Wednesday / July 24

Coming Back from COVID: Anything but Retention 

The negative effects from teaching during COVID-19 are a hot topic. While there have been positives resulting from the pandemic—especially the need to value and teach self-regulation and to have teachers engage in higher levels of release of responsibility (Hattie 2021)—there are many students who have failed to thrive, been denied access to technology, and been absent and not engaged. While these results occur also in regular schoolingtoo, COVID-19 spotlights the need to have a laser-like focus on providing solutions. 

One fast-emerging popular solution is to accelerate retention: Hold ‘em back a year! They missed a year so let them repeat it. They have fallen behind and not experienced a full year’s teaching, so let’s make them do it again. It sounds easy: It puts the blame squarely on the students, and by repeating the year we may make our accountability test scores look better. Many US states have already enacted laws preventing students from moving on to fourth grade until they are considered proficient ostandardized reading tests. Despite the pleas for retention, the percentages are declining and now hover around 2.4%. Retention is highest in first grade (>6%), for boys, and Blacks and Hispanics (Warren, Hoffman, & Andrews, 2010).  

Yes, 2020 was a strange year! And many students missed out—too often this reflected an imbalance in that Black and Brown children and those from families with fewer resources are overly represented in this group, due partly to lack of access to technology and starting lower on the achievement stakes even prior to the pandemic. Surely, we should have a policy that works for them in particular. 

The research on retention is quite simple: It sucks. There have been 9 meta-analyses, about 250 articles, 80,000 students, and 3,000+ effects with an average effect of -.32, all with very few or very small positive effectsOf all the influences over which we have some control, retention is probably one of the most negative, as demonstrated here: MetaXIn addition, the effects on personal and social outcomes are also negative, and these negative effects increase over time. 

Look at some quotes from non-meta studies: 

“Retention reduces high school credit accumulation and the likelihood of taking enough math and English Regents exams to graduate from high school, increases dropout, and leads to a significant increase in educational costs incurred by NYC DOE” (Mariano, Partorell, & Tsai2018). 

Holding young people back in schools holds them back in life” (National Academy of Education (1991, p. 53). 

Students are retained in rather arbitrary and inconsistent ways, and those flunked are more likely to be poor, males and minorities. …  It would be difficult to find another educational practice on which the evidence is so unequivocally negative” (House, 1989).

“School retention has negative effects on student achievement: and enrollment; attendance; drop out rates; stability rates (percent of students who stay in the same school for the entire year); suspensions; percent of male students; percent of black students; and reading comprehension” (Kaczala1991).  

“The confluence of research examining academic and adjustment outcomes associated with grade retention indicates that there are no benefits to student retention over promotion” (Jimerson & Brown 2013).

Too many retained students received the same experiences in the repeated year, so perhaps the effects of no change would be expected, but this is not the case: They actually go backwards as they live up to the expectations of failure. Being retained one year almost doubled a student’s likelihood of dropping out, while failing twice almost guaranteed it. Retention is the second greatest predictor of school drop-out.  

So, what is the alternative?  The traditional argument is automatic social promotion, but this sometimes maintains lack of progress (even though it cannot be worse than retention). The opposite of retention is acceleration. Levin (1998) has shown the benefits of acceleration for ALL students (e.g., increasing the rate of learning, particularly telescoping the curriculum). Those who have been absent or falling behind seem to profit from doing less more deeplyDarling-Hammond (1998) provided four options 

  • Enhance professional development for teachers to ensure that they have the knowledge and skills they need to teach a wider range of students to meet the standards 
  • Redesign school structures to support more intensive learning 
  • Ensure that targeted supports and services are available for students when they are needed 
  • Employ classroom assessments that better inform teaching 

There are also good arguments for supplementing regular teaching with 1:1 or small group intensive tutoring. Kraft and Falken (2021) outlined a comprehensive tutoring program for post-pandemic interventions. Such tutoring could support students’ social-emotional development, enhance their attachment to school, and accelerate the rate of learning. They estimated the cost and the implementation challenges and noted that some places like Maryland and LA Unified have already committed to tutoring. In my home state of Victoria, Australia, a condition of schools receiving tutoring funds is they must employ ex-teachers or initial teacher education students, thus ensuring this is not another feel-good amateur without expertise disaster.   

I would insist that the classroom teacher and school leader sign off on the success criteria for the tutoring, provide baseline evidence of the child’s status, and then insist on regular evaluative evidence of progress towards the success criteriaGiven our experience with online teaching, the costs of travel could be reduced and it should be easier to schedule tutoring sessions so as not to disrupt the student’s regular classroom instruction. Some tutoring could be 1:1 and some in small groups as students can feed and learn from each other, enhance their skills in collective efficacy, and build relationships. Tutors who are not efficient and effective at reaching success criteria could receive intensive ongoing professional learning, and in time the best training programs could also form the basis of educating new tutors and building a networked system for best practice. Kraft and Falken recommend developing a National Tutoring Institute to assist in finding, training, evaluating, and promoting best tutoring practices. It could curate a library of instructional materials and formative assessment for tutors, provide technical assistance and implementation support, and assist in commissioning program evaluations to inform ongoing improvement efforts. 

Every student deserves optimal learning interventions; if this was the case, there would be little need to classify students—a process often needed to chase the funding tied to the labels we inflict on children. Every student needs excellent diagnosis of what they can and cannot do (yet) and evidence that the more intensive instruction has the desired accelerated impact on the student. They don’t deserve to be blamed, to bear the costs of teaching failures, or to be put in a situation where they are even more likely to fail in schooling.  Anything but retention. 


Darling-Hammond, L. (1998) Alternatives to Grade Retention: Four complementary strategies to improve teaching and learning make more sense than holding students in grade.  The School Administrator,   AASA | American Association of School Administrators 

Hattie, J.A.C. (2021).   The unplanned experiment for teaching and learning ( 

House, E. R. (1998, November). The predictable failure of Chicago’s student retention program. In Rethinking Retention to Help All Students Succeed Conference, November, Chicago, Illinois. 

Jimerson, S.R., & Brown, J.A.  (2013).  In J. Hattie, & E.M. Anderman  (Eds.). International guide to student achievement. Routledge. 

Kaczala, C. (1991). Grade retention: a longitudinal study of school correlates of rates of retention. Cleveland Public Schools, Department of Research and Analysis, ERIC Document Reproduction.  

Kraft, M.A. & Falken, G.T. (2021). A Blueprint for Scaling Tutoring Across Public Schools EdWorking paper No. 20-335.  Annenberg, Brown University. See Research | Matthew A. Kraft ( 

Levin, H. M. (1998). Accelerated schools: A decade of evolution. In International handbook of educational change (pp. 807-830). Springer, Dordrecht. 

Mariano, L. T., Francisco (Paco) Martorell, & Tsai, T. (2018). The effects of grade retention on high school outcomes: Evidence from New York City schools. Santa Monica, CA: RAND. 

Warren, J.R., Hoffman,  E., & Andrews, M. (2010). Patterns and trends in grade retention ni the United States, 1995-2010.  Educational Researcher, 43(9), 433-443. 

Written by

Professor John Hattie is an award-winning education researcher and best-selling author with nearly 30 years of experience examining what works best in student learning and achievement. His research, better known as Visible Learning, is a culmination of nearly 30 years synthesizing more than 1,500 meta-analyses comprising more than 90,000 studies involving over 300 million students around the world. He has presented and keynoted in over 350 international conferences and has received numerous recognitions for his contributions to education. His notable publications include Visible LearningVisible Learning for Teachers, Visible Learning and the Science of How We Learn, Visible Learning for Mathematics, Grades K-12, and, most recently, 10 Mindframes for Visible Learning.

No comments

leave a comment