Will Brehm 1:33
I just like to welcome Nelli Piattoeva, Ezekiel Dixon-Román, and Noah Sobe to this Datafication of Comparative Education webinar. In many ways, this is a live FreshEd show. That’s the other hat I wear outside of Waseda. So, just to kick things off right away, I just want to ask – let’s start with Nelli: How do you conceptualize algorithmic governance or the datafication of governance, and why is this an important topic to study?

Nelli Piattoeva 2:02
Thanks, Will. Very happy to be here, and the first time for me ever to do a webinar. So, for me datafication is about translating a phenomena into quantitative formats so it can be then entered and re-entered into different numerical formats, and then interpretations can be made about the phenomena that has been quantified. It’s also about using data and datafication to redirect thinking and action of people just by the mere act of collecting data on one aspect and not the other. So, making something countable and something else kind of not countable, not important to be counted. Datafication also increasingly happens in implicit ways, so actually most of the time nowadays, we’re not even realizing all the data that is collected on our activities, either as private citizens or professionals. And so, it’s also very much an implicit process that we need to understand also as something very much intangible. And why this is important – I think it’s changing very much … the very basis of society and education. So, for instance, datafication is very much reliant on learning machines – machines that learn from data. And so, we start thinking about things like, “who is human?” if machines can learn, and sometimes they claim to learn even better than humans. So, what distinguishes a human from a machine? So, these kind of philosophical questions become quite timely. And also, data is a huge business. Data has been said to become the new oil, and in that sense, it’s a huge machine of value extraction and it’s a way also to redistribute income between people, between society. So, it is profoundly changing the basic conditions of life, and I think we need to understand how and what can be done about it.

Will Brehm 4:27
And Ezekiel, would you add anything to Nelli’s conceptualization and significance of this topic?

Ezekiel Dixon-Román 4:35
Yes. Thank you, and good morning, good afternoon, wherever you are in the world. And thank you all for organizing this webinar. It’s an important and great topic. So, I want to start with really thinking about the relationship of calculation and governmentality, and the ways in which the very logic and rationality of calculation has long been the practices of governmentality. We know that in Western Europe, the practices of bookkeeping for accounting was recognized and appropriated by the state as an effective practice of surveilling and managing wealth in populations. These logics of governmentality were informed by enlightenment ideas of reason, truth, and universal conceptions of being in subjectivity and even of the human to sort of give a head nod to some of the questions that Nelli was just raising. As well as theological moves, quite interestingly, in the philosophy of science that postulated number and probability to be the logic and language of the natural world, hence making the natural world what they call “the book of the universe”. This led to the later positivist development of the social scientist and the quantitative imperative. The same enlightenment ideas that led to the quantitative imperative for the social sciences is also what informed interest in appropriation and practices for calculation in governmentality. This also later informed ideas of cybernetics as the science of communication and predictive control. Cybernetics informed not just the development of the computer, AI, machine learning, but also system theories of governmentality or governance, I should say, policy and society. So, it’s quite an interesting sort of “natural”, if you will, connection between algorithmic logical thinking or calculation, and governance in and of itself. There’s a long history to this already. And so, in some ways, I’m also sort of laying this historical context in order to suggest that I want to be careful about exceptionalizing the current moment in a way that it almost can suggest that there’s something particular of the epiphenomenon right now, in this particular moment, that in fact, the arch has been a long history that has been moving toward this very moment that we’re in. The only major kind of difference that we might point to is the advancements of technologies, and the powers of computation that has advanced these very possibilities of engaging in different systems of algorithmic governance. Last thing I’ll say is, I tend to think about algorithmic governance as cybernetic systems of communication and predictive control, particularly that have become part of the logic and rationality of governmentalities that I sort of laid out a second ago. This topic, I would say, is deeply important. I think there’s an interesting way in which there’s currently in public discourse, and even scholarly discourse, there’s this kind of this divided or even dualistic kind of discourse, where there is technological optimism that we get. And then we also get a sort of technological pessimism on the other hand, and I think both of them are very dangerous. If I were to lean on Paul Ricœur here for a moment, I would say, you know, hermeneutics of suspicion without empathy can be quite dangerous. And so, I do think we need to be critically studying this type of sociotechnical systems of cybernetic predictive controlling communication, but at the same time, trying to make sense of them and to try to think of ways that we can enhance them for addressing issues of equity, inequality, power, and even any forms of anti-colonialist kind of projects or imperialist kind of projects.

Will Brehm 8:30
So, Noah, I’ll just throw it over to you as well. By way of introduction, how do you conceptualize this algorithmic governance, and why is it important?

Noah Sobe 8:40
Sure. Thanks, Will, and thanks Chris, and Ezekiel and Nelli. We got some great introductions there. I think it is important to pull out some of the pieces here, and I completely agree with Ezekiel’s point that we shouldn’t exceptionalize the present; we should see sort of things that are happening as part of some sort of longer-term historical trajectories in governmentality and how governance works. At the same time, I think it is worth pulling out some sort of stages, and some differences, because I think there is a difference between the datafication of education, and also the algorithmic governance of human beings. And then there’s also the question that’s sort of in the title of the webinar – the datafication of comparative education, as a separate idea than the datafication of education. So, I think we’ll probably get to that. But just to focus on the first. So, data. I’m in Chicago. In the United States, you can tell a story about the rise of data in education. I think you can tell a similar story in the UK. You probably go back to the 1980s and you’ll talk about the requirements to collect and analyze information that are connected to a lot of accountability mandates. And for me, one of the best illustrations of this is the whole practice of data walls. So, if you’ve got Google handy, just Google that phrase “data wall” and do an image search. And what will probably come up are going to be pictures of teachers in front of boards that have multicolored tabs that identify students and their learning in relation to one another, and you’ll probably see representations of educators, kind of analyzing, thinking about the positions. And this, in fact, I’m not sure it’s a hugely widespread practice, but it’s just an interesting phenomena in and of itself that, at least in some areas of the United States and perhaps in other parts of the world, there is this idea that it is good practice for schools to set up these data walls so that teachers can monitor the learning and progress of their students. It’s also a great example of the way that data is embodied. That data enters into cultural practices, changes the work of teaching, the business of expertise, right. So, to me, that’s a really good example of one of the ways that some education has been datafied. And then, of course, you can take this to like a school level and consider all the currents. I mean it’s almost like school principals and district superintendents are sitting in front of like a commodities trader’s set of screens nowadays, where they’re weaving through all this data that comes in. So, I think that’s taking that datafication a further step. And then I think we should also – and Nelli’s comments got at this too – we should also talk about “big data”, which is yet a different kind of data analysis that’s relying on machines, that’s taking the digitization of data a step further. So those are the examples I would provide of the datafication of education. And I think algorithmic governance is tied into this, but it’s something sort of distinctly different. And I was trying to think of an example, and I could tell the story of getting on the TSA watch list, because I think I was captured by an algorithm where I bought a roundtrip ticket and only used one way, and then for about a year afterwards, got pretty intense searches. But I think a good example would be the notion of geofencing, right. This is possible these days. There are some American universities that have an app, right. This is not an uncommon thing – an app on students’ phones that lets them check grades, or whatever it lets them do. If you have the alerts note activated, and if you have the geographic reporting activated, there are universities that will geofence and send messages to people’s phones based on an algorithmic calculation. So, for example, why might you want to do this? Well, if it’s a Saturday night, and a student is going to a fraternity row, right, you might want to send them a text about the university’s bystander policy, right, or the university’s Good Samaritan policy. So, I’m not sure that lots of universities are doing this, but I do know that it’s being considered and talked about and perhaps even rolled out in a couple of places. But it’s a perfect example of the use of algorithms to govern behavior because it’s an analysis of past data. And humans are involved in creating these, right? Because they are creating this model that looks at data to predict behavior, right, and then is intervening with some sort of nudge that’s meant to reshape behavior. So, I think that that’s the key thing. And then of course, the algorithm is going to learn, it’s going to expand. It’s going to look at other geographic areas where maybe there are noise complaints. But I offer that as an example of the way in which like algorithmic governance works, in that it trolls previous information and learns, and then really what we’re talking about are things that are attempting to shape and reshape human behavior.

Will Brehm 14:54
So, from the deeply philosophical about ‘What are humans?’ or ‘Who are humans?’ to the deeply material where data is governing our everyday behavior, and perhaps social relations. So, let’s turn a little bit to the focus on education specifically. When we think about a lot of educational reform today, it’s data or big data, or AI, or machine learning, or 21st century skills that either explicitly or implicitly are invoked as some sort of necessary foundation for addressing social inequalities. Let’s start with Nelli: Can you speak to how you understand the relationship of data and inequality?

Nelli Piattoeva 15:42
Yes, that’s a very crucial question, as you said, the datafication processes that kind of rolled out and throughs with the rhetoric of curing inequality. And I have to say that from the research that I’ve done, for instance, in the context of Russia with my colleagues, we can say that whatever data exists, and even when it is, for instance, data on learning outcomes that can be analyzed in ways kind of that count the context – the social, economical factors for different schools – still, data is not a panacea. So political decisions need to be taken in order to address the issues of quality of education or provision of textbooks or good teachers. So, in the context of Russia that I know a little bit better, we can say that data exists, but data is not necessarily acted upon in ways that would be curing different kinds of inequality in education. So, teachers engage in a lot of reporting, students are tested on different kinds of contents, tests are invented for different ages, and still, we can say that lots of data on those lies in the graveyards, because still there is no political will to make things change. So that’s one of the kind of more like pessimistic views that I have about the ways how data is not actually … it can reveal things, it can show some patterns, but unless it’s acted upon, it’s not going to secure better equality for different student populations.

Will Brehm 17:50
And Ezekiel, what worries you about data and inequality or equality?

Ezekiel Dixon-Román 17:58
There’s a lot I can say on this, and I’ll try to keep it short. I’ll just begin with saying that I actually think one of the questions, or line of questions, that we have to deal with a little bit more in depth, as questions of the ethical, questions of fairness, and even questions of even what inequality is, and the context of data science. And let me throw another one in here – justice, even how justice even gets deployed in this context as well. There are – what is it? – over two dozen different perspectives on ethics, and so what ethical conception gets privileged when it comes to designing algorithmic systems for making particular kinds of decisions is always going to be not just a question, but actually intention. What communities are going to be interested in what ethical perspective or position? And also, what ethical decisions are going to be important in what context? So, whether we’re talking about in the context of college admissions, in the context of classrooms for making decisions around assessment or grades or whatever, but also in the context of child welfare, or predictive policing, where there are high stakes decisions that are being made on people’s lives as we speak. So, I think there’s some serious questions that we need to deal with around the ethical fairness questions of even what justice is. For me, I often tend to think about justice as a empty signifier that often becomes hegemonized by a particular dominant interest, and similarly that also have implications for inequality. But I want to be careful not to also fall into the trap of presuming data to be bad. Or even inherently non-humanistic. Data comes in all shapes, sizes, and forms. As we are having this conversation, we are producing data in and of itself. And we each are processing and analyzing that data in our own way, from our own perspectives, from our own sort of interest, if you will, and then making inferences off of it and finding ways of all sort of responding and engaging. Data and measurement from an anthropological philosophy perspective is part of our quotidian lives. So, to almost miss that is almost to make data something quite other than what it is: part of the very process and even enterprise of even being human. Obviously, the neoliberal discourse that has undergirded the discourse on data driven policy and practice makes particular assumptions about data as well as the social world through its gaze on outcomes of privileges that ends as a structuring force for regulating inequality, and inequities as well. I want to be very intentional about emphasizing the fact that it’s not just about inequality, the discourse around education, around the achievement gap. And educational inequity is often about … its focus is the language of inequity or equity. But in fact, what really is the focus in actual policy practices, and even research has been on inequality. There’s been this conflation between inequality and inequity. And so through its gaze on outcomes, it privileges the ends of the structuring force for regulating inequality and inequities while overlooking the means, right. So, in fact, it becomes a conception of justice that is completely focused on assuming that the ends can address all issues of justice and inequality in society. And those ends tend to be market based, focused ends. Moreover, the modernist gaze on data tends to be toward homogenizing difference, producing colonialist narratives about statistical difference that positions one group over and against the other. Finally, data driven not only assumes data to be self-evident, but also misses the fact that data and algorithmic acts have performative effects in shaping or reconfiguring the social world. So, in all of this, I think it’s critically important that we take seriously and really, I think there needs to be a really a serious potent on research, nationally or internationally. And some have been moving in this direction, such as the FAT* conference group – fairness, accountability, and transparency, and data science group that folk out of Cornell and other places have organized, and have started a few years ago, to really think about what does ethics mean in the context of data science? What is fairness in the context of data science? How do we address issues of transparency? And is transparency enough, particularly for addressing issues of equity, fairness, and justice? I’ll stop with that, but there’s a lot that can be said here; I think that might be enough, at least to stimulate some thought.

Will Brehm 23:38
So, I’ll move it over to Noah and kind of pick up on this issue of thinking through what is ethical? What are inequalities when we think about data? Is it data that is producing these inequalities, or is it actually more of the neoliberal discourse, the modernist gaze, that homogenizes differences? So, Noah, what are your thoughts based on what some of Ezekiel and Nelli have previously said?

Noah Sobe 24:04
Sure. So, I mean to try and bridge off that, I mean clearly algorithmic governance like enforces codes of normalization, and I think Ezekiel just went into a sort of a good discussion of some of the dangers of that – that we need to be attentive to. So, the second thing I would say is, also clearly data are socially produced; they’re not sort of out there to be found, right. And because they’re socially produced, they are laden with all the baggage that everything else that’s socially produced is laden with, okay. I guess what I would add is a third point, and this is, to me, where like a whole set of ethical issues come in.  Actually, Orit Halpern – there was a piece that was suggested for reading alongside this. I found it a difficult piece, and I was glad that I’ve actually read the full book, Beautiful Data, which makes a really, I think, thoughtful argument about where we’ve come such that knowledge, and the production of knowledge, has been recast as data analysis. So, I would say that the problem and what sort of the sticky part about datafication is that it’s not just about data, it’s about what happens with data and what data does, right. We are in a mode right now where like data analysis has become a knowledge generating practice, and we’re in a moment where data analysis is considered an ethical and truth producing practice, right. So, database decision making is good; like we should do database decision making. That’s the argument that’s made out there. I think Ezekiel got into a lot of the ethical questions that come up. I mean I absolutely agree on a certain privileging of the ends, but I would also add that there’s similarly like a “technicization” of the ends, where we do not actually like … or too rarely do we sit down and talk about what the end should be. And said, data analysis is just simply like the application of technical procedures to a problem that’s understood as technical without any kind of, in many instances, like careful engagement with like, “What should we be trying to do?”, right. And so, when data analysis basically colonizes knowledge, that’s what gets moved off of the table, right. Are the kinds of conversations we’re hoping to have here?

Will Brehm 26:49
So, Noah, I actually want to follow up with that, data as colonizing knowledge. One of the audience members from our conversation today has posted a question about “Does data actually look different across cultures?” Or is this a somehow colonizing effort that looks very similar?

Noah Sobe 27:09
I think in one sense, yes, absolutely. Data looks different. Different patterns of data analysis look different. If you think about accountability data, audit cultures look different in different places. At the same time, there certainly are like normalizing global institutions, actors and networks that are homogenizing what it is that is acceptable data, right. We see this in many communities around the globe, where parents and families and children have different views on what are the markers, the evidence, the indicators of successful schooling than teachers or policymakers or administrators do. So yes, I would say I think unquestionably, there’s a huge variety in sort of data in the world. But I think connected with that, or alongside that, there’s also clear differences in which kinds of data are imbued with authority, right. And certain kinds of data are imbued with more authority than others. So, a good example would be like all the indicator data that comes out of international large-scale assessments like PISA and so forth. I mean that data carries a lot of authority in the world, and yes so that’s the answer to that question. I’d be curious what my other panelists think on that.

Will Brehm 28:43
Yes, so Ezekiel, what do you think on that? I mean, do we see these big differences in data across cultures and, in a sense, is data actually some sort of lived social reality?

Ezekiel Dixon-Román 28:55
Yes. Period. So, I mean I hate to be so blunt. But whether it’s a measurement instrument, whether it’s an algorithm, whether it’s some kind of – I’m just going to say for a broader scope – anything that is designed to produce information, to produce data, there is a sociopolitical process that goes into it, there’s a sociocultural process that goes into it, and there are sociocultural assumptions that are made about the very construct or things that are the object of measurement itself. And as such, for me I tend to think about measurement as a performative process. So not in a modernist sense of trying to get at some kind of truth or singularity, but in fact, measurement is engaging in performative acts that are enacting, forming and shaping the subjects that they are, in fact, interacting with. And as such, the way in which an instrument of measurement of quantification is performatively enacting on one subject is not the way it’s going to performatively enact on another subject. And, it is in fact, the very sociocultural and sociopolitical constitution and architecture of the instrument and I even say sociotechnology of quantification that even produces that kind of variability. ‘What do we do with that, though?’, right. So, I think that’s a whole another question, right? On the one hand, the way we often would sort of write that off, if you will, is to say that there’s bias there, or critique it is to say that there’s bias there, that there’s measurement error there. And I think that’s also a problematic move as well. In fact, bias and measurement error come out of a discourse of modernist epistemological truth, right. So modernist epistemology, it assumes an element of measurement error, then would assume that there is some kind of truth in that very construct that we’re trying to get at, around with that … in accounting for that measurement error. And bias also is situated in the same line of discourse. So, I think we need to also do some work on rethinking what even measurement is doing to bodies, to humans, to … Let me be careful, I don’t want to say “humans”, to more than human ontologies. Because in fact, we’re not just collecting data on humans, we’re collecting data on even the algorithms itself and other entities itself, right. And so, I think we do need to tend with what in fact, particularly for those that are situated and whether it’s postmodern thought, whether it’s post structural and postcolonial theory, new materialisms, posthumanism, whatever it might be. Anything that’s working against modernist epistemologies needs to tend with, “What is this thing that measurement is doing? Is it really biased, or is it something else?” How do we make sense of that, and what do we do with it?

Will Brehm 32:09
Nelli, I actually want to ask you about one of these ways that the data can look different across cultures is its interaction with, or construction of, particular forms of citizenship and identity. So, I know you do some work on those notions. So, how are data redrawing the boundaries of citizenship?

Nelli Piattoeva 32:34
Yes, that’s a very interesting question. This is a question that I would say need to be studied empirically. Because again I think in different contexts, the facts of datafication on citizenship would be quite different. So, for instance, in the context of Russia, where kind of nationalistic, patriotic policies have been on the rise in the past 15 years, one of the kind of manifestations of the importance of data in education policies has come with the idea of increased testing and examinations, compulsory examinations, of students. And so recently, for instance, one of the discussions in politics has been about, “Why not make the history test compulsory for school graduates?” So, you can say that through making history as a test compulsory and at the same time, in the context of this kind of political pressure to narrow down the interpretation of history, to narrow down the understanding of the role of citizens in a kind of neo-totalitarian society, datafication actually becomes a means to construct a kind of one rigid, controlled identity of the Russian nation. And this, I think, speaks against some of the interpretations that datafication is a process of breaking down collective identities, because of the capacity of data to individualize, to identify people in very kind of particular, very kind of limited ways. So, again, like in the context of Russia, I would say that how datafication is recontextualized there, it is actually used by the government to bring about, together with other policies, a more unified, but also a more kind of limited and state related identity of citizens. But at the same time, when we look at some other processes, like something that the press has been writing lately about the Chinese social credit system, where people, individual citizens, will be scored through different data traces that they leave either as purchase power, or criminal record or some other issues. So there, like the datafication has a very embodied and a very individual effect, and I think it breaks collective identity of people in terms of imagining some kind of political action against the state, because you are just a private citizen scored in a way that you don’t understand, and your neighbor is going to be scored in some other ways. So there, I think, again, datafication can have a very different effect on citizenship identity. And of course, we can also talk about datafication from the perspective of citizen rights, responsibilities about privacy, or in terms of collective action. And I think it’s a really huge topic all in all.

Will Brehm 36:20
And Noah, in your opinion, what is this relationship between data and identity?

Noah Sobe 36:27
Well, I think Nelli gave some great examples of how data and sort of interpolation of people in terms of and in relation to the data that’s collected about them and the way that’s used. I think she gives some excellent examples that explain sort of the importance that this has for identity. Maybe what I would say is one of the things that I think we’re seeing is not just datafication, but also the digitization of data. I think that’s a really important, and this is like a bridge into big data. I think we’re seeing the sort of electronic collection and assembly of different databases. So certainly, like in relation to a kind of assessment, I mean we know that those produce identity possibilities, and Nelli spun through that nicely. But I think there’s also some important considerations about sort of what happens when the normalizing forces that make up people start moving into this digital realm.

Will Brehm 37:39
To me, that brings up this big issue of privacy. What data is being used, whose data is being used. So, Ezekiel, where does privacy fit into this conversation about data, and I guess, education, more largely?

Ezekiel Dixon-Román 37:56
That’s a great question. I want to add a couple things to some of Nelli’s comments actually. I find the project fascinating, and it also connects questions around privacy as well. So, there’s fascinating work around biometrics, particularly around in immigration, and the use of biometrics particularly, not just for identification, but also for the surveilling of boundaries – geographic and international boundaries – that I see as fascinatingly in interaction with the kind of work that you’re doing, Nelli. And I would imagine you’re probably looking at this as well. For example, use of facial recognition, and even fingerprint and even retinal scans for ways of not just datafication, but for identification. Identifying and surveilling what bodies are within, what bodies are without, and even finding ways of excluding, right. This also has some interesting questions, actually, to privacy. And one, who has access to the forms of data that is being collected? When do we even know when data is being collected on us? When do we have control over when data is and is not being collected on us? And even when it seemingly is given as a choice, for example, when we open up our new iPhone, or Samsung, whatever it might be, we always sign off to some form of agreement and terms that include your giving over of data. And in fact, how many people actually read those terms is one question, a whole other question is even if one read it, who would understand it anyways? But then on top of that, there’s a whole third layer: even for those that might understand it, do we feel like we really have an option? So oftentimes, it’s either you accept this, or you don’t get to use the technology. And this is a legal issue that needs to be taken up on legal grounds, where companies should not have that much power to be able to say, “If you’re not giving us your data, you can’t use this technology.” I think the boundaries of private and public have been reconfigured in our contemporary moment and are continuously being reconfigured. So, what becomes understood and constituted as private for my son, who’s five years old, is going to be completely different for what it is for me or for anyone else from minor or any other generations. The other piece, too, is that also is contingent on cultural context as well. The only other thing I’ll add to this is the discourses around transparency, which obviously is sort of the opposite discourse around privacy, or in relation to privacy, I think is very limited and, in some ways, problematic. On the one hand, there is a move in Europe toward trying to make more transparency and policies in data governance, as well as last August, New York City also implemented a policy of legislation in order to create more transparency in data and algorithmic governance, but so what if you’re able to see? So, what if I have access to seeing that algorithm that’s even producing my data? One, what am I going to do about it? Two, do I even know what the algorithm is doing that’s producing that data? Three, does the data scientist that he can wrote the algorithm even know what that the algorithm is doing? There are truly black box algorithms where even when the algorithm is implemented, they don’t actually know – the data scientists that program it themselves don’t actually know what the algorithm is doing. And so, the very notion of movement of transparency, it’s in an important direction, but it’s limited. It does not get us really in an important directions of equity, fairness, and justice that we really are trying to move toward and are interested in … I’ll stop on that.

Will Brehm 42:36
Thanks. Actually, I want to shift the conversation slightly now to thinking about comparative education. And so, Noah, I’d like to ask you about, does datafication of governance and everything we’ve been talking about today, does that change how comparative education scholars conceptualize the idea of education?

Noah Sobe 42:59
I think it changes our product, our work. I just want to say one thing about privacy, which is privacy is a historical accomplishment. The fact that we’re interested in privacy means we’re attuned to a certain kind of subjectivity that conceives of the possibility of having a public and a private self that are different from one another, right. And I think that’s fascinatingly what’s in the mix right now. And I think Ezekiel’s illustration of this is always changing, this is always historically changing. And for different generations. And yes, we should be concerned about it. Absolutely. But we should also sort of pause because we’re at the state we are; we should pause and think about, “Well, why is it that we’re concerned about privacy in the first place?” Okay, to your question. And I just say that because I found Nelli’s and Ezekiel’s answers really helpful. Ezekiel ended by talking about algorithms that we don’t really know how they work, and I think that’s kind of the danger that faces the field of comparative education right now. So, I mentioned the digitization of data, right, and I think that’s sort of most captured in the notion of big data, right. And I think it’s important to approach big data as more than large data sets, okay. When we talk about big data, we’re talking about a set of analytic techniques that go beyond analytics. We’re talking about a set of practices. Big data analysis, typically, is about collecting large volumes, putting it all in a database, and then running correlations with no research questions. Like Ezekiel was saying, you’re not even being driven by a hypothesis, you’re just mining this new “oil” resource to see what kind of learning, what kind of ideas we can pull out of it. And I think that’s the largest phenomenon or shift that faces the field of comparative education that we’re going to need to grapple with like how do we orient towards big data analysis? What’s the place of it in the field? And then I think as this datafication proceeds, and school leaders start to develop data dashboards – and I think of schools as data platforms – I think it’s important for us in the field to really think carefully about how we orient towards that data we use and which of these ideas we pull in, which we resist, which we reconfigure and so forth.

Will Brehm 45:57
And Nelli, in your mind, what are some of the biggest impacts that data and datafication are having on or reconceptualizing in the field of comparative education?

Nelli Piattoeva 46:11
Well, I’ll just start by saying that I think the whole idea of, and kind of practices of research, also are changing, I would agree with Noah. And I would say that, for me, it’s interesting how perhaps, as researchers, we used to imagine us going in the field, generating our data, coming back to the office, analyzing, and so on. And now, because of datafication becoming now this intensified process that affects the whole society, we are data as well, as researchers, in the different processes professionally, this dashboard. So also, we can them when I go to see some websites of my Australian colleagues, immediately I see their publication rates, their citation indexes, and I’m amazed, because fortunately in Finland, where I’m permanently based, we don’t yet have those systems. So, we are data, and we are acted upon also as data, and I think this is a fascinating development in what processes … So, it also blurs to me the whole boundary between who is the researcher, who is the research, who is governed, and I think we need to start thinking about how this the blurriness of this whole situation and us also becoming data in different processes in different contexts, how that also has an effect on us and on our research and on our being, also on our identity as researchers. And for comparative education, I think many of the questions that we were asking today about cultural sensitivity, about context, I think they arose because we are comparative education researchers having a conversation. So maybe I would answer this question from the perspective, “What is it that we can do?” “How can our field contribute to the understanding of datafication?” And I think it’s precisely bringing in the ideas of context, culture, how these processes affect different people, different geopolitical locations, differently. Also, the ideas of what Ezekiel was bringing into the conversation: the coloniality, the coloniality of knowledge production that we have discussed in comparative education already for a long time. So how datafication is a kind of new phase of maintaining these unequal practices of knowledge production.

Will Brehm 49:01
I want to sort of bring our conversation to a close, I want to think a little bit about the ethical use of data and what is datafication mean for ethics, and particularly in the context of Reed Elsevier, and Pearson. These are big education companies collecting all sorts of data that students don’t even really know that they’re collecting. And they’re marketing it and creating economic value in ways that I think researchers are only beginning to sort of understand. So, Ezekiel, how do we as researchers deal with the ethical issues that datafication bring up?

Ezekiel Dixon-Román 49:41
Yes, I appreciate that question. So, I’m actually going to try to make some connections to where Nelli actually left off, particularly the notion of all of us being, even as researchers, data. So I want to begin by sort of introducing the Lucian notion of dividuation here because what, in fact, has been removed, particularly in the datafication, and using Noah’s language of the digitization of data, is actually more of the collecting of data and information that’s beyond even the individual, that goes down to literally the compartmentalizing of even the body itself. And it becomes even greater and even more of a biopolitics that’s at play, that is collecting information, not just on us, but about the various parts and composite aspects of our very being – everything down to not just biometrics, but even biological markers, or biomarkers, to even the kind of algorithmic forms of sensitivity analysis, even sentiment analysis, whether it’s in writing, or even based on the encoding of facial expressions, and moods. And all this data, all of it, and it’s all of these various forms, is being used to inform policy and practice. And so, we then have to raise the questions as we’ve been raising and engaging in this conversation of “What does this then mean for the social?”, “What does this then mean for the kind of ethical implications that kind of unfold?”, “What does this mean for the political and economic?”, right. And I can almost choose any one example … one of the areas I’m doing work now on is in predictive policing, as well as in learning analytics. And you see, although different contexts, you still see there are still similar issues that arise around the ways in which the algorithms and missing more broadly the sociotechnical systems that are being built around these algorithms, if you will, are being imbued with sociopolitical relations. And I’ll say a system of sociopolitical relations that are hierarchizing and differentiating bodies in ways that are rendering certain bodies as fully human, and other bodies as nonhuman or even partially human. Also, a head nod to Nelli’s reference, and explicit sort of mention to colonialism and any kind of actually postcolonial kind of interventions that are being deployed with algorithms. So then, connecting this also to comparative education as well. So, I’m not a comparative education scholar, I should say, and I’ve always wrestled with this notion of comparative education. In fact, even when thinking about comparative education, I tend to think about maybe wanting to take a transnational studies lens in order to even deconstruct the very boundaries that are often even thought of in comparative education, or be the essentialisms that often even constructed into comparative education, but there’s a way in which datafication of societies, I want to say plural, I would actually argue there is increasingly more and more of a movement that is moving toward not just a sharing across states, across the agencies, but increasingly, it will become the sharing across national boundaries, and in fact, we’re already seeing this. We see it in immigration policy: You go through certain borders, certain immigration, or customs, and even though you’re going through just the US, it’s not necessarily the case of the US only has US data on you, they also have other countries’ data on you as well. And so, the spilling over, if you will, the globalization of even datafication, also is something that we need to also take, think about what are the potential implications for socially and politically for us, as well as thinking about this in the context of comparative education and the ways in which datafication in society is producing and preconfiguring a form of biopolitics in controlling, managing and populations. Last thing I’ll say is with regard to just quite directly, the questions around power and inequality, whenever we have any kind of system that is dominated by any particular perspective, that is an architecture, that is informed by any perspective, and I’m going to pull from critical disability studies for the moment to even try to fully inform where I’m going. So, one of the things I really appreciate that critical disability studies really pushes is the notion that the architectural design of space, of even material buildings in space itself or place, is often informed by an ableist lens. Or even, I might say, an able bodied or even lens, right. And so, what that then does is what it doesn’t mean that those, who are not able bodied, are disabled, but in fact, it constructs, it really materially constructs those very bodies as disabled. And in fact, if we were to flip it and suggest and actually think about what would it mean, what would it look like, and what would it do, if a material place was informed by a hearing, or visually impaired or some form of impairment. What are the ways in which it would begin to construct other bodies as disabled, even though as I’ve been constituted as able bodied? I think the same lens needs to be placed on to the context of algorithmic governance and data science and education, and really thinking about what is in fact, not the intentions, but what is the doing? What are the performative effects of such sociotechnical systems on education and society? And with that, thank you.

Will Brehm 56:27
It’s sort of reminds me of what Haraway says where we are all cyborgs, right. That we are not outside of data at all. So, Noah, I want to actually ask you, what are the big issues in ethics, in data science and datafication that you are concerned with as a comparative education scholar?

Noah Sobe 56:48
Well, I’ll just be real quick. I’m concerned, and Ezekiel’s comment about kind of the move to societies … I’m concerned about the construction of these sort of global scopic systems, right, that are attempting to bring human activity in the domain of education, but in other domains, too, right. Kind of within a single screen of vision for the purposes of intervention, extraction of value, and so forth, right. I’m very concerned about that’s what Pearson’s involved with, that’s what Ministries of Education as they start to reconsider their school systems as portfolios, literally drawing from the kind of investment rhetoric or discourse. So, I’m very concerned about that. I think we, as researchers in the field of comparative education, we have an opportunity, maybe a responsibility, just to sort of emphasize that that’s not the only way to produce value, right, that there are other ways to produce valuable knowledge. Nelli laid out a nice kind of mandate for the field of comparative education: Here are things we do well, we produce – I’ll say valuable, okay – valuable knowledge about context, about culture, right. I mean these are things that have value: they have a different kind of value than the value Pearson is after, right. But as researchers, I think we can make a strong case that we’re doing something that’s important.

Will Brehm 58:24
And Nelli, what about you? How do ethics fit into the work that you are doing in terms of datafication of education?

Nelli Piattoeva 58:35
Well, I think in in relation to this, I was thinking about your question about what can we do, right, as researchers? What makes us different from all these huge producers of big data with our small datasets that become too old, very soon. And I think the kind of conversation that we’ve been having today about the political, the ideological constructedness of data, how the production of data has lots of different layers, lots of different assemblages, that contribute to what data is or how it is performed, and how it performs also reality into being I think, this is precisely the processes that we need to pay attention to try to understand, to the extent possible. Ezekiel was saying about the impossibility to understand algorithms that all the time learn from new and new pieces of data. So, this is a challenging task, but I think understanding how data is produced is one of the tasks for comparative education, but also for educational research in general.

Will Brehm 59:51
Well, with that our hour has come to a close, so Nellie Piattoeva, and Ezekiel Dixon-Román, and Noah Sobe, thank you so much for joining this webinar on the datafication of comparative education.

Noah Sobe 1:00:04
Thank you all; I appreciate it.

Ezekiel Dixon-Román 1:00:06
Thank you.

Nelli Piattoeva 1:0:06
Thank you. Thanks. See you.

Coming soon

Coming soon