Category Archives: Blog Posts By Me

For EDC students who may wander on to my blog. These are the posts I think would be most suited for others commenting. Of course, you are welcome to comment on my other posts if you want.

What I really like at the moment (June 2018)

Tweeting at politicians and mucking around. @JYDrummer and @dabjacksonyang

Open educational resources on acting – click here  This website is really well run and has many interesting clips. The blog posts are thought provoking and diverse. It is called the Actor’s Pad. It is so interesting that I am considering getting into drama as a sideline from drumming. Because barely anyone seems interested in my drumming and no one pays for music these days.

I am also enjoying yoga. physio exercises  and how quiet it is in Edinburgh now the undergraduates are gone (but before the festival kicks in).

I’ve also got my fingers crossed that the challenges I have issued on HITRECORD will attract the interest of some illustrators. I found this platform via Twitter and my nostalgia for Third Rock From The Sun. The layout is a little confusing and overwhelming at first but once you get past the homepage and read the instructions it’s pretty good. As I said, I found the site via the horrendously combative clamour of Twitter. In comparison HITRECORD seems like a relative oasis of calm. It is a bit social and network-y but mainly it is nice creative people looking to do interesting fun things. I hope there is enough traffic on the site for my ideas to be picked up and worked on. We’ll see.

I found an old podcast I did for my brother back when he was DJing more regularly. He was such a good DJ. I still like listening to these now. This podcast includes an old demo I did called Twist and Crawl where I made a lot of use of a Aphex Twin style VST algorithm.

I’ve also got a fair few new music videos up on youtube. I intend to get better at video editing over the summer. Get them a bit more slick for tossers who judge music on the visuals rather than the content. Which is fair enough, I am a Bowie fan. I know visuals are important.

By using my double-barrelled name I have successfully Search Engine Optimised myself. It is a relief to not be so closely associated with Stargate related slash fiction.

Things I am currently engaging with that are not Twitter and Youtube…

Books – specifically this guy’s but then also Calvin and Hobbes when I want something a bit lighter

Long Walks

Coffee

Squats

Cycling

My Old School

Linked in

My former drum teacher’s Academy

My band played a gig at Henry’s Cellar bar the other weekend which was well documented if not exactly well attended. Watch this space/blog.

Getting in touch with old friends after losing my phone a month or two back then being busy.

I am all over twitter and youtube like a bad rash…

Because I can’t keep my opinions to myself when people say or do idiotic things. Myself included. I also need to find a bit more work as a drummer. So please like and follow….

(drum roll)

@JYDrummer

wooooooooooooooo!

I’ve also got a youtube channel as many hip young drummers are doing the youtube these days to show off their mad chops. I have no mad chops. Only mildly depressing post-30s jowls.

BUT!

You may still want to look at them here!

More woooooooooooooooooooooooo!

I tried to engage with reddit but I was called an arsehole too much for me to find it useful. I will probably just go back to lurking there.

WOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO!

My mum thinks I’m special and my wife knows that I am not that much of an arsehole so that’s good enough for me.

Festival of Creative Learning 2018 Reflection

Event Brief:

The event lasted one hour with roughly 25 students in attendance over the course of the entire event. Of those attending at least 5 had some degree of musical experience. This was evident in the fact that they brought their own instruments to the event. At least 6 people had already attended a previous “analogue jam” event, with most of the experienced musicians being previous attendees.

The event was opened and closed by myself giving a short speech explaining how playing music and drawing together in a non-assessed environment gives a much more authentic experience of teamwork than in standard university teaching. I also addressed the issue of recording and privacy, emphasising that only snippets of the event would be recorded in order to ensure a low pressure environment for experimentation and improvisation. This seemed to go down well with the attendees and I invited them to think about whether with the advent of lecture capture, campus based lectures are becoming restrictive, high pressure environments.

I took it upon myself to start and close the jams by loading up the different apps and/or starting simple percussion rhythms before handing the instrument to a reluctant participant and encouraging them to interact. I really enjoyed the event due to my predilection for the socioconstructive model of education when it comes to music. I place a great deal of value on the complex negotiation that goes on between participants in a musical improvisation, much more so than the actual aesthetic value of the musical output itself. I am after tension and release, the passion and drama rather than empty displays of musical expertise.


Digital Outputs:

A brief clip on Instagram which captured an impromptu act of shadow puppetry between myself and two students. In the background you can see the screen output of the Bloom app which was being operated by a further 3 students off camera. The hand percussion and melodica are being played live by humans, whilst the synth sounds are created by Bloom’s algorithm depending on where the students touch the ipad screen.

At one point we switched the Ipad to a simple paint programme to allow people to draw using the touch screen whilst others made music with the available instruments. This was what was created at the end of 10 minutes or so jamming. It was really exciting to see this abstract collage come together on a screen at the front of the room whilst everyone played.


Things That Went Well:

 I was particularly pleased with the Roland VT-3 voice changer. As I suspected If you give people the chance to mask their voices in swathes of digital effects and noises they become a lot less self-conscious about singing in public. Additionally, the easy to use interface with a limited amount of slide adjusters meant It had a toy like appearance that encouraged collaborative play.

There was a high level of participation from attendees, particularly if I made the effort to hand an instrument to them personally and showed them a suitable, simple rhythm they could try. There were only a few people who were so reluctant they would refuse an instrument in such circumstances.

The Bloom app worked perfectly through the projector and speakers. The touch screen and visual aspect of the app was off particular appeal to “non-musicians”, to the point where I had to keep an eye on things so everyone could have a turn.


Things That Could Be Improved:

It was difficult to gather information on the attendees for this reflection e.g. what year they were in, what they were studying, what level of musical experience they had, etc. as I had to make sure the room set up and cleared in time. Due to the improvisatory nature of the event and how easy the apps were to use I think the event still ran successfully despite me not knowing much about the attendees. It would have made this reflection easier though. In future, I could perhaps consider a short paper feedback questionnaire during the last 10 minutes of the event. This would allow me to tailor the marketing and promotion of the events as well as give insight to the type of people that drawn to these events.

Due to personal issues leading up to the event, I was unable to write a speech that would explain to the participants exactly how participating in a musical improvisation would help them critique other types of group work within their studies. I still gave the speech but it was entirely improvised and consequently not as coherent as it could have been. At the time I felt this was appropriate as it reflected the improvised format of event, now that I come to write a reflection I wish I had documented it properly. The essence of my argument was that, due to the lack of formal assessment, my event represented the true spirit of group work, that the output of the event would only be as coherent or “good” as everyone there wanted it to be. This meant that the event would teach them the “soft skills” of teamwork and cooperation that other formal teaching assessments could only simulate artificially.

I perhaps need to give people a little more guidance if I want the improvisations to explore more nuanced textures and dynamics. Some participants got so drawn into the apps that they did not listen to others and basically did free-for-all squalling noise, all the time. This demonstrates that making the apps as easy to use as possible is not enough to ensure that everyone interacts together, no matter what their skill level. Developing listening skills, musicality and self-control takes a lot of time and, in my opinion, is best developed by instruments not being so well designed that anyone and everyone can play them first time.


New Digital Skills I Employed:

Twitter moments (which I used to juxtapose my event with a silly rap poem about Foucault by my favourite music journalist, Simon Reynolds. It struck me as being in keeping with the playful yet critically engaging atmosphere I wanted the event to have)

Ninjajamm (easily my favourite app for using myself, although difficult for others to pick up and play)

Bloom (Most relaxing and fun)

Trope (Not quite what I was after, bit too dark and rumbling for the speakers in the room)

What was interesting to note was that despite the low barriers to participation these apps gave people there was still a lot of reluctance for some “participants” to have a go. I can’t really identify with that mindset but then not joining in is still being part of the music. It’s just a more passive way of participating.

 

 

 

Waves hands frantically…

https://www.ed.ac.uk/news/2018/british-sign-language-promoted-by-new-plan

woooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooop!

I’d give it a thumbs up but bloody facebook have taken that away from me in my mind.

I have also been asked to stay out of the School of Politics and Social Science whilst I am “sick”. Which presumably means for the rest of my life given I was born with no hearing on my left side and am not going to be getting any better anytime soon. I never liked it in there anyway. They can come meet me elsewhere…the meadows, Moray House, Online, Wherever, Whatever, Innit.

Final Assessment Off Cuts

I don’t think this is going to make it’s way into the final assessment script which is a shame. BUT! Nothing is wasted when you have a blog so here we go:

“How do the methods used to generate formal academic knowledge effect it’s representation in digital network media?”

We’re going to look at two apps, Litlong and Curious Edinburgh. Now ostensibly they are representing formal academic knowledge in a very similar manner. They are both apps for mobile phones that add content markers to a GPS map of Edinburgh. However, if you look at them in detail we find that there are vast differences between the apps. I will argue that the cause of these dissimilarities is the different methods the app developers used to generate their formal academic knowledge. From this we can conclude that is important to avoid making instrumentalist conclusions about what methods are used for generating knowledge. In turn, this will reinforce why it is vital to take a critical approach technologies in Education which looks at the wider social practice of a technologies production.

How are they similar?

GENERATE Both are content based apps as they give access to large amounts of data with no assessment or learning task built in. Cherner et al.

GENERATE Both are manipulable – Goodwin and Highfield 2012 no set structure in the order, time or place you access the information. Counters perception of algorithmic/digital culture being committed to procedure.

GENERATE Both required inter-disciplinary approach to their production –

http://www.morayhouseschool.education/public/dice/speaker.php?ID=BenWilliamson&vRes=HD&status=16 interdiscplinary

http://www.morayhouseschool.education/public/dice/speaker.php?ID=BenWilliamson&vRes=HD&status=7 algorthmists relate to interdisciplinary approach of both apps

For a single academic, or even a single department, to produce these apps they would need to be well versed in computer science, data analysis, English Literature, History, Sociology and so on. Far too much for any one person to be an expert in all the requisite fields. This means that the apps had to be produced by an interdisciplinary approach.

The example of these apps provide a possible answer to a question posed by William, namely, should academics learn to code to transmit knowledge? These apps seems to suggest they should just head teams who do the coding for them.

 

What causes the differences between the apps?

I would argue that the differences between the two apps are caused primarily by the way formal academic knowledge is generated and represented. The developers generated different academic knowledge as they are from different disciplines. This means they would have different ideas about what the content of the apps should be. We can critique this using some of the same methods that Tarleton Gillespie uses to critique algorithms. Specifically, we need to ask how the developers addressed:

  1. Patterns of inclusion
  2. Evaluation of relevance
  3. Entanglement with practice

So LitLong had a very broad pattern of inclusion. As long as the text excerpt found by the algorithm could be mapped onto a location in Edinburgh it was included. Every excerpt that managed this was deemed to be relevant. The workshops with different groups showed the varying ways the app could be entangled with practice.

In contrast, Curious Edinburgh’s pattern of inclusion was far narrower. The locations mapped had to be of verifiable significance to the history of science. The evaluation of relevance was carried out by the academic team drawing upon their knowledge of the field of the history of science. These academics were also in a position of institutional power to ensure it the usage of the app entangled with their student’s practice by making it part of a formal assessment.

CONCLUSION –

What this analysis of the two apps shows is that you can take what on the surface seems like a similar approach and end up with very different ways of representing formal academic knowledge. For me the most significant differences between the two apps were how they were used (in workshops, formal assessments) and the levels of funding they had to produce the apps (with subsequent responsibilities the teams had to their benefactors). These differences were caused by the way the developers chose to generate and represent their knowledge. What is crucial for our understanding of how digital networked media can be used to formulate this kind of knowledge is to consider the social context in which they are made and used. This in turn reinforces the importance of the critical approach taken to Education and Digital Cultures throughout this cours

The Top Ed-Tech Trends (Aren’t ‘Tech’)

Below is a wonderful post which I think reflects the themes of the EDC course. EDC was structured to cover the trajectory of digital cultures from its roots in cybercultures to its present day iteration in algorithmic cultures. This talk also aims to look the possible implications of ed-tech by critically considering previous trends.

I could possibly write some song lyrics on the “Californian ideology” whilst referring to other classic California name dropping songs e.g. California Girls – Beach Boys, Hotel California – Eagles, California Love – Tupac, California Uber Alles – Dead Kennedys, California Dreamin’ – Mamas and Papas.

 

his talk was presented at Coventry University as part of my visiting fellowship at the Disruptive Media Learning Lab

Every year since 2010, I’ve undertaken a fairly massive project in which I’ve reviewed the previous twelve months’ education and technology news in order to write ten articles covering “the top ed-tech trends.” This is how I spend my November and December – researching and writing a series that usually tops out at about 75,000 words – which I didn’t realize until print copies were made for my visit here is about 240 pages.

Now, all those words and pages make this quite a different undertaking than most year-in-review stories, than much of the “happy new year” clickbait that tend to offer a short, bulleted list of half a dozen or so technologies that are new enough or cool enough to hype with headlines like “these are the six tools poised to revolutionize education.” To be honest, these sorts of articles are partly why I undertake this project – although each year, when I’m about 15,000 words in, I do ask myself “why am I doing this?!” (This talk will hopefully serve as an explanation for you and a nice reminder for me.)

Last year, I gave a lecture at Virginia Commonwealth University titled “The Best Way to Predict the Future Is to Issue a Press Release.” (The transcript is on my website.) It was one articulation of what’s a recurring theme in my work: we must be more critical about the stories we tell and we’re told about the future of education. Indeed, we need to look at histories of the future and ask why certain people have wanted the future to take a certain shape, why certain technologies (and their stories) have been so compelling.

To be clear then when when I write my “trends” series, it’s not meant to be predictive. Rather it’s a history itself – ideally one that’s useful for our thinking about the past, present, and future in the way in which the study of history always should be. It’s a look back at what’s happened over the course of each year, not simply – to counter that totally overused phrase from hockey player Wayne Gretsky’s dad – to “skate to where the puck is going,” but to examine where it has been. And more importantly, to ascertain where some folks – those who issue press releases, for example – want the puck to head.

So I am not here to tell you, based on my analysis of ed-tech “trends,” what new tools you should buy or what new tools you should incorporate into your teaching or what old tools you should discard. That’s not my role – I’m not an advocate or evangelist or salesperson for ed-tech.

I realize this makes some people angry – “she didn’t tell us what we should do!” some folks always seem to complain about my talks. “She didn’t deliver a fully fleshed-out 300-point plan to ‘fix education.’” “She didn’t say anything positive about technology, dammit.”

That’s not the point of my work. I’m not a consultant hired to talk you through the implementation of your next project. My work is not “market research” in the way that “market research” typically functions (or in the way “market research” hopes it functions). According to the press releases at least, ed-tech markets are always growing larger. The sales are always increasing. The tech is always amazing.

I want us to think more critically about all these claims, about the politics, not just the products (perhaps so the next time we’re faced with consultants or salespeople, we can do a better job challenging their claims or advice).

As you can see, much of what I write isn’t really about technologies at all, but rather about the ideologies that are deeply embedded with them. I write about technologies as practices – political practices, pedagogical practices – not simply tools, practices that tools might enable and that tools might foreclose.

Throughout the year, I follow the money, and I follow the press releases. I scrutinize the headlines. I listen to stories. I try to verify the (wild, wild) claims of marketers and salespeople and politicians. I look for the patterns in the promises that people make about what technologies will do for and to education. And it’s based on these patterns that I eventually select the ten “Top Ed-Tech Trends” for my year-end review.

They’re not “trends,” really. They’re themes. They’re categories. They’re narratives.

And admittedly, because of my methods, how I piece my research together, they’re narratives that are quite US-centric. I’d say even more specifically, they’re California- and Silicon Valley-centric.

I use “Silicon Valley” in my work as a shorthand to describe the contemporary high tech industry – its tech and just as importantly, its ideology. Sticklers about geography will readily point out that the Silicon Valley itself isn’t the most accurate descriptor for the locus of today’s booming tech sector. It ignores what happens in Cambridge, Massachusetts, for example: the site of Harvard and MIT. It ignores what happens in Seattle: the home of Amazon, Microsoft, and the Bill and Melinda Gates Foundation. (The influence of Bill Gates in education and education technology policy really cannot be overstated. Bill Gates is not part of Silicon Valley per se, but the anti-democratic bent of his philanthropic efforts – justified through claims about “genius,” through a substitution with charity (which is also tax relief) for justice – I would contend is absolutely part of the “Silicon Valley narrative.”)

Silicon Valley is itself just one part of Northern California, one part of the San Francisco Bay area – the Santa Clara Valley. Santa Clara Valley’s county seat and the locus of Silicon Valley (historically at least) is San Jose, not San Francisco or Oakland, where many startups are increasingly located today. Silicon Valley does include Mountain View, where Google is headquartered. It also includes Cupertino, where Apple is headquartered. It includes Palo Alto, home to Stanford University, founded in 1885 by railroad tycoon Leland Stanford.

The “silicon” in “Silicon Valley” refers to the silicon-based integrated circuits that were first developed and manufactured in the area. But I extend the phrase “Silicon Valley” to all of the high tech industry, not just the chip makers. And those chip makers aren’t all located in the area these days. Arguably the phrase “Silicon Valley” obscures the international scope of the operations of today’s tech industry – tax havens in Ireland, manufacturing in China, and so on.

But if the scope is international, the flavor is distinctly Californian. A belief in the re-invention of the self. A “dream factory.” A certain optimism for science as the penultimate solution to any of the world’s problems. A belief in technological utopia. A belief in the freedom of information technologies, in information technologies as freedom. An advocacy for libertarian politics – think Peter Thiel (a Stanford graduate) now advising Donald Trump. A faith in the individual and a distrust for institutions. A fierce embrace of the new. A disdain for the past.

California – the promised land, the end-of-the-road of the US’s westward (continental) expansion, the fulfillment of Manifest Destiny, colonization upon colonization, the gold rush, the invention of a palm-tree paradise. The California too of military bases and aeronautics and oil. California, the giant economy. The California that imagines itself – and hopes others imagine it – in Silicon Valley and Hollywood but not on the farms of the Central Valley. The California that ignores race and labor and water and war.

The California that once could boast the greatest public higher education system in the US – that is until Ronald Reagan became governor of the state in 1966 after campaigning on a vow to “clean up that mess in Berkeley” and promising during his first year in office that he’d make sure taxpayers in the state were no longer “subsidizing intellectual curiosity.” We can see in Reagan’s pledge the roots of ongoing efforts to defund public education, something that enabled for-profit schools to step in to meet the demand for college. We can see too in Reagan a redefinition of the purpose of higher ed – it’s not about “intellectual curiosity”; it’s about “jobs,” it’s about “skills.”

Despite thinking of themselves as liberal-learning, today’s tech companies re-inscribe much of this. “Everyone should learn to code,” as they like to tell us. “Higher education is a bubble,” as Peter Thiel has said. “Disrupt.” “Unbundle.” “It’s like Uber for education.” And so on.

“The Californian Ideology,” as Richard Barbrook and Andy Cameron described all this in their terrifically prescient essay from 1995, does not tend to make many lists of the “top ed-tech trends.” But the ideology permeates our digital technologies, whether we like it or not. And if and when we ignore it, I fear we misconstrue what’s going on with Silicon Valley’s products and press releases.

We’re more likely to overlook the role that venture capital plays, for example. 2015 was a record-setting year for investment in education technology, with some $4 billion flowing into the industry globally. But the total dollars fell sharply in 2016 – “only” $2.2 billion. The number of investments fell by 11%. (It’s a bit too early to tell what 2017 will bring.)

I repeatedly select “the business of ed-tech” as one of my “top ed-tech trends” because I think it’s crucial to questions about investors’ interest in education and education technology. What sorts of companies and what sorts of products do venture capitalists like, for example? What’s the appeal – profits, privatization? (Turns out, lately investors like testing companies, tutoring companies, “learn to code” companies, and private student loan providers.) Why has investment fallen off? (Turns out that “free” might not be the best business model for a for-profit company, particularly one that cannot rely on advertising the same way that other “free” products like Facebook and Google can. Turns out too that a lot of the education startups that have been promising “revolution” or hell even “improved outcomes” for the past few years have been selling snake oil. Turns out that the typical timeline that venture capitalists work with – about three to five years after making their investment, they expect a return in the form of an acquisition or a public offering and very, very few ed-tech companies go public. Turns out that Pearson, which once funded and acquired a lot of startups, isn’t in particularly good financial shape itself.)

Now, it’s so very typically American to come to the UK to talk about ed-tech and to insist “oh really, it’s all about the US – our values.” “It’s all about the state I live in” even – to invoke Pearson, a company founded in Yorkshire in 1844, the largest education company in the world, and still insist that the “Silicon Valley narrative” and the “California ideology” are the dominant forces shaping education technology. (I’m not thrilled about this either, mind you!)

In Distrusting Educational Technology, sociologist Neil Selwyn identifies three contemporary ideologies that are intertwined with today’s digital technologies – my reference to “Silicon Valley narratives” are meant to invoke these: libertarianism, neoliberalism, and “the ideology of the ‘new economy.’” Selwyn writes,

Most people, it would seem, are happy to assume that educational technologies are “neutral” tools that are essentially free from values and intent (or, at most, shaped by generally optimistic understandings and meanings associated with educational change and improvement). In this sense, it is difficult at first glance to see educational technology as entwined with any aspect of the dominant ideologies just described. Yet, as was noted earlier, one of the core characteristics of hegemony is the ability of dominant ideologies to permeate commonsensical understandings and meaning. Following this logic, then, the fact that educational technology appears to be driven by a set of values focused on the improvement of education does not preclude it also serving to support and legitimate wider dominant ideological interests. Indeed, if we take time to unpack the general orthodoxy of educational technology as a “positive” attempt to improve education, then a variety of different social groups and with different interests, values and agendas are apparent. …While concerned ostensibly with changing specific aspects of education, all of these different interests could be said to also endorse (or at least provide little opposition to) notions of libertarianism, neo-liberalism and new forms of capitalism. Thus educational technologies can still be said to be “ideologically freighted”, although this may not always be a primary intention of those involved in promoting their use.

I’d add another ideological impulse that Selwyn doesn’t mention here: that is, a fierce belief in technological solutionism (I’m building on Evgeny Morozov’s work here) – if students are struggling to graduate, or they’re not “engaged,” or they’re not scoring well on the PISA test, the solution is necessarily technological. More analytics. More data collection. More surveillance.

I would point to this “ideological freighted-ness” in almost all of the trends in which I’ve written about since 2010. You can see neoliberalism, for example, in efforts towards privatization and the outsourcing of core technological capacities to third party vendors. (This is part of the push for MOOCs, we must be honest.)

I’m not sure there’s any better expression of this “Silicon Valley narrative” or “California ideology” than in “personalization,” a word used to describe how Netflix suggests movies to us, how Amazon suggests products to do, how Google suggests search results to us, and how educational software suggests the next content module you should click on. Personalization, in all these manifestations, is a programmatic expression of individualism. The individual, as the Silicon Valley narrative insists, whose sovereignty is most important, whose freedom is squelched by the collective. Personalization – this belief that the world can be and should be algorithmically crafted to best suit each individual individually (provided, of course, that individual’s needs and desire coincide with the person who wrote the algorithm and with the platform that’s promising “personalization.”)

Personalization. Platforms. These aren’t simply technological innovations. They are political, social – shaping culture and politics and institutions and individuals in turn.

In 2012, I chose “the platforming of education” as one of the “top ed-tech trends.” I made that selection in part because several ed-tech companies indicated that year that this was what they hoped to become – the MOOC startups, for example, as well as Edmodo, a social network marketed to K–12 schools. And “platforming” was a story that technology companies were telling about their own goals too. To become a platform is to be “the next Facebook” or “the next Google” (and as such, to be a windfall for investors).

Platforms aim to centralize services and features and functionality so that you go nowhere else online. They aspire to be monopolies. Platforms enable and are enabled by APIs, by data collection and transference, by data analysis and data storage, by a marketplace of data (with users creating the data and users as the product). They’re silos, where all your actions can be tracked and monetized. In education, that’s the learning management system (the VLE) perhaps.

I wondered briefly last year in one of those posts on my 2016trends.hackeducation.com blog, where I ruminate on what really is “trending,” if we were seeing a failure in education platforms – or at least, a failure to fulfill some of the wild promises that investors and entrepreneurs were making back in 2012. A failure to “platform.” Despite raising some $87.5 million in venture capital, for example, Edmodo hadn’t even figured out a business model, let alone become a powerful platform. Similarly MOOC startups have now all seemed to pivot towards corporate technology training, but certainly all corporate training isn’t running through these companies. Neither Coursera nor Udacity nor edX have become corporate training platforms, although perhaps that’s what Microsoft hopes to become, as a result of its acquisition of the professional social network LinkedIn, which had previously acquired the online training company Lynda.com.

Platforms haven’t gone away, even if specifically education technology companies haven’t successfully platformed education – yet. Technology companies, on the other hand, seem well poised to do so – not just Microsoft, but Google and Apple, of course. And even Facebook has made an effort to this end, partnering with a chain of charter schools in the US, Summit Public Schools, in order to build a “personalized learning platform.” From the company’s website:

The platform comes with a comprehensive curriculum developed by teachers in classrooms. The base curriculum is aligned with the Common Core, and each course includes meaningful projects, playlists of content and assessments, all of which can be customized. Teachers can adapt or create new playlists and projects to meet their students’ needs.

“Playlists” – this seems to be one of the latest buzzwords connected to personalization. “Students build content knowledge by working at their own pace and take assessments on demand,” the Summit website says. But while students might be able to choose which order they tackle the “playlist,” there isn’t really open inquiry about what “songs” (if you will) they get to listen to.

AltSchool is another Silicon Valley company working on a “personalized learning platform.” It was founded in 2014 by Max Ventilla, a former Google executive. AltSchool has raised $133 million in venture funding from Zuckerberg Education Ventures, the Emerson Collective (the venture philanthropy firm founded by Steve Jobs’ widow Laurene Powell Jobs), Founders Fund (Peter Thiel’s investment firm), Andreessen Horowitz, and others.

The AltSchool classroom is one of total surveillance: cameras and microphones and sensors track students and teachers – their conversations, their body language, their facial expressions, their activities. The software – students are all issued computing devices – track the clicks. Everything is viewed as a transaction that can be monitored and analyzed and then re-engineered. Stirling University’s Ben Williamson has written fairly extensively about AltSchool, noting that the company describes itself as a “full stack” approach to education. From the AltSchool blog,

As opposed to the traditional approach of selling or licensing technology to established organizations, the full stack startup builds and manages a complete end-to-end product or service, thereby bypassing incumbents.

 

So why take a full stack approach to education?

 

“You want to own the total outcome,” says A16z General Partner and AltSchool investor, Lars Delgaard. “We are building the world’s biggest private school system. To make that experience the one we want – one that is more affordable, better, and revolutionary – you need to have full ownership.”

While the company initially started as with aspirations of launching a chain of private schools, like many education startups, it’s had to “pivot” – focusing less on opening schools (hiring teachers, recruiting students) and more on building and selling software (hiring engineers, hiring marketers). But it retains, I’d argue, this “full stack” approach. Rather than thinking about the platforming of education as just a matter of centralizing and controlling the software, the data, the analytics, we have this control spilling out into the material world – connected to sensors and cameras, but also shaping the way in which all the practices of school happen and – more frighteningly, I think – the shape our imagination of school might take.

John Herrman recently wrote in The New York Times that

Platforms are, in a sense, capitalism distilled to its essence. They are proudly experimental and maximally consequential, prone to creating externalities and especially disinclined to address or even acknowledge what happens beyond their rising walls. And accordingly, platforms are the underlying trend that ties together popular narratives about technology and the economy in general. Platforms provide the substructure for the “gig economy” and the “sharing economy”; they’re the economic engine of social media; they’re the architecture of the “attention economy” and the inspiration for claims about the “end of ownership.”

Platforms are not substitutes for community. They are not substitutes for collective political action. We should resist the platforming of education, I’d argue. We should resist because of the repercussions for labor – the labor of teaching, the labor of learning. We should resist because of the repercussions for institutions, for the law, for democracy.

And these are the things I try to point out when I select the “top ed-tech trends” – too many other people want us to simply marvel at their predictions and products. I want us to consider instead the ideologies, the implications.

from Hack Education http://ift.tt/2nwX9OP
via IFTTT

Learning Analytics – I’ve been asked for my opinions

UoE is gearing up to put LA systems in place and surveys are being sent out to students as a gesture towards consultation. I was happy to provide the patina of consent to this institutional shift as it gives me some material for this blog. In bold are excerpts from the survey and italics are my comments and questions. I would also like to juxtapose the survey with a twitter conversation on LA systems from previously in the course here and here.

 

In the forth-coming years, learning analytics will be increasingly prevalent in higher education.

< This is a reasonable statement in isolation but within the context of the survey, especially as the opening line, it makes it sound that the application of LA is fait accompli and uncontroversial.

 

The educational data is used to implement support services that are used to aid student learning such as the development of early alert systems for those who may be at-risk of failing a course or dropping out, personalised learning environments, and improving student feedback processes.

< Positioning LA as being in line with learner-centred discourse. Possibly as a way of preventing critical engagement with the implications of LA?

 

As students will be the main beneficiaries from learning analytics, it is important for their opinions and expectations are accommodated into the design and implementation of any developed services.

< A very definitive statement. Are we so sure that students will be the main beneficiaries? Is it possible that the main benefits will be managerial control and perceived administrative efficiency and accountability?

 

The university will explain all the learning analytics service processes as clearly as possible (e.g., how my educational data is collected, analysed, and used)

< The operative words being “as possible”. How clear will these explanations be? How clear can they be if the university chooses to use propriatory software with blackboxed algorithms? Where will these explanations be made available to students? Will it be obvious or hidden somewhere in the course handbooks (which as a course secretary I can tell you no student looks at, ever)

The university will ask for my consent to collect, use, and analyse any of my educational data (e.g., grades, attendance, and virtual learning environment accesses)

The university will ask for my consent before using any identifiable data about myself (e.g., ethnicity, age, and gender)

< Will consent be asked for once to cover the duration of the degree? Should it be asked for at several points once the students have a chance to live with the implications of their decision? Can I opt in to some data collection but not others? E.g. you can see my grades but not my VLE access?

The university will provide support (e.g., advice from personal tutors) as soon as possible if the analysis of my educational data suggests I may be having some difficulty or problem (e.g., I am underperforming or at-risk of failing)

< Failing is not as subjective as underperforming. If you get less than 40 on an undergraduate course you have failed by the criteria set out by the university. But underperforming is far trickier. This will vary from subject to subject e.g. if I took Maths I’d be overperforming to even get a pass, whilst I would expect to ace a History course. Will the system reflect this?

Underperforming is entirely contingent on circumstances. The best I can do at one point of time, in a particular environment might be completely different in another.

The university will ask for my consent before my educational data is outsourced for analysis by third party companies

< Will it spell out what these third parties intend to do with the data? Will the university profit from selling the data set to third parties? What does it intend to do with said profits? Can they guarantee that the data will be held securely by third parties and not sold on again?

The university will give me the right to opt-out of data collection and analysis even if the action reduces the opportunities to provide me with personal support

< This gives insight into how opt-out will be framed to students. The university will do its best to make it sound like a bad choice so it can preserve the integrity of its data collection.

The university will request further consent if my educational data is being used for a purpose different to what was originally stated

< Again, how will these explanations be framed? Can students opt out later on?

The learning analytics service will collect and present data that is accurate (i.e., free from inaccuracies such as incorrect grades)

< Interesting to see no “as possible” in this sentence. It WILL be accurate, no question. We wouldn’t want to undermine the idea that the analytical data will be infallible.

The learning analytics service will show how my learning progress compares to my learning goals/the course objectives

< And what if I have completely different learning goals from the course objectives? Under what kind of circumstances will the students make their learning goals? Will there be an element of public performance in these goals if they know they are going to be rated on them later on?

The feedback from the learning analytics service will be used to improve the educational experience in a module/course/programme (e.g., identifying problems in the feedback, assessments, and learning activities)

< No mention of whether the introduction of analytics may worsen or just significantly change the education experience in an amiguous way. It will apparently only improve things.

The teaching staff will have an obligation to act (i.e., support me) if the analytics show that I am at-risk of failing , underperforming, or if I could improve my learning

< Which staff? Act in what way? In what time frame? Will the teaching staff be measured on the outcome of their actions? Will their workload be balanced to reflect such expectations?

 

Tweetorial TAGS analysis

So happy that Clare got this working. Let’s have a look at what’s happened then:

First of all the numbers haven’t changed, Phillip Downey still has the most tweets and was linked to the most conversations. This is a quantifiable fact so it is not surprising that it shows up in both of the analytics systems.

Another question that reappears when I look at the TAGS sheet is how useful can analytics be if we don’t understand the reasoning behind the calculations. I can see Ben Williams in the centre of the mess fairly near me, Dirk and Nigel, but what does this proximity signify? Did we really mention him or tweet him that often? In reality I imagine he has a lot more to do with Jeremy Knox on account of them writing e-books together and using each others’ lectures in courses, but the data doesn’t show that.

One thing that did surprise me was that amongst the mess of nodes I could see both Coldplay and Ed Sheeran. I instantly equated this with physical presence in a conversation, “when the hell were we talking with Coldplay? What on earth about?”. This reflects my instinctive, embodied conception of Twitter rather than a data led view. By clicking on the node I actually found out they were just mentioned once by Dirk in passing but because he hashtagged them it showed up in the data. I’m still a little unsure why anyone would bother to hashtag things like that. Or in fact hashtag anything at all. Why do people choose to reveal themselves to algorithms? I mean it’s not just a case of writing in shorthand. You could abbreviate things without hashtagging them. Maybe they imagine they’ll search through the data at a later date, but does anyone ever do that?

I’m such a Twitter noob.

Anyway I do appreciate the bizarre modernist poetry of the courses top hashtags. Another way to abstract algorithms perhaps:

Ed Sheeran Dwarf fortress!

Quads, quads, quads!

Duckduckgogoduckduck

lastel! Cyborg emmental!

Rollerderby Time Machine!

Stilton Analytics! Brie! Poss! Poss!