A few years ago, myself and other members of the school community were invited onto an episode of David McWilliams’ TV3 series titled “David McWilliams’ Ireland: Could school be making you stupid?”. About midway through the discussion, a young man working in tech declared with evangelical certainty that Direct Instruction was the best teaching method. You know, like, ever.
According to The Glossary of Education Reform, “direct instruction (DI) refers to (1) instructional approaches that are structured, sequenced, and led by teachers, and/or (2) the presentation of academic content to students by teachers, such as in a lecture or demonstration. In other words, teachers are “directing” the instructional process or instruction is being “directed” at students.”
Well, the other day, a friend of mine posed a familiar question about how things are taught in Wicklow Democratic School. It reminded me of that TV show, and prompted me to write this post.
So, the tech guy on the TV show claimed that DI’s superiority over other educational methods had been proven as an incontrovertible fact, and he said so in a way that suggested anyone who didn’t know as much already had no business being in the room in the first place.
I confess, the forcefulness of his pronouncement put me on the backfoot. I was sitting there thinking ‘Has there been a new study? Did I miss something? Had I got it all wrong?’
That night, I started trawling the internet to find the edict from on high, the celestial memo, that the tech guy, and presumably everyone else working in education, had already received and which identified the one pedagogy to rule them all. But it turned out there are quite a few such edicts, and they don’t all agree with one another.
If DI Is Kong, Then What’s Godzilla?
The most prominent educational alternative to DI is called Constructivism, with a variety of subcategories such as Inquiry Based Learning, Problem Based Learning, Project Based Learning, and Active Learning. I’ll just call it PBL from here on. This type of education posits that, rather than just passively taking in information, learners construct knowledge and “build their own representations and incorporate new information into their pre-existing knowledge.” Constructivist teaching aims to “provide experiences that facilitate the construction of knowledge” and engage students in the learning experience.
PBL has no shortage of advocates claiming that research shows it is superior to DI, not to mention those who claim a combination of DI and PBL is the best approach to teaching.
To understand constructivist education a bit better, let’s consider the following experiment by Held and Hein, described by Michael Anderson:
In this experiment, two sets of kittens were raised in the dark, and only exposed to the light under very controlled conditions. When they were in the light, one set of kittens was allowed to roam freely, although each kitten from this first group was fitted with a harness, itself attached to a basket in which a given member of the second group of kittens was placed in such a way that only its head was free to move about. Because the kittens were raised mostly in the dark, both groups developed the same motor capacities and physical repertoire. Likewise, kittens from both groups were exposed to the same amount and sort of visual stimuli. However, only kittens from the first group were in a position to move about and see at the same time.
The results were quite striking, and indicated that the kittens from the second, constrained group had not developed an appreciation of the physical significance of their visual experience. For instance, when a normal kitten is held and brought close to the floor or a table-top, it will reach out its paws in anticipation of the contact. The constrained kittens did not display such reaching behavior. Likewise, kittens from the constrained group were more likely to bump into walls and fall off the edges of things, apparently not recognizing the significance of the relevant visual input.
The authors conclude that “practical activity, understood as a mode of epistemic access to the world, is a necessary underpinning of our general referential capacity.”
In other words, we have to meaningfully explore, and interact with ideas in practical ways if we are to meaningfully integrate them into our understanding of the world. This is essentially the constructivist, or PBL, view on learning. By touching and navigating their environment, the kittens in group 1 were able to understand it and therefore ‘construct’ it into knowledge they could use, but the kittens in group 2, as passive recipients of the information, couldn’t apply it in a way that was useful to them.
However, we (apologies for this bold assumption) are not kittens, and as a result the above conclusion doesn’t necessarily apply to us humans. So, let’s sift through what the research says and see if we can get a clearer picture of what does.
What does the research say?
There are few creatures more insipid than the radical moderate. The compulsive fence-sitter. The type of person who seeks out the two most popular opinions and decides theirs by finding the halfway point between the two. And then smugly reminds the room of how sensible they are, bemoaning the fact that no one else is as reasonable as them.
However, just because Mary from up the road says “2 plus 2 is clearly 5,” and Joe from below the cross says “No, no, no it’s obviously 7,” that doesn’t mean that the correct answer is 6. But sometimes it seems there’s no shortage of people whose entire analytical strategy amounts to little more than being seen to take the reasonable middle ground.
Of course sometimes the truth is in the middle. What I’m trying to avoid is coming down in the middle because it’s the safe thing to do and because it will make me seem reasonable whether the truth is in the middle or not. However, in this case, it is actually pretty difficult to sort the wheat from the chaff and identify the best approach to learning. That’s because when it comes down to picking a victor between PBL and DI, there are some problems with the nature of educational research itself. You see, educational research is arguably more difficult to do well than research in any other field.
Educator Donald Clark explains this well in a critical review of one particular comparative study of DI versus Constructivist teaching (which favoured DI):
Both the authors and the constructivism movement are guilty of jumping on theories before they are fully understood. But why do we do this? Joel Michael writes in Advances in Physiology Education:
“…it is important to recognize that educational research is difficult to do; this has been cogently highlighted by Berliner in “Educational research: the hardest science of them all.” Berliner points out that unlike a physics experiment, in which it is possible to readily distinguish between the independent and dependent variables, and also possible to isolate and control all of the independent variables, in educational experiments all of this is problematic.
Researchers may not agree on which variable is the dependent variable of greatest interest or importance. There may be disagreements about which independent variable(s) are to be manipulated. There may be disagreements about how to measure any of the relevant variables. And, finally, it may be extremely difficult, or even impossible, to isolate and manipulate all the variables suspected of being involved in the phenomena being studied.”
Rather than waiting for eons to pass before all the research is available, we (the learning, training, and educational community) often jump into a new theory because we simply do not want to wait until we are dead and buried before we can fix and/or improve our methodology.
A quick google search explains dependent and independent variables as follows –
- The type of soda – diet or regular – is the independent variable.
- The level of blood sugar that you measure is the dependent variable – it changes depending on the type of soda.”
Berliner is explaining that in educational research one rarely knows for sure which is the type of soda, and which is the blood sugar.
In summary, educational science is not as clear cut as we might like, so there are many pedagogical questions that do not (at least not yet) have clear, empirical answers. This means that, until they do, people’s values, experiences and a certain amount of conjecture are going to come into play when making decisions about the best type of education in a given context. One method might be right in one context, and wrong in another. Results from DI might be poor in a particular instance, but that could be because of a number of variables that have nothing to do with the teaching method. It’s not that we can’t know the right answers or that there aren’t any, there just may not be enough research of sufficient quality to base a conclusion on.
So, there are a number of pitfalls at play here that prevent the DI vs PBL debate being a simple google and gloat, open and shut case.
For example, some research that supports DI depends heavily on the contested theory of Cognitive Load Theory. A 2018 study called “The Effectiveness of Direct Instruction Curricula: A Meta-Analysis of a Half Century of Research” analyzed 328 studies from 1966–2016, and concluded that DI is an effective teaching method. This study has been used to assert that DI is the preeminent form of teaching, or at least the most empirically supported.
However, as educational author Larry Ferlazzo points out, “as the paper says, the paper ‘did not attempt to compare the results of each of the DI programs with specific other approaches.’ It seems to me that meta-analyses comparing the specific impacts different kinds of instructional strategies might have been much more helpful to educators. Yes, direct instruction might be effective (and, of course, we all use it sometimes), but could other methods be more effective?”
He adds, “I don’t think it’s at all clear – at least to me – from the paper what in practical classroom terms they use to define ‘direct instruction.’” What constitutes DI or PBL seems to vary quite a bit across the different studies that I’ve read so far. Imagine trying to discern which is the tastier food, fruit or vegetables – you look at all the studies on the subject to get the answer, but if some of them compare roast potatoes with pears, they could end up with very different results than others that compare brussel sprouts and mangos.
Again, I am not saying there are no answers to be found, but studies and headlines and tech guys on TV shows should be taken with a pinch of salt, and their methodology should be investigated before we find ourselves with very firm views on the issue. The same is true of this article. I’ve endeavoured to be thorough here, but this isn’t a formal literature review and it’s quite possible I’ve missed some important research or a relevant perspective.
There is certainly enough research on both sides to simply cling to the work that supports your bias. For example, Alfie Kohn highlights an instance where a 2004 study found that students taught with DI did better than their classmates who were allowed to design their own learning experience.
Kohn points out that the PBL element of the study was not set up in line with “the strategies most experts recommend for promoting discovery and exploration” but it nonetheless “may have given pause to progressive educators — at least until another study, published three years later, looked at the same issue in the same discipline for students of the same age. The second study, however, investigated the effects after six months instead of after only one week, and it also used a more sophisticated assessment of the students’ learning. It turned out that any advantage of direct instruction soon evaporated.” In some cases, he added, the 2007 research suggested “that direct instruction can be not merely ineffective but positively counterproductive.”
How do I know another, better study won’t come along and flip things on their head again? I don’t. That’s why it’s so important that as educators, or parents, or students, we keep open minds and avoid becoming dogmatic on these issues like the infamous tech guy from TV.
How many people saw a headline about the 2004 research and made up their minds, never hearing about the 2007 research, or being closed to it when they did?
Long Term Impacts
One study, however, did give me pause.
A number of articles anecdotally suggest that novice learners do better with teacher-led instruction, and once they’ve developed a basic competence, they are ready for more freedom in directing their own learning.
Alfie Kohn reviewed the research around this idea and demonstrated quite convincingly that many of the most famous studies supporting DI are an absolute mess in their methodology.
More revealing is the following, which is so striking I’ve included the quote from Kohn in full:
Frankly, given how much happens to us over the years, it would be remarkable to find that any single variable from our early childhoods had a long-term effect. That’s why the results from another such study are nothing short of amazing. Back in the 1960s, a group of mostly African-American poor children from Michigan were randomly assigned to DI, free-play, or High/Scope constructivist preschools. They were followed from that point, when they were three or four years old, all the way into adulthood. As in the Illinois sample, the academic performance of the DI children was initially higher but soon became (and remained) indistinguishable from that of the others. By the time they were 15 years old, other differences began showing up. The DI group had engaged in twice as many “delinquent acts,” were less than half as likely to read books, and generally showed more social and psychological signs of trouble than did those who had attended either a free-play or a constructivist preschool.
When the researchers checked in again eight years later, things had gotten even worse for the young adults who had attended a preschool with a heavy dose of skills instruction and positive reinforcement. They didn’t differ from their peers in the other programs with respect to their literacy skills, total amount of schooling, income, or employment status. But they were far more likely to have been arrested for a felony at some point and also to have been identified as “emotionally impaired or disturbed.” (Six percent of the High/Scope and free-play preschool group had been so identified at some point, as compared to a whopping 47 percent of the DI group.) The researchers also looked to see who was now married and living with his or her spouse. The results: 18 percent of the free-play preschool group, 31 percent of the High/Scope group, and not a single person from the DI group.
None of the studies I’ve come across so far on DI investigate the social emotional impact of teaching young people in this way, only its impact on their ability to memorize information and then regurgitate it. The study above, and indeed my experience in both the Irish and Spanish mainstream education systems, which rely on DI heavily, show that it may have the potential to inflict actual harm. In Wicklow Democratic School, it sometimes feels like half the job is helping students heal from their experiences in mainstream education.
I’m quite open to DI, I think it may have its place in a full repertoire of teaching skills. Mainly because it’s easier. Given ample time and resources to prepare a perfect activity or workshop, I would rarely choose to use DI. However, such Goldilocks circumstances rarely arise, and so, DI can sometimes be the ‘good enough’ solution to fall back on in a pinch. That, to me, is probably okay, so long as it doesn’t become a crutch. But the research above raises serious questions about pushing DI as the best method for an entire course, subject, or curriculum, and it certainly casts doubt on the idea that it’s the best approach for novice learners. “First, do no harm” isn’t just for doctors.
Gold Star Research
Further evidence emerged in favour of PBL with the recent publication of two gold standard (which means the studies were randomised and controlled) studies involving 6000 students across 114 schools, where over 50% of students involved came from low-income households:
In one example, students in Amber Graeber’s AP Government class took part in a simulation of an electoral caucus. Meanwhile, instead of simply reading about Supreme Court cases, students in Erin Fisher’s class studied historic cases and then took on real-world roles, arguing the cases in mock court, acting as reporters, and designing campaign ads and stump speeches to make their case.
Researchers found that nearly half of students in project-based classrooms passed their AP tests, outperforming students in traditional classrooms by 8 percentage points. Students from low-income households saw similar gains compared to their wealthier peers, making a strong case that well-structured PBL can be a more equitable approach than teacher-centered ones. Importantly, the improvements in teaching efficacy were both significant and durable: When teachers in the study taught the same curriculum for a second year, PBL students outperformed students in traditional classrooms by 10 percentage points.
When taught with PBL, struggling readers and highly proficient readers alike outperformed their counterparts in traditional classrooms. Significantly, this held across both primary and second level education, providing further refutation to the idea that DI is better at earlier and lower skill levels and suggesting that PBL is better in general.
Again there could always be another angle, but this is looking at the most comprehensive research on the issue to date.
It’s important that I clarify any biases I may have. Of course, working in a democratic school, my hope was that the research would reflect the benefits I have seen PBL have. That said, I have, as I always do, attempted to avoid any confirmation bias in writing this article, and indeed, about midway through writing, I thought I’d have no choice but to come down on the fence.
However, after diving deeper, the data seems, at least for now, to be in PBL/constructivism’s favour. The research demonstrating the most convincing support of PBL is of a high quality, particularly the recent gold standard studies and the longitudinal work cited by Alfie Kohn. In contrast, research supporting DI has a tendency to be limited in its measure of learning success and lacking in consideration of social emotional outcomes, and it sometimes straw-mans PBL by comparing DI to PBL methods that are not considered best practise. The fact that PBL is ever able to outperform DI on standardized tests designed with DI in mind is interesting as well, like one plant outgrowing another that is planted in more fertile soil.
Furthermore, I find it interesting that, in my own facilitation of classes, I do sometimes use DI, but it tends to be something I fall back on when I have less time to prepare a class or activity. The more time I have, the more my classes tend to look like PBL, and the better they go. This is of course, totally anecdotal and could mean anything or nothing at all.
Once again, it’s not an open and shut case. As we saw, educational research is difficult to do well and there have been a few cases of a study supporting one method, only to be refuted a couple of years later by a new study. Nevertheless, I believe it is worth trying to figure out what the most effective learning methods are, and as Nobel laureate Carl Wieman says, to cultivate the “mindset that teaching should be pursued with the same rigorous standards of scholarship as scientific research.”
At Wicklow Democratic School, we will continue to try and be as open as possible to different perspectives and ideas. We will continue to learn and search for ways we can improve, whether that involves subtle tweaks or occasionally upending the tea table, with the only constants being our democratic commitment to every voice being heard with empathy, and of course, change itself.
For now, most of what we do looks a lot more like PBL than DI. We have our co-created curriculum and most classes are mostly constructivist. We have some activities that are structured and others that are less so. I believe we have had some great success, and I can also see tremendous room for growth. We started out with a very laissez faire approach, but over time, students and staff felt more stunted than freed by this. The beauty of a democratic school, as I’m always saying, is that we have been free to adapt and evolve beyond that dip in the road.
Now we’re more excited than ever to explore all the ways we can improve as we continue to play our part in reimagining education for the 21st century. Could we apply more of an inquiry-based approach to creating the curriculum together? What can we learn from PBL approaches like POGIL, Scale up, High Scope or the work of researchers like Michael Fielding on democratising public schools? We don’t have to imitate any of these methods wholesale, rather we can pick and choose the bits that work for us and adapt them to our specific, Irish context.
At Wicklow Democratic School, we embrace the idea that it’s not only the students who are learning – the staff members are learning right along with them, as any good educators should. So here’s to ‘constructing’, a ‘direct’ route to our future!