< home



Trust me I’m a Doctor…


A talk I recently gave to the British Computing Society in Cheltenham was an inflection point for me. Having previously given a little time and thought to the ‘knowledge’ provided by large language models (LLMs) and AI agents during my PhD in 2017 (e.g. Alexa, Google Home Assistant et al) I was pushed to give my thoughts at the talk on ‘trust’. Question arose such as: Do I trust the knowledge of LLMs; Should we encourage our younger students to trust information provided by corporate institutions; Can we validate information from AI as trustworthy; What knowledge is lost or promoted and who do we trust to curate it; Is synthetic text trustworthy? Trust theory explains how we as human people form trust in various scenarios. It posits that human people need trust to form social structures and social interactions. Trust is the willingness of a ‘trustor’ to be vulnerable to the ‘trustee’.
Key facets of trust theory are the extent to which the trustor has competence, skills and knowledge (ABILITY). The trustees goodwill, care or concern for the trustor (BENEVOLENCE). The trustees honesty, consistency and reliability (INTEGRITY). A central premise is that human people continually learn over time (from infancy to adulthood) whether or not their basic needs are met by their primary care givers (Erikson, E.H., 1963. Childhood and society (Vol. 2). New York: Norton). This has been interpreted as being a tenet of forming ‘trust’. Trust provides a framework for humans to be amongst each other. Weigart and Lewis (1985) have explained in detail how trust is a property of collective units (rather than being of isolated individuals). For me this points to how a lack of trust (or weak level of trust) weakens social bonds and this can be micro/mesa/macro level: between you and me or a synthetic agent, between us as a society or across populations globally. Weigart and Lewis (ibid) posit that “trust is the mutual faithfulness on which all social relationships ultimately depend. Consequently, trust may be thought of as a functional prerequisite for the possibility of society in that the only alternatives to appropriate trust are chaos and paralysing fear” They go on to explain that trust is emotional, cognitive, and behavioural and trust is evident at the micro-level of ‘interpersonal trust’ within societies (e.g. members of a subculture) and the macro of ‘system trust’ (e.g. Government). ‘Lying’ (Bok, 1979) has been discussed as being counter-productive to trust. I am assuming for this article that lying would include dis- or mis-information (i.e. ‘alternate truth’, by which I mean people can provide their own truth, interpretation of facts/statistics or simply their own opinion)

Moving forward and connecting to my own research interest in technologies, trust I’ll classify into two broad categories: trust between people and trust in technology. Each requires different heuristics, but these are then the basis upon which we as humans form this ‘feeling’ of trust and are also able to follow quite similar cognitive and importantly physical pathways – this latter pathway would include voice or spoken interactions (where we ultimately give facts, data, information (knowledge) to others. I’ve briefly talked about human trust above. I’d like to move on to talk about trust in machines. I’ll be using the term synthetics from now on to refer to machines, robots, agents, digital technologies, AIs, LLMs. For ‘synthetics’ then what are the heuristics of trust? McKnight et al., (2011) captured the following ideas when discussing technologies: reliability; functionality; helpfulness. If therefore a synthetic is to be trustworthy it needs to perform or operate consistently and properly.
A synthetic should also have the capability to do what you or I need to get done. A synthetic should also be regarded by us as helpful. To operationalise this in everyday life. I interact with my Google Home Assistant often through voice. I do not regard it as trustworthy when I apply the synthetic heuristics. It is inconsistent and mostly does not play audio tracks I ask for or respond to the question I asked. It is limited in the range of features it has often asking me if I want a URL sent to my phone. It provides inaccurate information, it fails to respond and I work hard to get it to be assistive framing how I speak and the volume and the clarity. The Google Home Assistant (GHA), for me, lacks credibility. If I think of GHA using the human heuristics (McKnight et al, ibid) I feel that GHA is not trustworthy. However, when I consider ChatGPT et al., using the McKnight et al., heuristics is it a different outcome. I did not want to trust an LLM but they are always reliable, they function very well and are incredibly helpful. Do I trust the LLM given the reason I use it – yes. ChatGPT feels credible to me, I can’t help feeling this way. But we can also see that the reason for use or the motivation for social interaction frames the way in which we evaluate the synthetic. What happens when I apply the human heuristics to the synthetic LLM – does it remain trustworthy? To recap: Trust is the willingness of a ‘trustor’ to be vulnerable to the ‘trustee’. For humans trust invokes ability (1), benevolence (2) and integrity (3) (and competence?).
ChatGPT and I may add the many many other AI agents out there to date can be understood to be social actors that are now included in social structures particularly those involving knowledge and digital communications. Is there a trustor/trustee relationship between myself and something like ChatGPT? I believe there is. So, firstly, do I regard ChatGPT as having competence, skills and knowledge – for certain – it may not be perfect but given proper prompting by someone with a little literacy in prompting it has ability. Tick factor 1. Does it or can it have my care at its core and show concern for me (2)? This is perhaps where some bending of what this means is needed. If it understands that I am asking it for information that is essential for me for some role or other then I position it as being able to be benevolent. It can provide me with information I need to perhaps resolve a problem I have encountered - it is helpful. I believe it to be so. Tick tentatively 2 above. Is ChatGPT honest, consistent and reliable? It tries not to lie and I know that it doesn’t actively engage in lying (Bok, ibid). It may hallucinate but if prompting were perfect would it lie to me? We won’t go into depth on lying. Is it consistent in that it is there when I need it, performs the same outputs and can be considered reliable – yes of course. Tick 3 above. Do I want to know that I feel trust for a synthetic? Not really. I am uncomfortable with the idea of trusting a synthetic. Perhaps here we can deploy the uncanny valley principle but through the lens of it being about trust and knowledge. For instance to what extent, when and how does a synthetic that can think make us uncomfortable and appear creepy? I have arrived at my endpoint. Trust is not complex, it involves a scenario and process that probably includes goals to be reached. In this process trust can be formed as human-human or as human-synthetic. Each has a set of heuristics that can be applied to understand credibility. When considering credibility it becomes then possible to feel that a synthetic (voice) might move into ‘uncanny’ territory and through a lack of credibility and low trust become creepy or be perceived as untrustworthy. When you factor in cyber threats such as data poisoning or human hacking trust becomes more important because of the ways in which knowledge, information and data might be used as a threat vector to manipulate or influence thinking and behaviours that emerge from thinking about knowledge that is synthetically generated. We see a blurring of the boundary between truth and fiction and an interrelated existence or uncertainty between reality and the virtual. Look at the family portrait… then look again closely…

Ethnograms - see sound (and or smell)

I'm an academic because I work in a university and do research and publish. But, as a creative also, someone who dabbles with the visual, I'm increasingly fascinated by the intersection of humans and material culture. Whether it be a person looking at a piece of graffiti, or taking part in sport or using an 'app', the interactions between people and things provide plenty of opportunities for 'data' (see my research outputs for more). An area I have become increasingly interested in is where creative practice (let's call it art) and research merge. What I mean by this is that the boundaries between the two can be perceived as quite clear but I pose the following as a provocation... I walk around an urban center and as I do I can smell all manner of different odours and smells, some nice and some not so nice. In my mind I can know where certain places are because of their smell. For instance, the tyre garage which smells strongly of rubber but only when you are very close of it. The fast food shop, the smell of which permeates across a wide area. If I note down where all the various smells occur, register how strong they are and on which days of the week I can smell them, I am collecting data. If I then map this data producing a visual artefact perhaps this is both the data analysis and a piece of 'art'? The research 'output' is then a exhibition of said works of 'art' and not a journal publication? Where does art's boundary end and research's boundary begin?
heatmap
This avenue of thought (as above with smell mapping) leads me to consider the 'transmediation' of data - the change from one sense to another. As in the example of smell and the urban center, being able to see the city as a constellation of smells. This is my current obsession, trying to get data from one sense and transmediate it to another sense so that different perspectives can be gleaned.
How I got to this point began back in 2017 when I began to think about digital artificial intelligence voice assistants. Things like the 'Echo Dot' or 'Home Assistant', devices which make just ones and zeros to form sound that the human ear can recieve and body understand. This in itself is magical. On and off patterns produce a wave that a human can draw meaning from - computer and human interacting. I can ask Alexa a question and the machine can help me learn. I digress, my point is that data exists in one form and what can occur when it is changed into another form whilst retaining its integrity? In terms of transmediation, I was stumped when doing my fieldwork in classrooms. How do I get a 'picture' of what talk and questions were occuring in lessons so that I can really get a proper sense of whether students were being curious or just 'downloading' facts from the teacher (for memorisation and then recall in examinations at a later date). To get this picture of the classroom I needed to see talk. I went through all the various methods of portraying data visually but could find no single method that met my needs. In the end, out of frustration, I invented my own method called 'ethnograms'. (For a step by step guide see my published article here). So, by collecting and portraying data in an ethnogram I was able to see inter-relationships between people through their talk. What I revealed was startling and worrisome - oftentimes some students were silent for whole lessons and this was a pattern across the school day. Also, the teachers talked the most using short sentences and few words and their talk was performance oriented.
So, returning back to the blurring of the boundairies between research and art. Currently I am mapping graffiti tags over a large town. My aim is to develop what I have called a 'senseplace' to portray specific spots or patterns across a large geographic area where people feel saddened by 'vandalism' and where they report that places "look bad" or are "ruined by this graffiti". In this way, I can memorialise bad senseplaces and give an eagle eye view of a town and look for further patterns. For instance, perhaps some areas are more 'bad' than others? My plan to further blur the art/research boundary is to exhibit these visualisations of 'tag-maps' in the summer. I am creating a 4m x 4m white cube on which graffiti artists will tag and paint to decorate (sic) it and then my visualisations will hang over this for a month. I am using the exhibition as a data collection activity - to blur the boundary between research output and data collection.

Digital More Knowledgeable Others and Performance-oriented Talk

Dr R. Cook. October, 2023
Introduction
Novel research I carried out (in 2018 with students in school classrooms using voice technology devices with a digital voice assistant) established that voice technology had some positive educational use cases for students. Firstly, in terms of affordance, they found the devices engaging and were keen users. Secondly, they talked more with the voice assistant and with each other and this was about learning. Thirdly, they asked more questions some of which were evidence of epistemic activity. For example, 87 questions (not requests for music, jokes etc) were asked by students over a period of 90 minutes. A lack of student questions is a something many educators will be aware of. I posited that the voice assistant performed the role of digital more knowledgeable other but also considered that the voice assistant may however, have been simply a digital more informative other. After all these off-the-shelf assistants have a role to play in facilitating data production and harvesting and promoting consumerist behaviours rather than providing education, teaching and learning to support knowledge creation.
graffiti painting
In terms of less positive use cases, the devices themselves could be problematic and unreliable. For example, students were often not heard, misheard and misinterpreted or information relayed back to them was content from web pages read verbatim. Despite this, and other anomalies, students displayed resilience and persevered, continually asking and making requests seemingly unbothered by the voice assistant’s failings and shortcomings. What was gained by ease of use appeared to outweigh students’ perceived effort and time investments. It was said by students that asking the voice assistant was easier than searching through a book.

Classroom observations of work students were set by the teacher and digital data from their talk with the voice assistant (snippets of which were pushed to and stored in the Corporate cloud by the voice technology devices) showed that much of what was sought by students were ‘answers’ to worksheet questions, tasks and bookwork they had been set to complete by teachers. Surprisingly two other very interesting features emerged from the analysis of the data, namely ‘performance-oriented talk’ AND a potential for ‘epistemicide’. I will discuss these below.

Performance-oriented talk // Page 102, Section 4.2.1
Despite hypothesising that the voice assistant would develop students’ agency, promote curiosity and increase question asking (thus altering teaching and learning) the majority of talk between students and teachers that occurred in the classrooms (not the material and subject content being disseminated via didactic methods) centered around grades, levels, marks and exams and so on. The lexicon of language strongly suggested that there was an orientation of talk around performance and measurements of performance both in the present (as perhaps assessment of learning and as learning) and in the future (e.g. forthcoming mock exams, end of term tests and ‘summer exams’ and GCSEs, summative assessment and ‘high stakes’ testings). Interactions of talk between teachers and students were often short (few words, single sentences) and functional with question asking by students being either a confirmation or clarification type: how long is left, what do we do here, how many marks is this and so on. This is not to say that teachers knowingly or consciously, as part of a pedagogical approach, actively engaged in performance-oriented talk (although methods such as AfL, AaL and AoL might engender it) the overarching aim and goals of the school, frame ‘teaching and learning’ in performative ways so that this type of focus and resultant pedagogy is ultimately inescapable. Schools, teachers, students and parents ‘buy-in’ to objective, quantifiable measurements of ‘learning’ and measurement and judgement making about school effectiveness. Progress, the target or goal for teachers and trainee teachers, is evidenced by metrics and data becomes central to processes and procedures both administrative and educational. It could be argued that the ‘tail then begins to wag the dog’ as appetites for data generation, collation and processing increase and pedagogies adapt accordingly, often it seemed unconsciously and unknowingly. In this way educational policy (and perhaps economic imperatives) at structural level permeate schools at orgnisational level which in turn seep into practices at classroom level. It should therefore perhaps not be a surprise to hear talk in classrooms that has a necessary performance-orientation given what ‘education’ and schools are conceptualised as being set to achieve.

Epistemicide // Page 159, Section 5.1.2.9
As mentioned earlier, students liked using the voice assistant preferring it to say a book or subject text book. They also talked more with the voice assistant than they did with the teachers. However, there are concerns that need to be considered when information that is delivered directly to students by a voice assistant comes from a business (‘corporate’ knowledge) and which also bypasses the teacher (teacher out-of-the-loop). Herein then lies the potential problem if one considers the following thought experiment. As a corporation I have a canon of knowledge derived from my corporate aim (profitability) that I wish to disseminate. My knowledge, I consider, to be in need of being the canon that people learn and absorb. It just so happens that my ideology is based around profit making. So, by disseminating my knowledge, I can make more profit. What it means is, if I can assist consumers to know more about my products or services in a way that promotes consumerism then purchases will follow and so will profits. I therefore only need a way to share my knowledge that is both easy and accessible. Here is where a digital technology that is cheap and easy to use but can push and pull data (perhaps via IoT) can become the vehicle to achieve this. The kick-back is that data on behaviours and shopping patterns and resultant profiling of customers can feed back into my system to inform successful strategies and perhaps also become a commercial entity itself. If my method of dissemination is ubiquitous and I have reach into millions of homes and engagement from millions of consumers then my canon of knowledge may become the de facto standard. What I say is what is believed and other canons of knowledge are usurped or squeezed out.

So, for a something like the voice assistant in terms of the teacher being out-of-the-loop when information is relayed directly to students from a corporation where does this leave students? It means that students who have come to school for shared social learning experiences that aim to enable them to build their own knowledge through engagement with a trained pedagogue, in fact may only interact with a corporate agent. What does this say about what we value within the walls of a school and ultimately what does the word ‘learning’ mean in this context? Really, we are saying that methods that produce data are valid, and learning is progress measurements derived from data. Alongside this, data, that which schools now require is placed (saved or stored) into corporate hands and brands and products are built in and placed in classrooms and a corporate canon of knowledge (as per profit-based goals) is directly connected to young and emerging consumers. It is not too much of a stretch of the imagination to then suggest that tweaking or altering the canon of knowledge in favour of a product, point of view, political perspective or to subdue, block or remove another may be perfectly legal and legitimate at a corporate level. Disseminators of a canon of knowledge do not have to confirm or adhere to professional educational standards if used in education.

Summary
When I placed a voice assistant into classrooms with students what I hoped to explore became secondary to issues that surfaced around performance-oriented talk and epistemicide. Some of the issues I had to deal with were also moral, I was left feeling that despite the research being ethically approved, morally certain aspects had become questionable. I had hoped that voice technology would give students agency to engage with rich and varied knowledge sources representative of or from worldwide cultures and develop a more critical perspective about the world they inhabit and function in. Naively I started from the point of knowledge is power. I had hoped that alongside the teachers they worked with students would co-construct knowledge using instant on-demand information that would enrich and augment their educational experiences and make them informed consumers of knowledge to contest, contend and hypothesise: to be more curious about our world and societies. Instead I had to confront the notion that that corporate devices select, curate and disseminate knowledge they deem to be relevant and important with no accountability for their degree of openness and diversity. It is currently not possible to identify the source of knowledge that is used by voice technologies. I have to take responsibility for having placed devices with students (opting them in by proxy) who then unquestioningly worked with Corporate knowledge as ‘the truth’ and in doing so produced data (as a product) for Corporate to use free of charge for profit.

Magazine Writings

graffiti painting
QUALIA

Phenomenology is ‘the study of phenomena, things as they present themselves to, and are percieved in consciousness’ (Hockey, 2021). Husserl and Mealeau-Ponty have described a way of seeing or perceiving the world through a first-person descriptive (often via narrative accounts) of what one senses. This philosophical approach places the world, as we live it and construct it, as a ‘sensorium’. In this way, experiences can be thought of in terms of intentionality (towards an object, or phenomena) and meaning (derived from its virtues). For instance, a rider might think about her bike (intentionality), consider what the bike is (components, moving parts, etc) and speak of it as how she feels riding it, or emotions or memories generated (meaning).
Phenomenology has been defined by Smith (2013) as ‘a study of structures of consciousness as experienced from a first-person point of view’. Hockey (2021) has outlined how researchers might participate in sport and record other participants and the researchers sensory data using a method of data collection called participant observation. Here then, through a phenomenological approach, is a way to look at and interpret ‘adventure cycling’. In essence, when one participates in adventure cycling, what qualia [sensory data] emerges? What meanings, feelings, emotions are produced and what senses are evoked and what interpretations can be generated?
I lift from the saddle to stop the impact. I loosen my grip on the hoods, relax my elbows and try and have little contact with the bike. Any part of me connected to the bike is connected to the terrain. My speed is transferred into the rocks which send force back into the bike and these are expelled through me. The terrain is rattling every part of me. My head is being shaken, I can feel my neck tensing. Every part of me is being rattled. My vision is being disupted, the violent shaking of my head means I cannot focus, I just try to get glimpses of what is coming, and look as far forward as I can. Instinct draws my view to the line I need to take, the smoothest, the fastest, the least dangerous. Experience has taught me that where you look is where you go. Seeing a large rock means you inevitably go over it. When I reach the bottom, a mix of emotions. First, orientation. My eyes adjust. I can see clearly. Tears stream from my eyes, clearing dust also. The wind had blasted my eyes which were strained open to see whatever I could. My back, hands and feet hurt. The back is stiff, so stand to straighten it. Hands are hurting and hot. Palms are red and angry looking. Feet, especially toes, hurt. Big toenail, left foot. Hurting a lot. I have snagged it on my shoe, Perhaps it is bleeding again? I am back to normal. 50 miles of this so far. Another 30 miles left.
Tomorrow I know which parts of me will hurt and how much they will hurt. I know the balls of my palms will have little feeling except for a soft tingliing. These are now less sensitive from many years of cycling. This sport ‘hardens’ you in many ways. Sitting, shaken and battered for hour after hour stiffens, roughens and reduces sensations across the body parts. One or two days after a ride like this normal sensations return except for my palms. It seems that even with rest, they have less feeling. A sort of numb, dead area on each. With a finger I can feel more around the dead-spot, less in it. Putting this to one side, the trade off is worth it. What is gained mentally overshadows what might be lost physically. The risk of harm both during and after riding these events is surpassed by the intensity of sensory feedback gained whilst doing them.

[Sensory Data] “My field of perception is constantly filled with a play of colours, noises and fleeting tactile sensations which I cannot relate precisely to the context of my clearly perceived world, without ever confusing them with my daydreams” (Meleau-Ponty, 1962)

SEMIOTICS
Husserl's notion of primal impressions should not really be seen (as is sometimes done) as some kind of elemental building blocks or the contents of our perceptions or cognitions. Rather, one should think of the primal impression-retention-protention as that form of consciousness that presents itself as time - time as we live through it - as the living present before it has been appropriated by reflection. Primal impressional consciousness points to the corporeal and temporal nature of existence. At the level of primal consciousness there is not yet objectification of self and world. Lived experience is simply experience-as-we-live-through-it in our actions, relations and situations. Of course, our lived experiences can be highly reflective (such as in making decisions or theorizing) but from a Husserlian phenomenological point of view this reflective experience is still prereflective since we can retroactively (afterwards) subject it to phenomenological reflection.
Only through reflection can we appropriate aspects of lived experience but the interpretability of primal impressional life is already in some sense given by its own givenness. Post-ride, perhaps the day after, the dust on a bike, the empty packets from food, crumpled clothing stuffed into a bag. Encrusted bottles with the last drops of water, all these things tell the story of time that passed. In this way, something as ethereal as the light coating of dust is time, a representation, a symbol of time. Places and moments are evoked. Dust becomes a semiotic. Dust brothers. Shared time.

DUREE
Time. A finite resource. Time is percieved by modern humans as from a beginning to an end. This is a linear conception of time, or life, as a journey forward, from a start to an end. Beyond the end, the idea of another life, an after life. These notions crystalise thoughts as a linearity. An alternate perspective, one which rejects linearity, is the conception of time as a circle. The cycle or cirlce perspective lends itself to the constructs of infinity and perpetuation. One might argue that we are born of the earth and return to the earth. We are formed from raw materials of earth and become earth again. Similarities can be drawn with objects in nature. The tree. An acorn comes from the earth, grows into a sapling, then a tree. 200 years of growth, represented in one ring per year, then 300 years of ageing until the tree’s life cycle is ended. It falls. It lies on the ground and mycelium claims it back to the earth. During its life it has produced acorns, which have blown on the wind to land and have then germinated to produce new trees. The same might be thought of for flowers or animals or insects.
Born of the earth when an acorn pushed through the soil in early summer, a child has a syncronisity with the tree. Each year of the child’s life is represented by a new ring in the tree. At twenty rings the child has turned into an adult. After fifty rings the child becomes aged, considered old. Perhaps at seventy rings, the tree durée will continue but the old adult will not. Human life, like human time, is finite. The measure of human life in cycles of spring, summer, autumn and winter is a circle. Circles within a cycle. An ancient oak tree might see 500 summers and 500 winters, a human less than 100. It’s probable that a human may see and experience only 40 adult summers.
Forty of anything is a limited resource. Forty friendships. Forty days. Relative to human life on earth or the age of the rocks, trees and rivers, human life is temporary, finite and brief. When we concieve of time as a circle we can draw solace knowing that we are earth. We return to where we came from, we become of the earth. Earthlings. Both time and us are present in that which we live on. Like rings in the trees, we are parts and particles of the earth. Human beings, being humans. Aboriginal people exist according to ‘here-now’ time. Past and present are one and the same. An individual is the centre of ‘time-circles’ and events are placed in time according to their importance rather than organised or concieved chronologically. more important events are considered ‘closer’ in time.

EAST AS NORTH
North. Up. The top. Why do we use north as a way to orient ourselves to our place in the world. Consider how we desire south facing gardens, go south for the summer sun. When we consider what really shapes human behaviours we may look at that which we overlook. We are predisposed to orient towards the sun. Our buildings are lain to capture its’ warmth and light. Human drawn maps provide a perspective or world view that supports the notions of a top and thus a bottom. These constructs are understood more easily, through a process of socialization, in our actions and behaviours, as a hierarchical or stratified society. The top and bottom of society, nationally and globally. Northern hemisphere and southern hemisphere. The top of the world and the bottom of the world. What if there was no up or down? What if north as a construct had never been concieved? The sun rises from the ‘east’. Consider there is no east and no north for a moment. If you had to orient yourself with no map or constructs of the map, how would you gain your place in space and time? The constant is the sun. Shadows from the sun provide a mark on the earth that provides some sense of place. The movement of the sun through the sky provides some semblance of time. Together the sun and shadows provide a natural method of measurement and thus orietation. The sun provides some sense of order and control of human behaviour. We cannot see in the dark, but can only see in the light of the sun. it is how our eyes work. It follows that at night when light is absent and we cannot see, we should rest or sleep. When the light rises and we can see, it follows that we should wake and move. The sun provides heat, the night cold. It follows that we should shelter and stay warm at night and rise when the heat rises. The first glimpse of the sun, rather than the light from the sun, occurs not at ‘east’ but at the place that varies through the year, roughly where south east would be. Here then is our new ‘north’. Here is where we may now take our bearing of the ‘top’. However, rather than concieve of this as the top, perception as three dimensions, view where the sun is first viewed as where human action and behaviour, or lived time, begins. The cycle of human behaviour begins for a period bounded by the sun, its light and its warmth. Not clock time, but human time or natural time.Perhaps better thought of as simply, life. Presented is a perspective of a new north, an orientation to the sun. Viewed in this way, alternate and opposiing perspectives on time, space and place can begin to be considered.

'Shadow-work', 2022, November 28th.

The underlying enabling feature of technology is the Internet - without which data remains siloed on a device until it is ‘connected’ and data collected from it. The internet, to which technologies, or things, are connected in a growable network, facilitate data-harvesting from individuals and groups to which ‘analytics’ give interpretations and insights about people. Hence why the internet (how data is shared) and technologies (how data is made) are so important - both are integral to how some people understand how some other people think and behave. Data is a useful way to know things that people won’t or can’t tell.
I completed a ‘knowledge exchange project’ (KEP) involving both an Age charity and Charitable Trust. I explored ‘digital exclusion’ using interviews with age-specific groups throughout a county, a scoping exercise and production of a heatmap of digital infrastructure. Several interesting notions emerged which require more thought and unpacking because they were not within the scope of the KEP. For instance, in the data analysis it emerged that there were people who resisted or rejected technology. The vignette (data) below is an example of how this can be seen.
graffiti painting
Vignette 1:
I watch Malcolm at work, logging in, getting into his staff email and so on. He’s asked me for “help with something on the computer”. Looking for information to send someone he awkwardly moves the mouse cursor over the scrollbar and “moves the page”. He moves the cursor back to the search bar, types in a search word with one finger (on one hand), looking at the keyboard for each letter and then to the screen. Malcolm then moves the cursor to the search icon and clicks on it. Back to the scroll bar to ‘move the page’ again to see the “page”. It’s a laborious process that demands his full attention and occupies his mind - I see how slow and hard it is for him to do simple things others do effortlessly with tacit knowledge. For Malcolm using technology involves a lot of ‘shadow work’, additional labour and effort that is cognitively costly. Technology gets in the way he tells me, is another set of demands on him and he has no desire or motivation to become digitally literate he tells me.

The above vignette illustrates an intersection between technology and people. Many people manage this intersection because they draw from help, assistance and support from younger, more digitally literate relatives, colleagues or friends. They resist but the technology demands their attention because it needs to be fed data and interactions with technology are orchestrated by organisations to facilitate engagement - Malcolm has to login to work and search to find information as part of his work. However, as can be deduced from Malcolm’s work, the smallest digital-breadcrumb a login or search, has some value to someone. So, given that technology demands us to attend to it, what does that mean for people who resist or reject it? Who wants the data and why? Also, can someone reject technology? If the answers to these questions might be no, is the technology ethical or moral?

Using these questions as a way to problematize technology what does this mean for technology in education. Young people in schools are not adults, are seen as children and privacy rules differ as do notions of consent and assent. Some legislations identifies 13 as a boundary for access to certain media and often these media are part of digital technology systems in schools. Can students reject technology - for example, not login to access a homework or assessment?
Or, for example, not submit an assignment electronically or look at messages from the teachers without logging in? What data does the school need and what data should the school have. For example, data that speaks of socio-economic status or which speaks of an impairment or past misdemeanour. Perhaps data about gender or nationality or race? There are ethical and moral dimensions therefore to children's engagement with technology in schools.

This ethical and moral dimensions is a theme in the literature. More recent education technology studies signpost a moral dimension to the potential and possibilities within education associated with an ongoing ‘datafication’ of education, which has been explored by Perottwa and Selwyn (2020), Selwyn (2021a; 2021b) and Pangrazio and Selwyn (2021). In light of this moral dimension and datafication, Krutka, Smits and Willhelm (2021) proposed and applied a ‘technoethical’ framework used to critique Google Classroom, a platform that can be used to form part of a school’s education technology provision for staff and students.

Research has explored technology and described how, as technologies disappear (Streitz and Nixon, 2005; Plowman, 2019). More recently, it seems that individual agency and power have disappeared with people becoming less user and more ‘the used’. This user/used metamorphosis occurs as physical interactions (a push) such as clicks, taps and swipes become redundant or unnecessary and ‘shadow technologies’ such as non-tangible or invisible sensors, machine learning and analytics orchestrate and produce from people what is required (a pull), namely data. For example, whereas you may have needed to login to separate organisational platforms, you login once as a ‘single sign-on’ and all platforms are automatically available without login. What occurs through ‘shadow-tech’ and ‘pulls’ then is as Plowman (2019) has discussed, a disappearing occurs and alongside this as Krutka, Smitts and Willhelm (2021) have warned about the impact on democratic technology, pedagogy, power and knowledge are affected. Simply put, what is occurring, what is happening and what data is being produced and shared is no longer overt and obvious. For instance, when you are asked to login in and enter your details you consciously know you are accessing a platform. When you bounce from platform to platform via single sign-on you may not even notice you are logged in and have an account and account profile. What you click on, what you search for, where you are when you access it, and how you spend your time can be processed. Some organisational platforms now push a ‘digest’ back to the user telling them how they work, when they work with suggestions on how they may work better. This data may not be fully understood as to where it came from or how it was collected and importantly - who owns it. A hyper-emphasis is placed upon data, its production, collection and processing as Selwyn (2021a; 2021b) has warned. However, little is known regarding the moral dimension of deployment in school settings and what the perceptions and thoughts might be of the used.

Research Designs

I lecture/teach on several Masters Level 'Methods' modules across 3 Universities in the UK. I am involved with students at Undergrad through to PhD and common questions I get asked are...
What is a research aim?
What is a research objective?
What is a research question

I will answer these in short and this might be a place for you to begin to think about these before your supervisions with a supervisor.
Research aim - the broad high level view of your research and what ultimately you want to achieve at the end. A statement that gives the big picture of your research's purpose/intent
Research objectives - these are statements that are specific and have a clear intention or outcome connected to parts of the aim. They are the aim broken down.
Research questions - these are questions based on the objectives that are answerable, that your research will provide knowledge (to some extent) about. You can answer the RQs in your discussion chapter.

A quick and dirty (far from perfect) example to separate them out (hopefully!).
The aim of my research is explore students' perceptions of anxiety in group work and the role that technology might play.

RObj1: Explore anxiety in group work
RQ1: What do students self-report as anxiety

RObj2: Provide examples of a range of group work
RQ2: What features of group work cause anxiety

RObj3: List the technology used in group work
RQ3: What technologies are useful for group work/anxiety reduction

It becomes quite clear that given the RObjs and RQs (and overall aim) there is some clues or assumptions we can make methodologically speaking. Some options open up to us and others are perhaps closed off.
For example, we might discount a survey in favour of interviews as we are seeking 'self-reported' data. We might turn to phenomenology to interpret individual's perspectives about the things they interact with.

Literature Reviews
So what are you supposed to do in a 'critical review of the literature'?. Firstly, what you should avoid. Listing and summarising paper after paper telling the reader what the abstract already tells you.
Secondly, avoid a long prose absent of a logical structure which is poorly (or not) organised.
Finally, fall into the trap of criticising features of the research design. E.g. not much info about ethical approval. Or, 'small' sample sizes. These are common pitfalls for students. Taking small sample size as an example, the authors may be seeking thickness or depth of description so a small sample is fine.

The literature review has to have a connection to your aim and objectives. Literature selected should be therefore related.
Here are key things I look for when marking literature reviews...
Scope
Briefly tell the readers what literature you are reviewing to demonstrate a connection with your research aim and research objectives. Otherwise you are not defining your area and thus open yourself to a reader asking you why you did not consider the paper by leading author X. They may also ask you why you did not look at literature in the area of Y.
Funnelling
Start broad by laying out the knowledge in the wider field. Example, your aim is to explore student stress during the run up to GCSE exams in the UK education system. You don't need to talk about Education as a whole but lay out knowledge from secondary, just in eurpoean school system perhaps. Then focus in to a section on UK then Finland then France and Germany. You may then drill down into literature more closely related to your aim (perhaps UK GCSE examinations and student stress).
Focus
This is where you take your aim (GCSE stress) and dig deep into the literature drawing out key arguments, perpsectives and draw from highly cited or influential and key papers. You can demonstrate criticality and present your own academic voice.
Gap
Here, the final part of your lit review is where you begin to summarise what you have found out about knowledge in your area and present an area where further knowledge is need. It might be from a methodolgoical perspective, it might be to extend current thinking about a concept or it might be to apply a new method to gain further or new data. It might also be that there is a methodological gap.

Becoming Analytic
This involves connecting the data you have collected (and analysed) to theory and concepts. For example, if you have data from interviews about student anxiety and leaving school for University or employment you might draw from Markus' 'possible selves' theory or concepts of 'what I want to be'. You might draw from Jenkins theories on identity. What your aim should be is to begin to conceptualise YOUR data using others theories or concepts.
Too often I see the presentation of data that has seemingly been 'analysed' but this is where it ends. The analysis has been limited and there is little connection with theory or concepts.

Planning to Write
In literature reviews, students might form some methodological assumptions which need to be outlined in a Methodology chapter. Here is where the struggle for some students starts.
What is a methodology?
What are the connections between lit review and Methodology?
What is IN a Methdology chapter?
What is epistemology and ontology?

These questions are usually emerge during supervisions with students and require some dicussions. I'll answer each of these in turn below.
Question 1
Methodology can broadly be thought of as the process (or the how) of your research. It would lay out key elements that gives a reader the overview of your study in practice. To provide an example, I am specifying a qualitative research study, that uses an interpretivist paradigm, and will subjectively determine (to build towards theory inductively) thoughts and feelings about workplace stress. I will conduct an auto-ethnography that will collect data and via ethnomethodology, thematically analyse the data collected over a year long period of fieldwork in the workplace. Ethical approval will gained and gatekeepers of the workplace will be informed and consent will be gained. All the data stored will be kept under GDPR guidelines. It is a low risk study with minimum potential for harm. Inside this example you can spot METHODS.. instruments to collect the data.
A common mistake Masters students make is to overly focus on methods. e.g. Methodology chapters often start as "I will use interviews..." Something also often overlooked is philosophical assumptions and positionality (which is connected to reflexivity) more on this later.
Question 2
Your literature review can be used to foreground your methodology chapter by highlighting where further knowledge is needed. What I mean is that during your review you may outline that studies have been mostly quantitative. There is a need for a qualitative perspective therefore. Prior research has looked at one stakeholders views and through survey presented results of this. Perhaps now interviews or participant observations might surface new knowledge. What becomes evident then is that your literature review (along with your research aim and research questions) will signpost the direction you take in the form of methodological assumptions.
Question 3
See question 1! And refer back to research aim, objectives and questions at the top.
Question 4...
Common across all the courses, modules and Universities I work for are the struggle that students sometimes have with philosophical assumptions.
Epistemology, one’s theory of knowledge. One can have epistemologies about various things.
In academia, we generally take the stance that there is Interpretivism or Positivism.
Epistemology for Education as a system, or a philosophical idea. One of its main tenets would be ‘Education is a good thing’ simply put, productive for humans.
These are 'beliefs' (what we think is right or the truth) and each person has an individual perspective. Some may believe that knowledge is measurable and one truth can be accounted for. Another may believe that all reality is once-removed and thus an interpretation of multiple points of view or there are many realities (and these are socially constructed).
Take for example a teacher's belief about sanctions that can be given for 'bad' behaviour - what bad behaviour IS is a personal perspective in essence. (bad behaviour is also on ontological claim... as is sanctions)

Ontology, can be thought of as claims or theories about what is possible in the world, in existence. NB: As above, we saw that each epistemology contains numerous ontological claims (bad behaviour, sanctions).
Thus Religion: one claim would be there are certain objects in the world which are ‘relics’ = bones of the Saints etc.
Ontology, the nature of things or what is in existence, is a philosophical notion.
Take for example the DfE, Vygotsky or Mickey Mouse. The DfE is an organisation, part of a Government, Vygotsky is a person and Mickey Mouse is a cartoon character. Mickey Mouse 'exists' in a different way to Vygotsky who existed in the past and neither exist in the same way as an organisation like the DfE. Our understandings of existence, the nature of things, are then somewhat different given what they are (or were) and what they mean or meant to us. In school an ontological claim would be ‘time-tables’ exist. Or ‘learning’ is ongoing. You might theorise that timetables facilitate order and control that facilitates learning as episodic.

Conclusion Chapters
These are often asked about. Often students ask how can I write up to 2000 words? My take on this chapter is as follows.
You should not begin to talk about 'new' ideas, introduce new arguments nor enter into detailed discussion. Think abut this chapter as firstly being the summing up of each chapter. For example, you might structure your conclusion chapter around your dissertation / thesis structure: introduction, literature, methdology, findings, discussion.
So, you might begin by writing what your overall research aim was and the objectives you had. You might then talk about the literature and the gap you identified. You may write about your methodological decisions (perhaps even talk about limitations you have indentified).
Then you can state your main findings and key contribution(s) to knowledge. You can then recap on the key thrust of your discussion - what were you arguing for and mention the theoretical framework or theory you drew from. You may finish with providing your thoughts on to what extent you answered each of the research questions and then hint at where perhaps further research might be needed.

If you think this page has been useful and your module lecturers are open to an external speaker (I am for my modules as other perspectives are always valued) then I am happy to guest speak. Email me here

What is Masters level writing?
It requires you to demonstrate 3 things: criticality, breadth of knowledge and depth of knowledge.
Along with this you are supposed to learn the 'craft' skills of designing and conducting a piece of empirical research (something that collects primary data). If given the chance to do desktop research, personally, I would advise against, it's more fun to do data collection.
Criticality in short - prsenting and considering arguments from opposing perpsectives and providing your argument / opinion on these.
Breadth - a wide range of academic sources consulted with demonstration of synthesis of arguments, debates, themes and categories.
Depth - a high level of knowledge about a very specific topic or area that critically considers knowledge in this area.

Common errors
"it is clear that all teachers talk too much" - this is a claim and this type of claim makes me wince. It is extreme and definitive. So really what you are saying is that every teacher in the world talks too much. Well can you empirically say every teacher? What is too much? Isn't it context specific? Every claim really should have some sort of empirical basis and be contextual.
If the research studied 30 teachers and the sample of 6 students reported that teachers talked for 99% of the Maths lesson then you might still say...
In the sample of 30 teachers in this school in the Maths lesson observed in this instance SOME students asked reported very high levels of talk.

Citing works
It's simple really...
Jones (2020) said this
It is assumed this is true (Jones, 2020)
Jones (2020, p.1) said "Gosh, it's true".
Authors usually found it to be true (Jones, 2020; Smith, 2021, Wang, 2022).