top of page


Grouping students by ability is often suggested as a way for teachers to provide greater diversified instruction to students, and to better teach to the individual abilities within a class. After all, as teachers, we often have students with abilities that differ by multiple grade levels. That all being said, grouping student’s by their ability does present some possible risks, especially for students in the lowest-performing groups. Research into labeling theory and the impact of expectations has shown that students who perceive themselves to be labeled as less intelligent are likely to create a self-fulfilling prophecy. “Labelled students” often try less hard and expect less of themselves. Moreover, research has shown that teachers can lower their expectations for “labeled students” spend less time helping “labeled students” and also generally contribute to a negative self-fulfilling prophecy for students identified as being less intelligent. Advocates against ability grouping, also often argue that having gifted students work with struggling students benefits the struggling students, both by increasing the expectations of the struggling students and by inadvertently providing some peer tutoring.

On the other hand, it could be argued that having gifted students work with struggling peers, for the sake of their struggling peers, could be unfair to the gifted student. After all, what is best for the class as a whole, might not be the best for the gifted student. Ultimately, this might be not a question of scientific efficacy, but of philosophical value. Do we take the utilitarian perspective and do what’s best for the group? Or do we place the needs of the individual on a higher pedestal? 

This debate is not a modern one, but rather one that has been ongoing for the past 100 years. Indeed there have been at least 11 meta-analyses of the topic over the past 30 years and more than 280 individual quantitative studies of the subject. 

Kulik and Kulik did a meta-analysis of the topic in 1987, and 1992. Their studies found Hedges g effect sizes of .22 and .29. Robert Slavin conducted a meta-study in 1987 and found a Hedges g effect size of .39. Monstellar and et al did a meta-study in 1996 and found a Hedges g effect size of .19. On average these studies found an average effect size of .27. This is a statistically significant effect size; however, it is still a low one. Examining this data would suggest that grouping students by ability might provide an overall benefit. However, I would argue that this benefit is too small to be worth the opportunity cost for teachers’ time. Moreover, I would personally recommend a more fluid approach based on learning goals for grouping, not a fixed approach based on ability. For more information on this idea, please read our article on Podding. Or listen to our podcast episode on Podding: To contextualize these ability grouping effect sizes, I have graphed these effect sizes below, but also included some high yield strategies, to provide a comparison. 

Ability Grouping IMG001.png
Ability Grouping IMG002.png

While the overall effect size of ability grouping suggests a minor benefit. We cannot just consider how ability grouping affects the whole class, but rather we need to consider the impact of ability group on each ability level. Kulik and Kulik’s 1987 study looked at precisely this. Their study found a .33 effect size for the high group, a .34 effect size for the medium group, and a .23 effect size for the low group. As I alluded to at the beginning of this article, I found the idea of grouping students, concerning, as I could see some potential detriments for the lowest-performing students. However, upon looking at this specific data, it does not appear that those concerns are reflected within the data. That being said, just because the students did not show negative results within the context of these studies, does not mean that placement in the lower levels of groups did not negatively impact student self-esteem.

Ability Grouping IMG003.png

One final concern, with ability grouping, is how it affects gifted students. I believe that most teachers would find it reasonable to hypothesize that grouping gifted students together, might benefit those gifted students. Grouping gifted students together can allow for teaching those students more advanced curriculum, for more normalized high expectations, and for more sophisticated classroom conversations. There have been six meta-studies looking at this specific topic. Kulik and Kulik did a meta-analysis of this topic in 1982, 1984, 1985,  1987, and 1997. They found Hedges g effect sizes of .32, .47, .32,  .32, and .40 respectively. Goldring conducted a meta-analysis of the topic in 1990 and found a Hedges g effect size of .35. The average effect size of all these studies is .36. While this is statistically significant effect size, it is still quite a low effect size. Ultimately, I think it is fair to say that grouping gifted students together might provide them with a slight academic benefit. However, I am not sure whether or not this would outweigh any potential negative socialization repercussions. 

Ultimately the nerd in me finds the whole debate of this topic fascinating. I can definitely see how the execution of ability groupings could define its success or failure. On the other hand, the effect of grouping students by their ability appears to be so inconsequential, that I do not think teachers should really worry one way or the other about its impact! I am sure this stance will upset some scholar somewhere, who has spent their career studying the topic, but ultimately, I think realizing what interventions are not worth our time investment, is just as important as realizing which ones are!

Written by Nathaniel Hansford
Last Edited: 12/24/2020

Good, Thomas L., et al. “Expectation Effects: Pygmalion and the Initial 20 Years of Research                1.” Educational Research and Evaluation, vol. 24, no. 3–5, Apr. 2018, pp. 99–123, doi:10.1080/13803611.2018.1548817.

Steenbergen-Hu, S. (2016). What One Hundred Years of Research Says About the Effects of Ability Grouping and Acceleration on K–12 Students’ Academic Achievement. Review of Educational Research, 86(4), 849–899.

Gold, M. E., & Richards, H. (2012). To Label or Not to Label: The Special Education Question for African Americans. Educational Foundations, 26(1/2), 143–156.

J, Hattie. (2018). Hattie Ranking: 252 Influences And Effect Sizes Related To Student Achievement. Visible Learning. Retrieved from <> .

Copyright © 2018 Pedagogy Non Grata  - All Rights Reserved.

bottom of page