Should low achievers be held back?

Should low-achieving kids be promoted or held back?  The research isn’t clear, writes Daniel Willingham.

Until recently, comparisons of kids who were promoted and kids who were retained indicated that retention didn’t seem to help academic achievement, and in fact likely hurt. So the best practice seemed to be to promote kids to the next grade, but to try to provide extra academic support for them to handle the work.

But new studies indicate that academic outcomes for kids who are retained may be better than was previously thought, although still not what we would hope.

I wonder if anyone’s studied the effects of social promotion on teachers.

About Joanne


  1. Back in the day when schools held back students not meeting grade-level standards, retainment had a positive effect on the student population as a whole, particularly those who might not pass if they didn’t step up their efforts. When one of the pair of twins in my class was held back, prior to 3rd grade, a number of others got the message.

  2. Up until high school, it seems that holding kids back doesn’t help. However, pulling kids out who have failed and providing recovery programs does help. After that, at the high school level, they only go on if they pass.

    • Mark Roulo says:

      Up until high school, it seems that holding kids back doesn’t help. However, pulling kids out who have failed and providing recovery programs does help.

      What distinction are you drawing between (a) repeating grade X, and (b) being in grade ‘X+1’ but are ‘pulled out’ and are working on grade ‘X’ material?

      And why change at grade 9?

      • tim-10-ber says:

        But the credit recovery programs…they aren’t truly legit, are they?

        I am in favor of doing away with required seat time or “grade levels”…let them move forward at mastery, but, yes, they have to be pushed to excel…education is not a given, advancement is not a given, failure is okay but it is only failure if not learned from, kids need to earn the advancement…

    • CaliforniaTeacher says:

      I’ve seen quite a few children helped by being held back in kindergarten or first grade.

  3. Although I’ve read numerous studies against retention, my own experience is that it is a complete failure. I’ve seen one student after another at my own middle school fail the next year after being retained. They are also even more disruptive, in terms of behavior.

    Teach kids that they are failures and they’ll believe it.

    A much better solution would be to change the system that is used, so the students don’t fail in the first place.

  4. wahoofive says:

    The problem with retention is that it’s all-or-nothing. If you learned 75% of the third-grade material, you’ll be behind in fourth grade, and probably fall further behind. But if we make you repeat third grade, 75% of it will be a waste of time. If we could find a way to have kids repeat just a semester or a quarter, it might help.

    • Anecdotal, but merely offered as evidence that it didn’t use to be “all or nothing”:

      When I was in elementary school, it was not uncommon for pupils to be held back – there were several in my classes that had been held back. In particular, one boy in my first grade class started off the school year with us as a repeater, but sometime around the middle of the school year he was bumped up to second grade to go back with his cohort. Evidently whatever deficiencies he had were remedied, and for the remainder of his school career he stayed on track.

  5. Stacy in NJ says:

    The artificial construct of “grades” is the problem. Instead of assessing and determining ability level of each child and then teaching them where they’re at, we start by placing them in a grade that may or may not have much to do with their abilities. We need more individualization in education. Technology and alternative delivery systems are tools that may lead to great individualization.

    MAP (measure of academic progress) testing is one step in this direction.

    • Roger Sweeny says:

      It’s almost as if the primary purpose of school was not education, but daycare.

    • Not going to happen. The demographic realities of such a policy would be unacceptable to those in control of education.

  6. palisadesk says:

    “Back in the day….”

    Oh yes, I know. I taught in a district, back in the day, that was really serious about “no social promotion.”

    Let me tell you about the “positive effect on the student population as a whole.” My fourth grade had 41 students, of whom 25 or more had failed at least once. Many students had failed twice or three times, and about six had been retained four, five or SIX times!

    Picture a fourth grade with a pregnant teen (age 15), several hulking teen boys of 14 and 15, several adolescent girls 12-15, and a bunch of kids who were more than a year or two overage. Luckily for me (from a management standpoint) many of these also had poor attendance. The poor little 8 and 9 year olds huddled on one side of the room, trying to do their work.

    In this paradise of rigorous high standards, the “positive effect” seemed to be that parents didn’t need to save for post-secondary education. By seventh grade, two thirds of the kids who had been in the Kindergarten cohort were gone. By grades 10-12, classes had only a half dozen students at most.

    Be careful what you wish for.

    As for retention in grade as a means of dealing with student learning differences — it does not address any of the fundamental problems underlying kids failing to “meet the standard:”
    — students do not learn at the same rate, and some will be unable to learn the expected amount in a given school year, regardless of the quality of teaching (and even regardless of parental ability to supplement or tutor).
    — students can vary tremendously in their mastery of different subjects. do we hold back the very bright kid who is clearly learning and participating in the curriculum but whose grasp of written language is limited? What about the ones who can barely count to ten in third grade, but can handle language arts and social studies at an acceptable level?
    — if the student was unsuccessful in Grade X the first time, why the bleep should we think s/he will be successful the second — or third or fourth — time? Just what is going to be “different.”?

    It’s true that most of us,. myself included, can point to a case where retention in grade — usually very early, in K or first — proved beneficial. If students were routinely provided with the academic support needed, retention might be a success — but then, if they had the support they needed they would not be “failing” in the first place.

    Grouping students for instruction based on the date they emerged from the womb makes very little sense, and enforcing it more rigorously makes even less.

    We used to shlep the slow learners and low achievers off to “special ed” — often these were “dumping ground” classes that perpetuated failure (see Lloyd Dunn’s seminal article, ca. 1975, on the lack of effectiveness pf special education placements. Meta-analyses since then have found the same). That bus isn’t coming around again any time soon. Full inclusion has many problems, too.

    So, the next bright idea to handle disparities in student achievement is…….


    • First, I did not use the phrase “back in the day” to imply some sort of Golden Age, but to mean a long time ago. I should have stated that the school had no kids who were unable to do grade-level work, with sufficient effort. I am opposed to simply letting kids drift along with little effort or mastery. That was also before ALL kids, no matter how severe their cognitive, emotional and/or physical disabilities, entered regular public schools. Certainly, the situation you describe is unacceptable. One of my biggest complaints about public schools is that they tend to a one-size-fits-all approach, as if all kids had the same instructional needs.

  7. My ADD son couldn’t handle the workload or pay attention when younger. But suddenly at age 16, everything clicked. The 8th grade teacher had recommended holding him back, but I’m glad I didn’t. He’s now a solid B student and handing in all of his homework without anyone looking after him.

  8. Palisa, I mostly agree with you about retention, but it’s worth remembering that all the support in the world won’t help a low level kid learn difficult material. So kids often fail because they aren’t cognitively able to do the work.

    That’s why we should ability group kids of like age. Kids who are in sixth grade but reading at 4th grade level should be able to work on their reading skills without being surrounded by 4th graders.

    • Michael E. Lopez says:

      You’re obviously racist, elitist, and several other bad ‘ists’.

    • In education, we call this differentiation, not retention. I teach 8th grade. I have students who read at an 11th grade level and students who read at a 5th grade level. Separating them by these reading levels is tantamount to separating students by their shoe sizes, a ridiculous idea.

      I do what I can to help all students improve by coaching them on choosing books that are right for them.

      If a student in 8th grade reads at an 11th grade level, should she be moved to the 11th grade?

      • Michael E. Lopez says:

        Separating them by these reading levels is tantamount to separating students by their shoe sizes, a ridiculous idea.

        Because what could students’ reading levels possibly have to do with their academic performance?

  9. Richard Aubrey says:

    Remember the Safety Patrol? I was one, once, in the sixth grade. Sent off to deal with some out-of-control kids who had been held back, while the teacher in charge of the playground resolutely looked the other way.
    Had a good view of the subject then. On the other hand, it motivated me to get into martial arts.

  10. There are 2 factors to look at:

    1) Do the “held-back” kids actually get help for their deficiencies? In my experience, seldom. Yet, the best predictor of success in school is the reading, writing, and math skills. So, why aren’t these kids getting that help?

    Pushing them onto the next grade level isn’t helping. Holding them back apparently doesn’t help much, either.

    Here’s a radical thought – why don’t we target those kids for intensive help, both before and after school, as well as during the summer vacation? If they can’t read, well, why don’t we give them intensive phonics instruction? If they can’t do math, why don’t we use a program like Kumon or Saxon Math for those students. Let them re-join their peers after bringing them up to speed – and ONLY after they have tested on grade level (outside evaluators – we have too many in the profession who will lie their a$$ off, and help students cheat).

    2) The classes that lose these “skill-deficient” kids do better – not just academically, but in terms of less lost class time. Too many schools have class time wasted, due to disruptive students.

    I had an opposite problem earlier this year – one of my kids was WAY more advanced in skill level than the rest of the class. He was a chronic disrupter. When moved to a more advanced class, he struggled (as he was used to getting by with little to no effort). But the class he left? WAY easier to manage, as well as the overall academic performance improved.

    How about looking at “Skill Level Grouping”? NOT tracking, just grouping students with others that have similar skill levels. It would reduce teacher time modifying lessons for various remediation/enrichment needs. It would have to be flexible enough to allow students to move from level to level, as they progressed.

    Failing that, stop dumping low achievers (who have the ability to succeed, but choose not to) into the classes of those who struggle with the material.

    • Roger Sweeny says:

      You, ma’am, are a subversive.

      • Better yet, use phonics and Saxon (or Singapore) math for everyone, from the start; with good curriculum and explicit instruction, perhaps fewer kids will struggle and fall behind.

        • Richard Aubrey says:

          Boy, talk about subversive.

        • Phonics is the absolute wrong way to teach people to read.

          • J. D. Salinger says:

            Well, OK; let’s see how well you do learning Russian, Greek or Hebrew using the whole language method, Mark.

          • Mark, I don’t recall that reading is among your subject areas? It is one of mine, and you are quite wrong. Using words like “absolute” is generally a sign of wrongness, but that aside, studies have shown that phonics is one of the quickest and most efficient ways to bring struggling high school and adult readers up to grade level. To me it is a tool, and to you it is an absolute wrong way. As we say in the military, stay in your own lane.

          • Uggh. I apologize for that parting shot, it was unbecoming.

          • Roger Sweeny says:

            Phonics is the absolute wrong way to teach people to read.

            What makes you say that?

          • “Phonics is the absolute wrong way to teach people to read.”

            Oh, but of course – it’s MUCH easier to memorize thousands of possible letter combinations, a la Chinese characters, than it is to remember the 44 phoneme sounds of English.


            Talk about setting up at-risk students for failure ~

          • I’m an ELA teacher, and I have spent years researching reading and literacy and putting this research into practice. So, although I may not be an expert, I know more than most, especially people who aren’t teachers.

            Actually, I’d do much better with a whole language approach to learning a foreign language than I would using phonics. What makes me say this, Roger, is the work of reading experts — Stephen Krashen, Nancie Atwell, Alfie Kohn and Donalyn Miller, to name a few. Kohn and Krashen, in particular, have spoken out against the effectiveness of phonics and other guided reading programs.

            Krashen is one of the world’s most-respected authorities on grammar and literacy. He says that reading itself teaches all of the necessary skill sets and that isolating skills not only doesn’t help improve reading, it can in some cases be harmful to the reader.

          • Michael E. Lopez says:

            Krashen is one of the world’s most-respected authorities on grammar and literacy.

            And Einstein was one of the world’s most-respected authorities on physics.

            But he was really wrong about some stuff, too.

            It happens.

          • Michael, are Krashen, Kohn, Atwell, Miller and Lucy Calkins (another widely-published expert on literacy) all wrong?

            These are people who have spent the better part of their adult lives researching and teaching reading and language acquisition. They’ve discovered that a whole language approach works best.

            Incidentally, I have 20 years of experience myself, and it began in the world of guided reading programs. Before I knew better, I used the programs faithfully. Then, I did some research of my own, changed the way I teach reading and, almost magically, I started seeing remarkable gains in my students’ reading.

          • Michael E. Lopez, Esq. says:

            Michael, are Krashen, Kohn, Atwell, Miller and Lucy Calkins (another widely-published expert on literacy) all wrong?

            Surely it’s not a necessary truth that they are all right, is it?

            Are we to think that no one on Earth really understood how to teach language effectively until Krashen et al. “discovered” it? Or, perhaps a weaker claim — that no one really knew which method was best until Krashen et al. “discovered” it?

            Are we to think that no one in academia disagrees with Krashen, that his results are replicable and irreproachable, his methodologies flawless?

            That’s obviously not the case. I’m not saying Krashen’s wrong about everything, or even about any specific thing. I’m just saying that it’s really a sort of empty assertion to talk about his expertise and credentials as any sort of indication that his views are correct. And I also think that there are some pretty intuitive arguments that he might be wrong — such as the accusation that whole language approaches mistake the subjective, adult experience of reading for the actual neurological process of acquiring the skill in the first place.

            “Reading by reading,” sounds nice as a slogan. But if you think about it as a practical prescription, it’s pretty obviously got a regress problem right off the bat.

            Garon Wheeler has a fairly balanced view of Krashen’s second-language work in 20 TESL Canada Journal 92 (2003). It’s pretty clear that, as bright as Krashen might be and as interesting and provoking as his theories are, that he’s not the last word in his field by any stretch of the imagination. He’s even, in some ways, in a little bit of disfavor.

            That doesn’t make him wrong, of course. Galileo was in “disfavor”, too. That the church officials were true believers in geocentrism didn’t make it true. And just because you happen to be a true believer in the Comprehension Hypothesis/Input Hypothesis, just because you are subjectively experiencing success in your teaching and attribute it to Krashen’s theories, doesn’t make it so.

            One can easily imagine that a teacher could attribute his success in teaching to Jesus. It doesn’t mean we all need to run out and convert in order to teach our students — even if he’s right.

          • Mark Roulo says:

            Alfie Kohn has not spent the better part of his adult life researching and teaching reading.

          • Roger Sweeny says:


            Thanks for your answer. I now read by recognizing word shape and letter combination. If I have to read something that is spelled out phonetically, I am slowed down. However, I began with phonics.

            I was once thinking, what if soccer was taught the way good soccer players play it? Don’t look at the ball, don’t think about your feet, notice where the other players are on the field, consider whether you should move left or right or pass. Beginners would constantly be falling over themselves and failing. They would start thinking, “Soccer is hard. I don’t like soccer,” and eventually, “Soccer is stupid. I don’t ever want to play soccer again.”

            That is how a disturbingly large number of my 9th graders describe reading.

  11. palisadesk says:

    Cal’s suggestion of “ability groups” might be appropriate for secondary — I wouldn’t pretend to know. But ” kids often fail because they aren’t cognitively able to do the work” is rarely the issue in the early elementary grades, and even in upper elementary, skill deficits, irregular attendance or language issues are more frequently the problem.

    In the very early grades, which is where most of the emphasis on grade retention is concentrated, there is a very LOW correlation between cognitive ability and mastery of foundation skills. Beginning reading, written language and math skills are far more dependent on various developmental aspects (language, motor skills..), working memory, ability to store and retrieve paired-associate learning quickly. In my low-low-low group of second graders recently, I had a student with an IQ of 160+, a student with developmental delay due to abuse and neglect, a late talker with possible language impairment, and a kid with poor attendance. These kids were all over the map cognitively, but were at the same instructional level.

    The kind of grouping that works best in the elementary years is based on the student’s instructional needs, not his or her presumed “ability,” which is often hard to measure accurately in the early years in any case. What DOES happen, though, is that students who fail to master those early skills show a steady decline in measured IQ over the years. So we need to ensure they get the basics to enable them to use and enhance whatever cognitive ability they have — it’s not static, either for better or worse.

    I call it “instructional level grouping” but I think it’s the same concept as LindaF above was proposing. It’s quite different from tracking,” because it allows for the possibility that a kid can be very “low” in some areas while s/he is advanced or average in others. The difficulties lie in timetabling and organizing the number of instructional groups needed, and making them flexible enough for students to advance as quickly as possible. Logistically, this is a challenging option.

  12. Richard Aubrey says:

    In the military, they say that amateurs talk tactics and professionals talk logistics.
    No matter how gung-ho you may be, if the loggie guys say you can’t get there…you can’t get there.
    See the Germans and WW I….forget it.
    Anyway, the logistics of your suggested arrangement do seem to be difficult, if only getting the kids down the hall to the next room at the appropriate time. IIRC, I could stand reading twice as long as I could stand math, and if that turns out to be educationally valid for most kids, A/B scheduling would look like a well-lighted path with no intersections.

    • palisadesk says:

      The logistics are difficult, but not impossible. Some charter schools do this kind of grouping (Arthur Academies and Charter Day School, for example) , and a simpler version is commonly called the “Joplin Plan” (which inspired Slavin’s organization of Success For All but the data on SFA don’t appear impressive to me, for other reasons). It requires use of paraprofessionals and non-academic staff to take on some instructional duties — the gym teacher, Spanish teacher, even the school secretary (in some cases) would have a reading group (or math group). Usually, successful implementations of such models require involvement of local high school or college students as tutors or group leaders, community support, and so forth.

      It takes will, commitment, and some extra $$$, but isn’t inherently un-doable. The problem is that schooling isn’t truly about results, so the will is lacking. Often, too, the knowledge of how to teach low performers effectively is *also* lacking.

  13. Lightly Seasoned says:

    This is what RtI is supposed to address.

  14. I was a student in high school who wouldn’t have graduated if I hadn’t passed a required course (which I had failed as a 10th grader) as a senior. Why should students be given social promotion just to be kept with their age cohorts, or to not hurt their self-esteem.

    In the real world of work, persons who are low or below average performers don’t get raises (unless you work for the government), or promotions for that matter, why should it be any different in public schools. Oh wait a minute, the kid might have their self-esteem damaged if we don’t promote ’em (just wait and see what happens when they try to get a job and can’t actually do the job in question due to the fact they never learned what they needed to learn in grades K-12).

    Retention is only effective to a point, and like others, I’m quite skeptical of credit recovery programs where students make up 180 days of classwork inside of a month (or less), but even with credit recovery, many states have exit exams these days, and if the students can’t pass ’em, they don’t earn a diploma, but rather a certificate of attendance (which is worthless in terms of trying to gain admittance to college or the military, and the military requires minimum ASVAB scores to even be allowed to enlist, let alone be a warrant or regular officer).

  15. But ” kids often fail because they aren’t cognitively able to do the work” is rarely the issue in the early elementary grades, and even in upper elementary, skill deficits, irregular attendance or language issues are more frequently the problem.

    Yes, of course, SOME kids who are cognitively able are still struggling. But NO kid with cognitive deficits are succeeding.

    The real issue with failure at all levels is cognitive ability, with the understanding that some high ability kids have complicating factors.

  16. Oops–I accidentally hit enter before fixing several typos and an unclosed bracket. e . Sorry. And I wasn’t even done!

    In any event, spare me and everyone else the excuses about the other reasons why kids might fail. Low cognitive ability is the huge reason. Everything else is an also ran.

    Moreover, the teaching issues involved in helping kids with specific learning disabilities or absenteeism are entirely different from the issues involved with kids of low cognitive ability.