retrieval. It is the second technique Dunlosky rated high utility, and Hattie puts spaced versus massed practice at d = 0.60.
Spacing costs almost nothing to run. It is a sequencing decision, not a resourcing one. The trap is that blocking a unit and moving on feels more efficient in the moment and tests better in the short term, which is exactly why teachers default to it. Planned properly, spacing turns your existing content into a review schedule rather than extra work.
A simple term structure that builds spacing in without new materials:
- Weekly review: every Friday starter revisits the current week plus one topic from two to three weeks ago.
- Monthly review: once a month, a low-stakes quiz sweeps across everything taught that term.
- Interleave related topics: alternate question types (for example fractions and decimals) so students choose the method, not just apply the one you named.
Spacing and retrieval are partners. You space the practice, and each spaced session is a retrieval session. Run them together and you get compounding returns from content you have already made.
How much does feedback actually improve student progress?
Feedback is one of the highest-leverage things you can do, and one of the most misused. In the EEF Teaching and Learning Toolkit it carries roughly +6 months of additional progress at high evidence security, and Hattie's synthesis puts it at d = 0.70, well above his 0.40 hinge point.
But the effect depends entirely on what the feedback does. Task-focused comments that tell a student exactly what to fix next lift progress. Grades and praise, on their own, do not, and can even reduce effort.
The short version: feedback only works when it names the next fix.
The workload risk is real: written feedback for every student every task is unsustainable. The evidence does not require it. Fewer, sharper, action-naming comments beat a marked pile that students glance at and file.
To get the impact without the marking marathon:
- Give whole-class feedback from a sample of books rather than marking every one.
- Name the two or three fixes that would move most students.
- Build in lesson time for students to act on comments before you mark again.
NSW CESE's What works best names effective feedback as one of its eight evidence-based practices, alongside explicit teaching, and its guidance is written for Australian schools.
What is explicit instruction and how is it different from discovery learning?
Explicit instruction means teaching new content in small, sequenced steps, showing students how to do something before asking them to do it alone, and checking understanding at each step. It is the opposite of discovery learning, where students are expected to work concepts out for themselves with minimal guidance.
For novices learning new material, the evidence favours explicit teaching clearly. AERO's explicit instruction practice guide reports meta-analytic effect sizes from 0.40 to 1.22, and Hattie puts direct instruction at d = 0.60.
The reason is cognitive load. Beginners have limited working memory, so unguided problem-solving overloads them before they have secure knowledge to reason with. Explicit instruction manages that load by breaking content into steps and building fluency before independence.
The practical shape is I-do, we-do, you-do. You model the skill, work through examples together, then release students to independent practice only once responses show they are ready. Check every student's response along the way, through mini-whiteboards, cold calling or quick quizzes, rather than reading the room from a few raised hands.
This is not lecturing and it is not rote learning. It is heavily interactive, with constant checks that tell you when to move on and when to re-teach.
What is metacognition and self-regulated learning, and how do I teach it?
Metacognition is thinking about your own thinking: planning how to approach a task, monitoring progress while you work, and evaluating the result. Self-regulated learning is the wider skill of managing your own effort and strategy.
In the EEF Toolkit, metacognition and self-regulation approaches carry about +8 months of additional progress at high evidence security, drawn from 355 studies. That makes it one of the best-evidenced strategies available and one of the cheapest, since it needs staff training rather than new resources.
A recent Toolkit refresh nudged the figure to +7 months as new studies were added, a reminder these estimates are periodically re-checked. Hattie puts metacognitive strategies at d = 0.60. The catch is that it has to be taught explicitly, not assumed.
You teach it by making your own thinking visible and then handing the routine to students:
- Model a plan-monitor-evaluate routine out loud as you tackle a hard problem, including the false starts.
- Give students a success checklist and require a self-check against it before they hand work in.
- Ask "what will you do if you get stuck?" before a task, so the strategy is chosen in advance.
- Prompt reflection after: what worked, what would you change next time?
What are worked examples and when should I use them instead of problem-solving?
A worked example is a fully solved problem, shown step by step, that students study before attempting similar problems themselves. Worked examples are most powerful when students are new to a topic, because studying a clear solution costs far less working memory than solving from scratch. Hattie puts worked examples at d = 0.37, near his hinge point, and they are a core part of Rosenshine's Principles of Instruction.
The key is not to leave students on worked examples forever. You use them to introduce a method, then gradually fade the support, removing one step at a time so students complete more of the solution themselves until they are solving unaided. Switch to independent problem-solving once they have the fluency to handle it without overload.
A simple fading sequence:
- Full worked example.
- A completion problem with the last step missing.
- Then one with the last two steps missing.
- Finally a problem to solve alone.
Alternate a worked example with a similar problem to solve, so students study then apply in quick cycles. The upfront cost is building clean, uncluttered solutions, which is the one moderate-prep strategy in the ranking below and a natural place to lean on automation.
How can teachers use these high-impact strategies without adding hours of prep?
The honest answer is that the science is not the hard part. The prep tax is.
Every high-impact move carries recurring prep:
- Retrieval practice needs fresh recall questions every lesson.
- Spacing needs old content re-sequenced into weekly and monthly reviews.
- Worked examples need clean, faded solutions built by hand.
Every one of those is recurring work, and OECD's TALIS 2024 survey again found time pressure among teachers' biggest constraints. That is precisely why teachers know these strategies work but cannot run them consistently.
The short version: the strategies work; the prep is what stops them.
The fix is not more willpower. It is removing the prep so the highest-impact moves happen by default, not on the days you have a spare hour.
The table below ranks the highest-impact strategies three ways at once: their evidence rating, their typical prep and marking cost, and one concrete move you can run in your next lesson. Impact and effort rarely line up the way teachers expect, and that gap is the whole point.
| Strategy | Evidence rating (EEF months of progress / Hattie effect size d) | Typical prep / workload cost | One concrete classroom move |
|---|---|---|---|
| Feedback | EEF +6 months (high security); Hattie d = 0.70 | Very low to set up, but time-heavy if written per student | Give one task-focused comment that names the next fix ("add a unit to your answer"), not a grade or praise |
| Metacognition & self-regulation | EEF +8 months (high security, 355 studies); Hattie metacognitive strategies d = 0.60 | Very low (staff training only) | Model and script a plan-monitor-evaluate routine, then have students self-check work against a success checklist |
| Retrieval practice | Dunlosky: 1 of only 2 techniques rated HIGH utility; Hattie practice testing d approx 0.54 | Low (a few recall questions, no marking needed) | Open every lesson with a low-stakes 5-question quiz on last lesson or last week, ungraded |
| Spaced practice (spacing) | Dunlosky: HIGH utility; Hattie spaced vs mass practice d = 0.60 | Low (mostly planning and sequencing) | Deliberately re-ask older content in weekly and monthly reviews instead of blocking it in one unit |
| Explicit / direct instruction | AERO meta-analyses effect sizes 0.40-1.22; Hattie direct instruction d = 0.60; a NSW CESE theme | Low to moderate (structuring lessons into small steps) | Teach in small steps with I-do / we-do / you-do and check every student's response before moving on |
| Worked examples | Hattie worked examples d = 0.37; a core Rosenshine principle | Moderate (building clean worked solutions up front) | Show a fully worked solution, then fade steps one at a time until students solve unaided |
| Mastery learning (contrast: lower evidence security) | EEF +5 months but LOW security (80 studies, few RCTs) | Very low to moderate | Require a set proficiency on a unit before moving on, with re-teaching loops for those not yet there |
This is the job tutero.ai is built for. The AI teaching platform bakes the evidence into the resources it creates: retrieval quizzes on prior content, spaced review sets that pull older topics back in, and worked examples with faded steps ready to go. You already know retrieval, spacing and worked examples work. The platform builds them for you, so the science shows up in your classroom without the extra hours.
Want the highest-impact strategies to run themselves? Create your first retrieval quiz or faded worked example free at tutero.ai. Start creating, it's free.
You already know retrieval practice, spacing and worked examples work. The problem is the prep tax.
A five-question recall quiz needs no marking and adds almost nothing to your workload.
They overlap heavily. Evidence-based strategies are any classroom methods supported by high-quality research. High-impact teaching strategies is the specific framing used by Australian education bodies for the practices with the strongest measured effect on achievement, such as explicit teaching, feedback and structured lessons. In short, high-impact strategies are the evidence-based methods that research shows move student progress the most.
Among the strategies in this article, feedback ranks highest at d = 0.70, well above Hattie's 0.40 hinge point. Across his full list, influences like classroom discussion (d = 0.82) rank even higher. Remember effect size measures potential impact, not certainty. A strategy only delivers that impact when it is implemented well, which for feedback means task-focused, action-naming comments rather than grades or praise.
Barak Rosenshine's 2012 paper set out ten research-based principles for effective teaching. They include daily review, presenting new material in small steps, asking many questions and checking responses, providing worked examples and models, guiding student practice, checking for understanding, aiming for a high success rate, scaffolding difficult tasks, and building in weekly and monthly review. They bring together cognitive science, classroom research and expert-teacher practice into one practical framework.
Yes. Retrieval practice benefits learners across ages, including primary students, as long as it is pitched at the right level and kept low stakes. For younger students, keep questions short and concrete, use spoken or mini-whiteboard responses rather than long written tests, and frame it as a game or a warm-up. The goal is honest recall and no anxiety, so quizzes should never carry a grade that students feel they must protect.
Dunlosky and colleagues rated common favourites like highlighting, underlining, re-reading and summarising as low utility, because the evidence that they improve durable learning is weak. Students rely on them because they feel productive and are easy, not because they work. Only practice testing (retrieval) and distributed practice (spacing) earned a high-utility rating. Redirecting students from re-reading toward self-testing is one of the cheapest upgrades you can make to how they study.
Explicit instruction is not rote learning. Rote learning is memorising without understanding. Explicit instruction teaches for understanding by breaking content into small steps, modelling each one, checking that every student follows, and building toward independent, flexible application. It is highly interactive, full of questioning and feedback, and it releases students to work alone once they are ready. The structure supports understanding, it does not replace it, which is why the evidence for it is so strong for novice learners.
Hoping to improve confidence & grades?
.png)
