Impact of Programming on Primary Mathematics Learning

In this episode I unpack Laurent et al.’s (2022) publication titled “Impact of programming on primary mathematics learning,” which describes a randomized control study that compared the impacts of learning mathematics with an integrated CS and mathematics class.

  • quote in summary the present RCT study

    shows that the use of computational

    thinking via programming as a vehicle

    for learning mathematics in grades 4 and

    where math is taught for itself the

    transfer of learning although critical

    is difficult to achieve even when the

    presumed best conditions are met to

    facilitate it by relying on a near

    mathematics in both situations and high

    road transfer optimized by an explicit

    guidance of the teacher thus visual

    programming languages should be

    introduced with caution if they are

    intended to replace regular mathematics

    teaching the presented results are a

    strong indicator that this might be

    deleterious to mathematics acquisition

    end quote that's from page 8 of the

    publication titled impact of programming

    on primary mathematics learning which

    was written by Manon Laurent rosemaria

    Christie Pascal braceu amid cha chawa

    Sicily nura Erica DeVries and Pierre

    chonikai I most likely butchered all

    those names my apologies for that here's

    the abstract for this paper quote the

    AIM of the study is to investigate

    whether programming activity might serve

    as a learning vehicle for mathematics

    acquisition in grades 4 and 5. for this

    purpose the effects of a programming

    activity an essential component of

    computational thinking were evaluated on

    learning outcomes of three mathematical

    Notions euclidean division n equals 1880

    additive decomposition n equals 1763 and

    fractions n equals 644. classes were

    randomly assigned to the programming

    with scratch and control conditions

    multi-level analyzes indicate negative

    effects effect size range negative 0.16

    to negative 0.21 of the programming

    condition for the three mathematical

    Notions a potential explanation of these

    results is the difficulties in the

    transfer of learning from programming to

    mathematics end quote biota summarizes

    this study into a single sentence I say

    that this randomized control study

    compared the impacts of learning

    mathematics with an integrated Cs and

    Mathematics course you can find a link

    to this paper in the show notes which is

    linked in the description of whatever

    app let you're listening to the Sun or

    you can visit jaredoleery.com where

    you'll find hundreds if not thousands of

    free computer science education

    resources including well over 180 other

    podcast episodes at the date of this

    recording which has solo episodes like

    this one as well as many interviews with

    some awesome folks in computer science

    education and outside of the field

    talking about education in general and

    if you don't know who I am my CV is

    linked on my website but I've worked

    with evergrade kindergarten through

    drill students in a variety of contexts

    from music education to Computer Science

    Education if you've listened to this

    podcast for a while you know that I've

    had a very strong interest in exploring

    integration having previously developed

    professional development used by

    hundreds if not thousands of teachers

    around the country as well as writing

    curriculum that's been used by a few

    hundred thousand students around the

    world I'm well aware that many people

    are very interested in the idea of

    integration however as I've mentioned in

    many of the episodes on integration it's

    not really clear what people mean by it

    kind of like how computational thinking

    is a very vague term where like if you

    say to one person they might have a

    different understanding than a different

    person same thing with integrating there

    are many ways that you can integrate

    I've mentioned this on many episodes

    says but I personally don't think the

    approach of just kind of like inserting

    or crowbarring computer science

    computational thinking into the

    classroom like a math class or a science

    class or Ela or whatever is going to

    improve both that subject area and

    computer science or computational

    thinking if anything I think it might

    lead to overwhelm but there's not a lot

    of studies that actually look at how the

    two kind of work together or don't in

    those kind of contexts so this study

    kind of actually explores that in terms

    of what does it look like if we take

    computer science Concepts and teach

    mathematics with it in comparison to a

    class that is only focusing on math

    without computational thinking or

    computer science now the authors in the

    introduction mentioned that there's not

    really any kind of conclusion on whether

    or not mathematics and CS integration

    has a positive negative or neutral

    effect if it were positive great then we

    can see that hey if you incorporate

    these two or integrate these two

    different subject areas they'll have a

    better result than if you do not do that

    and that better is kind of like

    depending on what you're looking at

    maybe it's more motivation or higher

    test scores or whatever ever there's

    many factors in there but another good

    result could be hey it has no negative

    impact it has a neutral impact on there

    even though you might be taking away

    some time from the subject area by

    incorporating another domain you're

    learning two things without at the cost

    of the first one but we also don't

    really have enough research to find out

    whether or not it actually has a

    negative impact which is what I would

    guess might happen depending on how

    things are integrated now the author's

    note in their introduction that of the

    studies that do look at integration it's

    usually looking at it from a computer

    science standpoint to see like hey if we

    integrate computer science into Ela

    classes what kind of computer science

    Concepts and practices can you learn in

    those settings not necessarily what kind

    of an impact does it have on Ela when it

    does look at that kind of impact it

    doesn't necessarily do it in a

    randomized control study which is

    admittedly difficult to get without lots

    of funding like NSF grants things like

    that now when it comes to computational

    thinking many people talk about how well

    this like pattern recognition and

    abstraction and Etc is all going to be

    helpful for understanding X Y or Z

    subject areas and while that might sound

    fancy those claims aren't always backed

    by Empirical research it's like for

    music educators like there was this like

    music makes you smarter campaign that

    was going on I believe that was the

    wording that was going on like a decade

    or two ago and that was actually based

    off of faulty research like the whole

    so-called Mozart effect basically any

    stimulant is going to make it so that

    you're going to perform better on a test

    whether you do jumping jacks or listen

    to Mozart and it's only going to have an

    effect size depending on how you listen

    to it and for a short period afterwards

    and the same can be said with some of

    the studies on computational thinking or

    not necessarily studies but some of the

    research it might be a hypothesis that

    hey this skill of abstraction is useful

    it would probably be useful in this

    other setting that would be called

    transfer of the transfer effect and we

    don't really have a lot of studies that

    look to see does this impact

    understanding of abstraction and does it

    actually transfer over into another

    domain a lot of the studies that do look

    at transfer find that the skills do not

    necessarily transfer over or the

    understanding don't transfer over and

    this is just from like integration

    studies in general and like different

    transfer effects and things that I've

    looked at in like sports psychology and

    educational psychology

    Etc everything that I was just saying

    you can find a study or two that is

    going to say actually Jared this says

    the other way and that's great but it

    all comes down to we don't really have a

    conclusion on this yet we're still kind

    of figuring things out I happen to be in

    the camp where I'm very skeptical about

    it not only from a learning standpoint

    but also from like a power structure

    standpoint and whatnot in terms of like

    are we putting domains and subservient

    relationships to other domains and I've

    talked about that in other episodes that

    I link in the show notes but I keep

    bringing up these integration studies

    because again I think it's important for

    us to explore these now here's a quote

    from page three that is kind of like a

    little bow on what I was just kind of

    saying quote moreover results on the

    effect of programming to learn

    mathematics should be viewed with

    caution because among the majority of

    studies interested in the effects of the

    use of programming through mathematics

    lots are targeting the effects of

    computational thinking skills and not on

    mathematical learning among the few that

    are targeting mathematical learning few

    are randomized controlled trials based

    on empirical measures of learning

    outcomes in mathematics end quote okay

    so why is it so important to have

    randomized control trials well you might

    have a study where you you're looking at

    case studies multiple case studies doing

    some qualitative research or even like

    looking in a specific District or class

    or school or whatever those are great I

    love those studies in terms of being

    able to understand perspectives and

    whatnot and to understand how does this

    very specific population respond to a

    treatment or whatever but a randomized

    control study makes it so you can

    actually make some generalizable

    statements to say using the statistics

    that we've went through we found that

    this would likely have an effect outside

    of this population for X Y and Z

    characteristics or situations context

    Etc so if we had more randomized control

    studies and specifically more

    longitudinal randomized control studies

    in education not just like something

    that happens in like a six week period

    or something like that but maybe a six

    year period and then we can actually

    start to look at cause and effect in

    education and that's something that we

    talked about with like many other people

    like Andrea steffic in our interview

    which is all the way in episode 27 which

    is titled accessible CS education

    through evidence-based programming

    languages with Andrea stefek so I highly

    recommend checking out that one as well

    as all the other interviews because

    there's some awesome guests on this

    podcast okay so now we've kind of talked

    about the introduction like abroad or

    overview of things so let's take a look

    at well what was actually happening in

    this particular study so in this study

    they looked at three different

    mathematical Concepts so these were

    euclidean division additive

    decomposition and fractions each of

    these had a pretest and a post-test that

    were administered by teachers to an

    entire class in total there were 107

    classes from 46 different schools and

    each one of the schools was assigned to

    a treatment and control group the

    treatment group would receive

    mathematics instruction integrated with

    computer science whereas the control

    group would only do mathematics like by

    writing algebraic equations Etc and this

    was with grades four and five so this is

    like ages 9 through 11 with a total of

    really interesting tables and a lot of

    interesting details that are kind of

    explored in terms of like what are the

    means and standard deviations between

    the treatment and controls Etc I highly

    recommend taking a look at that I'm

    going to skip through a lot of this

    because I don't think it'd be very

    interesting in an audio podcast I'm

    going to read for you a quote on the

    bottom of page six quote the effect of

    the programming group is systematically

    significant and negative for all final

    models belonging to the programming

    group leads to a significant decrease in

    final performance for each Concepts in

    comparison to the control group this

    decrease is 0.16 standard deviations for

    euclidean division 0.19 standard

    deviations for additive decomposition

    and 0.21 standard deviations for

    fractions all other things being equals

    it explains 10 percent of the between

    Class variants of the final score for

    euclidean division 6.3 percent of the

    same variance for additive decomposition

    and 17 percent of this same variance for

    fractions end quote against from page

    six so in other words the group that was

    doing math and CS performed worse than

    the group that just did math now in page

    seven in the discussion section the

    authors have three potential

    explanations for this so let's explore

    each one of those alright so the first

    reason that they may have scored lowers

    because as teachers are spending more

    time teaching scratch or teaching

    students how to use scratch than they

    were actually teaching mathematics so a

    question that might be explored is over

    time like an extended period of time

    several years let's see what happens in

    these treatment and control groups does

    the mathematics group only perform

    better than the mathematics integrated

    with computer science or computer

    science integrated with mathematics what

    about some of the other approaches that

    we've talked about in other integration

    episodes where they're going in parallel

    so we have mathematics teaching a

    concept which is then reinforced by

    computer science class which teaches the

    similar concept from a different

    perspective or different angle which

    approach to integration work better or

    worse again we need longitudinal studies

    to figure out okay how does this

    actually have an impact down the road

    but this is an important thing to

    consider maybe if this was a longer

    study and it was lasting for like at

    least a year the teachers had more

    experience with it or students had more

    experience with scratch maybe they would

    have spent more time on mathematics in

    the class and maybe they would have

    scored just as well as the class that

    only did mathematics or maybe they would

    have been more motivated to want to

    continue to learn mathematics like that

    in and of itself even if it had a lower

    score I would argue is going to make it

    so that students would want to stick

    with it longer have more grit and

    actually enjoy the subject area in the

    long run as opposed to someone like

    myself who scored well on math classes

    but I never pursued it Beyond any of the

    requirements that I was mandated to take

    maybe if I had an integrated computer

    science and math class maybe I would

    have found a passion in it another

    explanation is the amount of time that

    was on task so having worked with

    scratch with a variety of students from

    second grade all the way up through like

    college students there can be a lot of

    off task time if we might think of it

    that way where you are just exploring

    the different features and functions

    like well what happens if I reverse the

    sound effect or what happens if I change

    the different costumes as opposed to

    like focusing on mathematics and

    computer science integration you're

    focusing on all the little neat things

    that you can do in scratch which is

    great it gets people interested in the

    platform wanting to learn things maybe

    they'll create more stuff at home

    Express themselves blah blah blah blah

    huge fan of that but if we're just

    trying to figure out whether or not in

    integration with scratch and math is

    going to have any kind of an effect

    positive negative or neutral the amount

    of time that students spent like

    exploring those different features would

    have potentially taken away time from

    them actually exploring the mathematics

    concept skills understandings or

    whatever so that's a good second

    potential explanation for this the third

    potential explanation that the authors

    mentioned it has to do with like memory

    management so cognitive load Theory

    which is an idea that you can hold I

    believe seven plus or minus two things

    at any given moment so like if I ask you

    to memorize a phone number you might be

    able to do that hold on to that in your

    memory but if I ask you to do that while

    walking on a tightrope while reciting

    the alphabet backwards and like doing

    that well I don't know scratching your

    head while patting your tummy the odds

    of you being able to remember that phone

    number significantly decreased the more

    cognitive load that you have which is

    why the theory I believe was seven plus

    or minus two things it depends on like

    how complicated it is or whatever the

    task is ETC you can maybe do five and

    hold on to it or maybe do nine but

    usually it's around seven so going back

    to what the authors were visiting here

    is maybe with cognitive load Theory the

    idea of holding and learning a new

    mathematics concept while holding and

    learning a new computer science platform

    maybe combining those two is too much

    cognitive load on a student which is one

    of the things that we've talked about in

    many of the other integration episodes

    that I've done is that if you're going

    to teach students two new things

    simultaneously they're going to have to

    try and split their attention on

    something rather than teaching one

    either before you learn the other or in

    parallel with but disconnected so for

    example one of the things I've talked

    about quite frequently is if you're

    going to teach music and computer

    science it can be very difficult to

    learn music concept skills practices Etc

    while learning computer science one at

    the same time ask me how I know I've got

    a drum set right behind me have a PhD

    music education and I've taught Computer

    Science Education for almost a decade

    I'm very grateful I learned one before

    the other because I was able to apply my

    understanding with better ease than some

    of the students that I've seen try and

    do both simultaneously it certainly can

    be done but it can can be a little more

    frustrating which I've seen some

    research on that talks about how this

    can be very difficult for students when

    they are asked to learn two things

    simultaneously and it can lead to

    burnout or frustration faster than if

    they learn each of them sequentially or

    in parallel with each other rather than

    simultaneously but again other episodes

    kind of explore different approaches

    where you're learning stuff in parallel

    rather than simultaneously and that's

    kind of a very quick summary and

    introduction to the paper itself I do

    recommend taking a look at it it's very

    interesting and I do appreciate the

    three explanations that they provided at

    the end but at the end of these

    unpacking scholarship episodes I like to

    talk about some of my lingering

    questions and thoughts so one of them is

    what approaches to integration do you

    think work better than others we've

    talked about integration quite a bit and

    different approaches that you can take

    when you're integrating CS with another

    subject area Etc which of those

    approaches do you think would work

    better than others and why and if you're

    working with somebody who wanted to

    integrate CS with another subject area

    what advice would you give them would

    you warn them or would you encourage

    them why would you do one over the other

    and when would you do one over the other

    for example if the teacher is is an

    expert in both computer science and

    another domain that they are teaching

    would you encourage them to integrate

    the two if the teacher was novice to

    both computer science and another domain

    that they were being forced to integrate

    with would you encourage them to do that

    or would you talk to admin and go I

    don't know if we should be doing that in

    that way it might actually turn students

    away from both domains because maybe

    it's going to be taught in a way that

    would be discouraging or frustrating for

    students I pose these questions and I've

    certainly shared my own opinions on

    these in the different podcast episodes

    but there's not a right or wrong way to

    do this with education everything is

    great there are many different contexts

    and different nuances that would inform

    one District Source teacher's decision

    to do something different than another I

    just think it's important for us to

    think through these things and kind of

    share our discussions on it I also think

    it's important that we have more

    research studies like this that are

    actually looking at randomized control

    studies but also looking at it from a

    longitudinal standpoint like several

    years ideally unfortunately we don't

    really have a lot of those at least not

    in the field of Computer Science

    Education like in the NSF grants that

    I've worked on in terms of applications

    and stuff they often ask you to use like

    the what works Clearinghouse which

    usually has a bunch of like longitudinal

    studies like talking about here's the

    effects of this type of professional

    development on blah blah blah right but

    it's all on subject areas other than

    computer science so we often have to

    look at and say Well it worked this way

    in this other subject area so we think

    it would work this way in computer

    science professional development or

    curriculum or whatever and so it's just

    kind of a big guess we need more of this

    kind of research I enjoy studies like

    these and if you did too and you know

    somebody who might enjoy this as well

    please consider sharing this episode

    with them just helps more people find

    the free resources on my website

    jodaly.com thank you so much for

    listening stay tuned next week for

    another episode until then I hope you're

    staying safe and are having a wonderful

    week

Article

Laurent, M., Crisci, R., Bressoux, P., Chaachoua, H., Nurra, C., de Vries, E., & Tchounikine, P. (2022). Impact of programming on primary mathematics learning. Learning and Instrucction (82), 1-9.


Abstract

“The aim of this study is to investigate whether a programming activity might serve as a learning vehicle for mathematics acquisition in grades four and five. For this purpose, the effects of a programming activity, an essential component of computational thinking, were evaluated on learning outcomes of three mathematical notions: Euclidean division (N = 1,880), additive decomposition (N = 1,763) and fractions (N = 644). Classes were randomly assigned to the programming (with Scratch) and control conditions. Multilevel analyses indicate negative effects (effect size range -0.16 to -0.21) of the programming condition for the three mathematical notions. A potential explanation of these results is the difficulties in the transfer of learning from programming to mathematics.”


Author Keywords

Computatinal thinking, programming activity, mathematics learning, primary school, randomized trial, learning transfer


My One Sentence Summary

This randomized control study compared the impacts of learning mathematics with an integrated CS and mathematics class.


Some Of My Lingering Questions/Thoughts

  • What approaches to integration do you think work better than others?

  • If you were working with someone who wanted to integrated CS with another subject area, what advice would you give them?


Resources/Links Relevant to This Episode



More Content