 So, I would like to welcome everybody, I'm really delighted to see so many people here. So I'm delighted to announce the NCM annual lecture for this year. It's a very prestigious event, we've had it in the past before COVID, but obviously COVID happened and we were not able to do sort of face-to-face events and last year we've started to do some of these face-to-face activities again and so we have been able to run this sort of annual lecture this year again and we are planning another one next year as well. And I'm really pleased to see so many people here, both physically here in the room at the Royal Society, it's a fantastic venue that we've been able to to get and obviously to everyone that is online at home as well or in other locations. And I'm delighted to welcome today's key speaker, Liz, Professor Elizabeth Stoko from the LSE here in London and she has researched for many years conversation analysis and we will hear very shortly her talk about this sort of very exciting research method area and I'm really delighted to hear sort of new developments in this field. Obviously communication is extremely important and also I would like to introduce John Sutton, the editor of the psychologist who will be today's discussant. So I'm really pleased that this has worked out and okay so that was basically my welcome. I just want to take the opportunity to very briefly say a few words about NCM. What are we doing? You know quite a few of them of you will be familiar with NCM activities but maybe for some it's less prominent. So basically we are providing high quality training in cutting edge research methods. Obviously we have some core intermediate research methods training as well but the focus is sort of advanced cutting edge research methods and we really have a wide range of work streams ranging from really the exploration horizon scanning activities to informing a comprehensive training program and eventually obviously creating impact. Obviously that leads us to having to evidence impact as well and that's a very difficult task from training activities. Basically we are a consortium of three core universities that is the University of Southampton, Manchester and also Edinburgh and we are joined by nine centre partners across the UK and there are different expertise, different regional areas that are being covered here both from the academic sector but also from non-academic sectors as well. And basically NCM is a distributed model, it's based on a partnership. So a lot of these activities that we are doing are based on engagement, are based on collaboration, are based on activities together with external stakeholders. And just to mention the first phase of NCM actually started in 2004 a very long time ago and we've had various different phases of funding and lots of changes throughout. We're actually in the fourth phase of funding and throughout the remit and the organisation of NCM has changed quite significantly and in fact next year we have the 20 year celebration or anniversary so to speak. So we will be organising a lot of activities next year as part of that so watch this phase effectively and if you have any ideas or want to contribute to that obviously we are always very welcome to hear about that. I've just put the link to our website here. Just feel free to explore the activities that we have on there. So I just want to emphasise why, what is it that we are doing and why are we doing it. I just want to really emphasise the sort of importance of research methods training. It really is about the application and use of high quality and appropriate research methods that's really essential in basically all ways of research and life and it's really important to improve evidence-based policy decision-making. It's really underlying a lot of activities I'm really passionate about that I teach students about this using appropriate methods and coming to the right sort of conclusion and being critical about the methods you're using and how to interpret that and so on. So the aim is really to train researchers to a high standard using appropriate often innovative methods. It doesn't have to be always innovative but NCM is about innovation in some areas. A quite a wide range of backgrounds, so research methods areas not just quant and qual but digital mixed methods, visual methods, a whole range of activities across all career stages. So we're not just focusing on junior or PhD student junior career researchers but actually across the stages. So we have for example innovation for that are more designed for maybe more senior members of staff across sectors. It's not just academics, it's across a whole range of areas, businesses, industries and so on across all disciplines and also not just within the social sciences but actually going beyond as well. And obviously we have to bear in mind it's a fast changing environment and we have to adapt to some of these developments. And the key thing is sort of effectively to serve the research community and the common good and that's sort of the main driver that I really quite like to emphasize. And I'm proud to say there's also very experienced, committed, passionate NCM team behind all of this. Also I would like to take the opportunity to really thank the team here today that has made that happening, had made this happen today and also we've had the advisory board meeting this afternoon. There are lots of things that you know have to happen and be organized. So I'm really pleased to have that support there. Just to sort of emphasize, obviously it's a changing landscape. There are many changes. I mean there are lots of, you know, living in a time of uncertainty not just because of the pandemic, lots of other activities or things that are happening as well. There are clearly technological changes that are happening at a very fast pace. There are societal changes changing in the data landscape. We hear a lot about the digital footprints or big data and so on. And obviously all of these activities, new tools, research methods are constantly being developed or maybe they exist already but we are not always aware of them. So being aware of them and being able to use them in appropriate manners is extremely important. Obviously also the funding landscape is changing continuously. So all of these things really need, really emphasize that there is a need for social scientists to be basically equipped for these sort of changes and times of uncertainty to lead basically cross-science, to lead the sort of cross-science agenda. It's really important for people to work collaboratively across disciplines, across sectors. One or two people cannot achieve or cannot address the sort of challenges we are facing and different disciplines and people with very different skills have to work together. And that goes beyond the academic sector and obviously crosses across government, business, industry and so on. So it's really important to train social scientists to a high standard using appropriate methods, often innovative methods and also to have and to offer flexible delivery formats. So that's really of increasing importance, not just because of the pandemic. We can see that, for example, today we have a hybrid event. These types of things are now happening, of course. Just very briefly to mention a number of sort of highlights of NCM successes. We have a very right range of training and capacity building activity. So it's not just about courses or running a few events, but we're having a whole range of training formats that we are advocating. And it's really sort of thinking about the sort of fluid transition between research and training effectively. So we are trying to be creative, innovative, advanced and also ambitious in terms of our delivery. So we had more than three and a half thousand course participants over the last three years and more than four thousand event participants and more than a million website users. And we have a number of online learning resources on that that are really having a very high uptake. It's across all, yes, I see disciplines across the career life course and so on. And we have, for example, started to put together sort of impact case studies. And last year we collected about 14 impact case studies, which this year and next year we will continue to update and add to. So we are obviously always extremely keen to hear from you if you had experience with NCM, maybe you've attended a course or an event, what has come out of it, how have you used these methods, have you maybe changed policy in government or something like that. Please email us, we are really interested in hearing the sort of impacts and changes we have been able to put into place. And also we're having really positive feedback from course participants, from people that are using our online learning resources. So we're collecting obviously all of this type of data. There are a number of quotes there where people really sort of feedback that the structure of a course or the delivery of a course was of high quality. There are just sort of some numbers here that may be of interest. And we had about last year, over the last three years about 20% of our participants were from actually outside the social sciences, for example from health and so on. So we are reaching out to other disciplines as well. Just wanted to emphasize, obviously we have the portal, the website, which provides a rich online database and there are various different infrastructures and services that people can use. For example, if they want to themselves run training events or want to find out about training events, all of this is on our website. And we, as I said, we have a comprehensive training program so you can explore what there is an offer. We have a wide range of activities like spring and autumn schools, also large scale events, for example, MethodsCon, Vifran or InnovationFora. There are various collaborative events or events coming out from our key pedagogy work stream, like training the trainers, for example. And I want to emphasize the Research Methods Festival that is coming up on the 7th to the 9th of November this year. So you can have a look on our website and there are lots of activities there. Last time we had more than a thousand participants on there. And we've created a number of networks where people come together, discuss, meet regularly or community of practice or methodological special interest groups that are focusing on particular aspects like, for example, pedagogy or video interviewing or DATN, which is the data resources training network or survey data collection method, so the whole range of activities. So if you want to create a network yourself and want to generate some activities, then we are always open to try and facilitate that and help people to do that. Okay, so now I would like to introduce our key speaker. So I'm really pleased that Professor Elizabeth Stocco from the LSE has been able to tonight give our annual lecture. She has been at LSE since January and previously she was for, I believe, 20 years or so at Loveborough University and had many senior roles there already. She has been researching for many years conversation analysis and basically it sort of focuses on an understanding of how social interaction works. So I'm really keen to hear about that. It's sort of from first dates to medical communication and from sales encounters to crisis negotiation, for example. She's, in fact, one of our members on the NCM Advisory Board, but that's not actually the reason why you're here, because Ali Hanbury, our senior engagement manager, was listening, I believe, to BBC Radio 4, the Life Scientific and thought you were really an engaging, fantastic speaker and that's how this came about and we thought, okay, this would be a really nice opportunity. So your work is having a huge impact and, for example, your crisis negotiations and it's also something that Ali, our senior engagement manager, is particularly interested in, is, yeah, of particular relevance and of interest. You've published a book, I believe, last year on Crisis Talk and you've had 150 research outputs in fact and books of authors and publications, of course. And you've had various work or you've worked extensively with external partners as well, so that's obviously relevant to NCM with the sort of public, third and private sectors and you've given many talks, for example, on Google, Microsoft, TED and so on. So it's really nice to see this sort of, yeah, link between academia and obviously the sort of applied world. There are many things I could mention. I just wanted to also emphasize, obviously, during the COVID-19 pandemic, you've been able, I'm very impressed by that, to work on the Government Scientific Advisory Group or contribute to that stage and you've been, you've been having, yeah, a number of honours, for example, the Wired Innovation Fellow or you've been awarded the Honorary Fellowship of the British Psychological Society. I think I, there are probably lots of other things I could mention, but in today's talk you will be talking about sort of conversation analysis as a research method. So obviously that links extremely well to the topic of NCIM. What is it? What is it not? What can conversation analysis actually do? What can it not do? And whether it's a qualitative or quantitative approach. Again, we have many discussions at NCIM about different types of research area activities and how it raises also really interesting questions of how we find out about the social world, conversations that happen around us and so on. So maybe on this note I should hand over to Liz and maybe I'll just go through the slides. Okay, so on this note I'll hand over to Liz. So thank you very much, yeah. Thanks very much Gaby and good evening to everybody here in the Room at the Royal Society and hello to everybody online as well. Before we get going properly I thought I would start with something that should unite any academic in the Room regardless of your discipline whatever research method you use, whether you're a qualitative researcher or quantitative researcher and that is the ever present problem of reviewer 2. So this title, A Method in Search of a Problem, was inspired by a reviewer, not of a grant, sorry of a grant rather than a paper. Who asked me this question with the intonation that you can probably imagine? It's conversation analysis just a method in search of a problem and of course they hadn't really thought it through that I am a conversation analyst so I'm going to use this data, this question as data forever as an example of why yes no questions in this case a negative interrogative question are much more interesting than you might initially think. I also decided that this is actually quite a good way to describe what I've actually been doing with my own work in conversation analysis for the last 12, 13 or so years because it's been almost exclusively co-produced with and for non-academic partners with the aim of addressing some kind of conversational or communication problem from health care and legal partners through to technology companies and beyond. So damn you reviewer 2 but also thank you for this quite interesting way to think about conversation analysis. I'm not going to talk too much about individual projects today but I thought I would explain a little bit about what I do across three, sorry five problem areas that conversation analysis has maybe something interesting to say hopefully you'll find it interesting. So I'm going to start with a psychology problem, then a details problem, then a communication myths problem, a hard data problem and a conversational products problem and I'll explain what I mean by all of these problems as we encounter each in each section. So to start off with then a psychology problem and I'm saying psychology because psychology is my home discipline but you might substitute psychology here for other social science methods when it comes to what it is I want to talk about that is a problem in psychology at least that conversation analysis might address. So as Mattias Mel argues here, laypersons often think of psychologists as professional people watchers. It's ironic then that in our discipline actually very little naturally occurring observation takes place and this is something that of course anyone who is a psychologist will be very familiar with that we're thought of being kind of nosy parkers who just you know observing everything all the time. But in fact we tend not to look at naturally occurring behavior you know where all the juicy things of life actually happen. Instead psychologists simulate it they ask people to report upon it later on surveys in interviews and so on. And so as Mel continues the Psychological Scientist Toolkit should have something a method to directly observe daily life where moment to moment behavior naturally happens. Conversational analysis has been providing this toolkit for over 50 years. Its origins for those of you who don't know are in sociology. It was imported into my home discipline of psychology through discursive psychology and of course there are CA scholars in pretty much all disciplines but maybe particularly anthropology communication and media linguistics and so on with applications far and wide. So what actually is conversation analysis it's much more than a toolkit it's a theory of and method for studying human sociality. Conversation analysts study social interaction in the wild as it happens not simulated not role played not experimentally produced not reported upon post hoc in surveys or on interviews. Our primary data therefore are recordings and what we do with the recordings is then transcribe them using a universal technical system that permits a very fine grained analysis of what people are saying as well as how they're saying it and when in the conversation they're saying it at what pace with what intonation with what in aggregate embodied conduct all of the things that produce social interaction. And I've often used the metaphor of a conversational racetrack to think about how taught works that the idea that conversations have a kind of landscape to them an architecture a structure and they unfold turn by turn one thing coming after another and one turn at talk affording or constraining the next thing that happens. So in any given interaction we move through what we might think of as projects so like openings or reportings or advisings or offerings or requesting or closing. And people often think that real talk is just too messy to capture and study in any kind of scientific way and yet when you zoom out and look at the whole conversational racetrack the whole architecture the whole landscape you can start to discern the components and its systematic structure. When it comes to these sorts of projects that happen along the racetrack any one of them can be done smoothly with no misunderstanding or they can be full of miscommunication and tension and friction. So I've looked at lots of different racetracks from people on first dates through to police interviews with suspects crisis negotiation people talking using technology lots of different settings. So to bring us back to Mel's quote the whole point of conversation analysis is to study moment-to-moment behavior naturally occurring as it happens in the wild and for conversation analysts this moment-to-moment detail is really important and the Shagloff here says who's one of the founders of the discipline social science theory theorizing must be answerable to the details of actual natural occurrences. It's quite a challenge for a lot of social science. This emphasis on detail brings us to problem two the problem of detail. One of the reasons people often articulate for doing qualitative research is to enfranchise and empower the real voices of participants so eliciting people's accounts and experiences in their own words in their own voices to generate this rich data. Yet when you look at reported research in qualitative methods quite often not always but quite often the detail of what people say their actual voice and especially how and where in an encounter they say things those things are stripped away along with the occasion they have for saying it in the first place by which I mean what the interviewer asks or how the question was worded. Conversation analysts for some is one of the soggyest of social science methods with its focus on details and gaps and pauses and these tiny things in social interaction and we're accused of you know rather than stripping detail away of including far too much detail and therefore avoiding sort of being able to see or say anything about the big picture. This this thing about CA having a focus on too much detail often reminds me of the apocryphal story of Mozart and what apparently Emperor Joseph II supposedly said after the performance of one of his operas too many notes dear Mozart too many notes and Mozart's reply apparently was just as many as necessary your majesty so here obviously there's there's quite a lot of detail on this manuscript. If you haven't seen one before here's a conversation analysts transcript produced using a system developed 50 years ago by Gail Jefferson another founder of the field and actually I quite like music as an analogy for some of what we do because like Jefferson both systems are standard they're universal and if you know how to use them you can represent sound and then do things with that representation as well as the original score. I want to show you why details are important. So this transcript that is here you won't be able to don't worry about seeing it in detail at this point. What you're looking at is a transcript of a leaked audio recording of an off air conversation about five years ago between John Humphrey's BBC Radio 4 today presenter at the time and John Sopal another journalist and they were talking off air leaked about another colleague Carrie Gracie and you might remember Carrie Gracie had resigned from the BBC in the context of a gender pay gap at the organization. So this was a this was a big story the recording was was online and Buzzfeed reported the story and made a transcript and already what I'm just hoping you can see just by eyeballing is that those transcripts are quite different. So to keep going with the music analogy just for a minute one is the easy to play version whereas the other is the original composition. So let's now look at just one little bit of the transcript. So again in a way it doesn't you don't really need to see the detail too much but what you can hopefully see is that according to the Buzzfeed transcript at this point in the conversation only John Humphries was talking whereas when you transcribe it using more get put more detail in what you can actually see is that John Sopal was participating quite a lot in this encounter. So lots of basically sort of supportive laughter, affiliative laughter and so on and we know for example from other conversational research that if someone's saying something obnoxious one thing you can do is stay silent and that will probably stop them and make them change their trajectory or something like that but if you laugh along with it then they'll probably keep going. So this is one of those moments where you know the detail is telling you a slightly different story compared to the transcript that was the one that was sort of popular online and even more interesting maybe and I'm not remotely defending John Humphries heaven forbid but all the headlines pretty much were about Humphries rather than Sopal. Now the reason that I'm showing you this is that conversation analytic methods partly get their power by simply exposing, specifying and explicating things that are happening and sometimes these differences in something like an orthographic transcript and a technical transcript are really important. So my colleague at Loughborough Emma Richardson has shown how different police interview you know recorded police interviews are when they're transcribed orthographically for the legal process compared to a conversation analyst transcript and you can probably now imagine the kinds of things that would be different there that might be quite important to know. The technical detail is also important for doing analysis. So I'm going to give you an example of how a tiny detail in a transcript cracked open a whole analysis. So a few years ago, well quite a long time ago now when I think about it, I was conducting some research on neighbour disputes with a colleague Derek Edwards and our initial interest in neighbour disputes was okay, how do we study something like neighbour relationships? How do we study them in a naturally occurring environment? So we got access to community mediation sessions where mediators were trying to help the parties in dispute and then initial inquiry calls into mediation services, calls into the local authority, local council, calls to environmental health services and police interviews where suspects have been arrested in the case of a nuclear, sorry, a neighbour dispute gone nuclear or semi nuclear. Our research originally focused on how matters of identity creep into neighbour disputes. So how an ostensible noise dispute starts somewhere along that conversation to turn into something about the kind of person who's living next door. We got really interested in matters of ageism and racism and particularly how they were handled by institutional parties. Towards the end of the study, I became particularly interested in these initial inquiry calls to community mediation services themselves and the fact that despite mediators offering a free service to people in dispute, many of these calls ended with the person saying no to mediation. So somewhere along this conversation people either got engaged with the process or said no to the process. So what is it that's going on in this sort of naturally occurring laboratory situation where the outcome is an endogenous feature of the encounter? And these are all real calls. It's no actors pretending to have a neighbour dispute. We're not asking people about their neighbour disputes. The mediators aren't doing training. Everyone is in a real call where the stakes are those particularly real stakes. So initial inquiry calls sort of zooming out and thinking about the whole arc of these telephone conversations. They started off with project one opening the call establishing you're in the right place. Project two explaining what the problem is with next door. A big story about somebody horrendous who lived next door to you. And then project three where the mediator starts to explain what mediation is. And because people don't know what mediation is this always happened at some point along the telephone call. And so I wanted to zoom in on project three because this seemed to be one of the points at which people started to disengage with the idea of mediating. So I'm going to give you an example as I say of why detail matters in understanding how mediators failed or succeeded in getting this project three right this explanation of mediation. So I found two things. First of all mediators always explain mediation. And they did it in one of two ways. They either explained mediation with a sort of philosophy behind it the sort of ethos of mediation. They'd say you know we don't take sides it's impartial it's voluntary we don't have any power. They'd explain the sort of why they do mediation. Or they would explain mediation as a process. So this happens and then this happens and then this happens. So that was the first thing. And callers typically engaged in a process type of explanation but not an ethos or a philosophical type of explanation. The second thing was that callers would resist mediation and start to disengage in maybe three or four fairly routine kinds of ways. And one of them was saying that they don't want their neighbour to know they've been complaining about them which you could probably imagine is a very common thing that you might expect people to say. So I'm now going to show you a mediator explaining mediation to a prospective client in a way that proofs successful the process based explanation. And as it comes out you're going to see the clip come out line by line with an anonymised recording and a transcript. And I want you to just watch for this little tiny detail that would probably not be present in an ordinary transcript but it turned out to be quite important for where I started looking at these calls. So hopefully the audio will work and here it comes. We're a mediation project in the area. And what we try to help the neighbours that are in dispute what we do first send a letter out to your neighbour straight away to say that we've been in touch with you and ask him whether they would get in touch with us so that we can discuss it with them. And it goes on and this person agrees to mediate at the end of the call. So I'm hoping that you can see that the detail really matters because what we have at the start of line 6 is this little thing that conversation on this call repair where the mediator starts to say probably what we do first we write out to the letter to your neighbour straight away just to say that you where was that headed you've been in touch with us but immediately deletes that and swaps it for we've been in touch with you making the mediation centre the agent of things happening in the life of the people in dispute to try and sort of design out one of the regular objections which is but won't they know I've been complaining if you say it the other way around. So the detail here really matters and this was one of very early bit of transcript that I was looking at and listening to and just noticing not only does this tiny detail matter but it's also starting to show us something about the mediators sort of tacit expertise and experience that the mediator here is kind of catching herself before she does the thing that is most likely to lead to one of the very things that people object to along these conversational race tracks. So the next step was to go through all of the materials and have a look at all of the project threes all of the explanations of mediation and start to figure out what is it that seems to get people turned on or what do they object to. And so I did lots of talks to mediators I started you know saying I've got something interesting to show you mediators who are some of them I've been working with some by word of mouth and at some point I crossed over from community mediation to family mediation and was finding lots of similar sorts of barriers to people agreeing to mediate in those initial inquiry calls. And this was me presenting 10 years ago at a conference for family lawyers and mediators and at the time the Ministry of Justice had a new campaign to promote family mediation and I was pretty critical on this stage of some of the messaging that was in this animated video because I knew that in real calls you know rather than maybe like in a focus group or just by like how do you put how do people put these things together this sounds right or this will probably get them in or I'm a mediator and I know what works but actually I kind of knew that these ways of explaining mediation lots of ethos probably aren't going to get people to engage in the process. And so it turned out that somebody from the Ministry of Justice was in the audience I wasn't really expecting but shortly after they invited me to come and work with their communication team and so we redesigned the video and related posters and it even found its way to the US Superior Courts Mediation Service because I'd gone to the states and talked to Mediation Services over there as well. So overall this work underpinned lots of training for mediators which turned out to be very much in demand because really until this point no one was even focused on what goes on to try and get people into mediation all of their training and development was all around being a mediator but of course if you don't have clients if you can't get people into mediation then you won't be able to survive as a service. So the detail matters and another quote here from the founder of Conversation Analysis if you can't deal with the actual details of actual events then you can't have a science of social life. Detail Matters This moves me on to my next problem a communication myths problem. So it's a very strange thing in a way to do any kind of public science or science communication about conversation because the phenomenon itself is very ordinary. No one needs to be a conversation analyst to have a conversation. So talking about talking to audiences like yourselves but probably not like yourselves but just like more generally it's a weird one and it's not like being a scientist of something like a black hole because a black hole doesn't exist in the first place for humans to understand it book, conversation, communication, discourse it's only there for us to understand each other and so we all have our lifetimes experience of interacting with each other and thus we have a lot of anecdote about how we think it all works but something odd does happen between us actually doing the talking and then sort of going meta to talk about how talk works and it's not just in sort of our everyday anecdotes that communication myths begin to solidify and take on a life of their own so this is one of my favorite facts about communication you've probably seen something like this on a slide somewhere else before or someone will have told you that didn't you know that communication is 93% non-verbal and only 7% verbal so this kind of thing that I see all the time especially when I'm in other training events so you may or may not have seen this if you've seen it you may or may not know that the original writer of the research that this gets kind of you know massively transformed from Albert Moravian has spent at least a decent chunk of his career trying to get people to stop quoting the thing there is a very compelling stat we all do love a stat and a pie chart and you feel like you've learned something oh communication 7% words but actually in this particular interview he's talking to a sociologist Max Atkinson and one of the things that they discuss is how easy it is to make this thing fall over because if communication was only 7% words how come radio is so popular or how come we can talk in the dark or how come when I go to France I can't get by 93% that I'm just fine I do need some words this is not of course to say that embodied conduct and intonation and those things are not also crucial but sadly this detail tells us that this stat is a bit of a myth so there are lots of these myths around and at least one of the things that I've been trying to do over the last few years is get people to just engage with people actually talking what are we so scared of let's look at what people are actually doing when they talk to each other and this brings us to the fourth problem the problem of hard data so this is hopefully going to start to put together some of the themes that we've encountered so far in the talk so a couple of years ago I wrote an article with with some other conversation analysts called the softness of hard data and a bit like this talk today the title of that article was inspired by a critic in this case a critic of qualitative research so a professor on Twitter was being critical of a piece of Swedish research on children and long COVID and this is what they said is there any data in there or is it just a list of anecdotes so they were taken to task by another professor and eventually the critic came back with this I appreciate this was poorly worded my point was more whether we should give that much weight to qualitative studies on long COVID in children before we have collected any hard data the sort of binary of hard and soft data words versus numbers qualitative quantitative it's very familiar to many of us but what does it actually mean to generate hard data and how does social interaction and conversation maybe get in the way so a great deal of research data as we all know is generated by questions so qualitative and quantitative data comes often in response to questions and it also comes from the interactions between the researcher and the participant the experiment and the participant but we know not just from conversation analysis but just from our own sort of tacit understanding of language that the way questions are constructed each word and their grammar afford and constrain the things that are going to happen next the data that is collected next in response so one of my favorite examples of this isn't conversation analytic research but a classic cognitive psychology study that I remember learning about during my own degree and it's the work of Elizabeth Loftus and colleagues on eyewitness reports and eyewitness testimony in experimental settings so they basically asked their participants to watch a video recording of a car accident and then they asked the participants to estimate the speed at which the car was traveling and then they read the verb in the question and you know you can sort of see where this is going so depending on the verb choice people made different estimates of how fast the car was traveling one week later they then went back and asked participants whether or not there was any glass at the scene whether they'd seen glass there was no glass at the scene but those people who'd heard the smashed version of the question were more likely to report seeing glass at the scene than those who'd had the bumped version of the question so of course there's all sorts of very interesting applications and implications of Elizabeth Loftus's work and she's done tremendous things to make us understand you know eyewitness testimony and also you know things that we might call leading questions and so on but let's take this down a slightly different trajectory what if we now replaced any in this question with some and was there some glass at the scene what might we might we expect if we just change another word in the question well it might increase the number of participants reporting that they observed glass why am I saying this so um one study that's become really famous I think in in Conversation Unlimited Circles conducted by John Heritage and colleagues in the US showed that in primary care settings they conducted a randomized control trial one word difference in something that the doctor said to the patients reduced patients unmet concerns in this particular setting so in primary care doctors typically open the consultation with what can I do for you today and patients patients typically formulate one thing in response and this issue that patients might typically say one thing even if they've got more than one thing on their mind bearing in mind this is a US context where you know we we know that we're all kind of discouraged from having anything's currently in the UK but anyway what can I do for you today we know that people say one thing so doctors are trained to ask is there anything else that I can do for you today or anything else on your mind and what they did was establish patient groups with more than one thing to discuss by asking them before on a survey and in one arm of the trial the doctors asked after the initial problem presentation and is there anything else and in the other arm and is there something else variations like that and this is because conversation analytic research shows that the polarity of questions will tilt the response that you get so you're more likely to see no responses to anything questions and yes responses to something questions due to the fact that you can basically say yes I have something but you can't say yes I have anything it's that kind of grammatical push this kind of pushing and pulling of language that we're not really very aware of and yet it's happening and this is what they found a substantially higher number of people responding with another concern if they are asked the something version of the question than the anything version of the question so some or any altered what happened next on the sort of conversational race track of a primary care consultation some patients leave with unmet concerns feeling dissatisfied and then maybe needing to make another appointment so in my own work and in particularly in collaboration with Rhinova Sikvoland we've studied different race tracks different contexts and found some other surprising things about the difference that words can make so for example in work on crisis negotiation where we were studying police negotiators talking to or trying to just even get talking to people in crisis maybe they're stood on a roof or they're about to do themselves some serious harm we found that when negotiators started with I'd like to talk to you they got far more resistance than if they formulated a sort of dialogue proposal with the verb to speak and so at the start of the negotiation if they use speak rather than talk they would get further faster than if they started with talk we were using recordings of actual negotiations happening and the crisis negotiators will eventually find their way through to the thing that is going to engage the person in crisis and clearly they don't already know that one thing is making a difference because if this was already available to them as a strategy especially at the start of the negotiation they would just start with it but what we're doing is able to look at the actual conversations what is it that people are doing those people who are really experienced and expert what is it that they're doing and then identify and describe and share it something else that we found at the other end of the conversation back to those mediation intake calls is that callers who are asked if they were willing to mediate were more likely to become clients than those who were asked if they were interested in mediating or would like to mediate and the placement of this example at the end of the conversation isn't just a design thing about the slide to fill the space it's that actually willing worked at the end of the call so if mediators asked people at the start of the call whether they were willing to mediate they didn't get a strong and uptake or they would get resistance whereas at the end of the call especially after resistance willing would get that turn around point and the caller would become a client so it's not just the words you use and of course it's also the kind of action that you'll put you can't just say talk and people will talk or speak or willing there has to be it's embedded in something but where it happens in the conversation matters as well another place where interaction might get in the way of hard data collection is in the laboratory itself in the sort of the gold standard environment for collecting reliable and valid quantitative data sets but there's very little scrutiny of what actually happens inside the black box of experimentation and I guess most of you will be familiar with the clip that's playing here on the screen it's Milgram's work on obedience to authority very widely discussed across psychology and the social sciences for all sorts of reasons but in the last 10 years or so researchers have gained access to Milgram archives which included lots and lots of recordings of the experiments taking place and until this point as one of the researchers says we we didn't really know very much about the detail of what was going on in the Milgram laboratory and what they found was that basically unlike what we might expect when we read about experimental methods where everything is kind of done the same way this over and over again to kind of reduce problematic variables creeping in in fact participants could draw the experimenter into a process of negotiation over the continuation of the experimental session resulting in radical departures from the standardised experimental procedure there are many examples across conversation analytic research across all of our disciplines that show all sorts of things like we think this is standardised but when you look at the encounter isn't or we think there's a rubric being followed to the letter on the page but when you look at the interaction it isn't the interaction of imperative gets in the way and hard data becomes increasingly less hard or just more maybe more interesting so to go back to the critic of qualitative data and the proponent of hard data maybe we're going to flip this criticism and instead say how much weight should we give to quantitative studies that relate to complex real world problems and processes until we understand something of the ways in which the data were collected okay final section final section focuses on what I'm calling conversational products by which I mean things that people do to sort of leverage a conversation and communication in various applied ways so if you're designing communication guidance or training or you're assessing people's communication skills or you're building AI augmented conversational systems what kinds of talk should we be training or assessing or scripting especially since conversation analysis analysis tells us that it's we don't really know how we talk when we go meta on it and our research yields in purely grounded results at variants with how we sort of think we do things in conversation so let's start with a very basic action something that we all do in conversation a request and I want you to imagine that you need to ask for a lift because your car's broken down so you're phoning a friend you're asking for a lift how are you going to do it let's have a look at people doing that hello yeah hi Donnie that's what what my car is what yeah and I know you want and I will and I would but except I've got to leave and it'll be about five minutes okay I've got to cross the bay I'll throw it away okay okay Donnie bye okay so all you hopefully need to see about this is a very basic point which is there's no explicit request and there's no explicit declination yet it's completely understandable by the parties there's a request for something and a declaration of that thing and if you're not picking that up we need to talk later about your communication skills okay so there's a request for a lift what about making a request for service at the vets here's a couple of examples of callers phoning the vet about their pet I can't need them probably the other day I'm wondering how much it costs to get the job done please hello I wonder if it's possible to make an appointment for my cat tomorrow for a follow-up in that operation so you might see some differences between these two calls first one requires the receptionist to provide a price second one requires the receptionist to navigate the appointment system and try to navigate that with the callers calendar the second one is the most common kind of request the appointment the first one is not a common request unless you're a mystery shopper phoning to test out and report back on how good this person is at taking your call and it turns out that mystery shoppers playing the part of pet owners aren't actually very good at playing the part of pet owners if you just hear one you probably don't know that you're talking to a mystery shopper but when you start to look quite a lot you start to see that mystery shoppers just don't give the same kinds of requests in the same kinds of ways to the people who are taking their call yet at the same time they are going to go back and report on how good that service was you could argue that's quite outrageous no one's really looking at what the mystery shoppers are doing and then you can also maybe see through those technical transcripts that there are some other little details there that also varied systematically so if you're a mystery shopper you're more likely to ask about a pet but if you're phoning about your animal you're more likely to talk about my dog if you are a mystery shopper you're more likely to stumble in the production of the kind of animal as you can see there new puppy whereas if you're a real pet owner you are still likely to do an umbs and ars but they're going to be somewhere else they're more likely to be in the bit where you talk about the service that you want so we're starting to see again the detail matters and maybe it doesn't matter that much but you know I think some of these things really do matter and the more you know the more serious and high-stakes settings that you start to look at you start to see is it okay that we are passing and failing people for example in a simulation in this particular environment this particular cohort of medics or whatever it is on the basis of things that we imagine people should be doing in conversation okay what about if you're designing the way Alexa or a conversational user interface asks a question you might give its voice rising into nation right have a listen to this this is a patient these are real don't worry I'm not good these are these are real patient and a real receptionist asking each other questions on the phone through the anonymization you should be able to hear that those are both questions and they both have falling intonation at the end so while some questions do have rising intonation many don't and it was this this nugget of something that is you know every conversation analyst knows that I'd written about in a book on conversational analysis and which had a section on myth busting that a woman from Google Cathy Pearl who was at the time the lead conversation designer for the Google Assistant had found and got in touch to say I'm going to tweak the voice of the Google Assistant when it's asking questions and guess what now it sounds more authentic so myself and Cathy and my other colleague here Saul Albert are starting to try to bring together the worlds of conversation design which everywhere writing your chatbots and conversation analysis so that you can just you know figure out how it is that people are talk so if you're getting these assistants to ask questions that they sound you know authentic it's the way we actually do it Alexa and Amazon haven't haven't encountered our services yet so here's an Alexa asking a question of the day and this is from Saul Albert's data which is the following country flag does not engage across if that doesn't sound weird to you then again you're not really listening to how where the intonation is generally going at the end of a question so if we actually do falling intonation we produce it when we're asking questions but we imagine something else because people just tell us that questions have that rising intonation then of course like here where we're risking building all sorts of biases and myths into conversation design as well as failing to just you know let's again why are we running away from looking at how people actually talk and trying to build some of that you know this 50 years worth of research we could be building in to these kinds of products find an example is a call to 911 emergency services I'm going to play it and just let you think about it as it rolls out and as it rolls out I want you to ask yourself at what point and how come the police dispatcher is starting to figure out what's going on in this call and I suppose I especially want you to watch out for points of overlap where the talkers are talking at the same time so here it comes 127 been there okay what's going on there I'd like to order a pizza for delivery Bamm you've reached 911 this is an emergency line you have a large with half pepperoni half mushroom um you know you've called 911 this is an emergency line you know how long it'll be okay ma'am is everything okay over there do you have an emergency or not yes and you're unable to talk because right right is there someone in the room with you just say yes or no thank you you're giving up it looks like I have an officer about a mile from your location are there any weapons in your house no can you stay on the phone with me no uh see you soon thank you so people often imagine that the reason that this police dispatcher knows that this is a genuine request is because they're calling the police so you can ask for a pizza and because you're calling the police they will kind of figure it out because the context gives you the answer but actually you may not know that that a lot of calls to the police are not genuine they're malicious they're accidental they're a nuisance type of call and so one of the sort of big tasks of the dispatchers is to figure out which of these calls are genuine so just because you put some words into a conversation to the police doesn't mean that you're going to be treated as making a genuine request for for assistance as in as we see in this case here we can also start to see the the points at which the dispatcher starts to figure it out and one of the things that is very interesting that we have started to see in other calls as well is these moments of overlap so how do you sound like you're ordering pizza to the person who is overhearing you and is in the house but at the same time the person who is on the other end of the conversation can hear you as as as genuinely requiring assistance and you can see that these overlaps these placements of these turns is is is is definitely key to it because when we look at other calls where there is as far as we know it's a genuine nuisance call if you like or we see calls like this the overlaps are present or absent and it seems to be a real component feature of how do you be how do you show that you're being something else to the person on the other end of the phone so we've written a couple of things about this so with Saul and William Housley we've looked at this call in relation to you know how do you think about this in relation to designing conversational user interfaces and so on so what happens if you ask Alexa I'd like to order a pizza for delivery you can all go and try it tonight but it's not going to send the police and how would you ever get an Alexa to understand that level of sort of pragmatic purpose in a conversation maybe we'll get there but you can see the challenges when you think about back to even how do you ask for a lift and how do you say no to a lift when no one asked for a lift and no one said no to a lift and yet we totally understand that that was happening and then in a more recent paper with Emma Richardson we've looked at how is it that callers manage and call takers somehow together manage to hear despite the fact that I'm not asking for help you are able to hear that I'm genuinely asking for help and so we've got another nice example of very like the pizza one in British data where the caller starts by saying hi how are you you know that totally pointless thing that happens at the start of conversations that tell you nothing and you're everyone's meant to lie if they weren't there at the start of conversations in ways that are totally hearable as talking to a friend then that caller wouldn't have been able to sound like she was talking to a friend to the person who was in the room so to come to a conclusion then three quick points to sort of reiterate we talked about the problem of psychology this issue that's very interesting to me as a psychologist over the years and to see how much of what we imagine psychologists are up to doesn't really figure in academic psychology you know it's a constant source of disappointment to friends that I'm not actually analysing them all the time and I can't help them with their dreams and all of those things but at the same time conversation analysts are there to kind of do some of the work that we might imagine psychologists are doing when it comes to things like conversational products conversation analysts can identify describe and share what communication or experience or expertise or effective practices what do they actually look like and then finally we all talk but we maybe don't really know how post hoc on reflection it's really strange when we go meta so Nullius in Verba which I learnt when I was putting this talk together today is the Royal Society's motto and it means or is taken to mean take nobody's word for it until you've seen the evidence so thank you okay well now now we get just just five minutes of me I'm Dr John Sutton I am a a chartered psychologist as well as editor of The Psychologist the magazine of the British Psychological Society you can find us on Twitter at psychmag where I've done a bit of a thread there while Liz was talking about some of the articles cover her work but there's I'll start with a bit of a confession I suppose a bit of a irony maybe the I'm acting as a discussant here because I left academia 23 years ago in large part because I felt hopelessly out of my depth when it came to research methods so it feels strange being discussant national centre for research methods award lecture even back then as a PhD student and in the year or so that followed I wanted to tell stories I knew that methods and analysis were going to be central to how I told those stories around my own research and even to knowing what stories were out there waiting to be discovered and I understood that scraping by with a bit of Spearman's rank and ANOVA wasn't going to be conducive to telling stories that really engaged and informed a wide audience so I jumped ship to edit the psychologist which was effectively telling stories of psychology and back in 2013 we got to hear Liz's story in the magazine and ever since then I've been drawn to her work and to conversation analysis in general and I thought that some consideration of why might allow me to reflect briefly on the lecture that we've just heard and set us up for questions so there's three reasons I think and the first is that I'm engaged by change as Will Storr writes in the Science of Storytelling we all are from the very first single-celled organism we have evolved to detect and respond to change we are change detectors and it's there it's everywhere in stories and once you know that and you start seeing it particularly at the start of all sorts of books and films it's also there in many of the articles that have kind of done best on our website on the psychologist's website so for example Annie Hickox wrote for us in 2019 and I'm just going to read you briefly how that piece started and I always point out that I didn't edit this at all this was exactly how it how it came in so she starts I was cooking dinner one barmy sweet-scented summer evening when our daughter Jane called me phone calls particularly on the house phone were not our two daughters preferred form of communication I was accustomed to long strings of WhatsApp messages or even days of silence when Jane and her older sister Alice were busy with studies work or friends in London I figured Jane was calling to discuss the plans they had made for their upcoming trip to work in the States over the summer sometimes a real-time discussion is best for fine-tuning and planning after her initial hello I couldn't hear her voice are you okay? I asked there was a pause and then mum I can't do anything I gently asked her what she meant but my heart already ate with knowing so in that example something has clearly changed and actually I think is is about to change and of course I also use this example because I thought it's the kind of exchange that Liz might study herself she might want to examine the length of that pause she'd probably notice the hello as a kind of stark an unusual way of opening a call and that leads me to the second reason why I think I'm particularly drawn to Liz's work I love that she studies those moment-to-moment behaviours little vignettes and scenes which tend to be part of a bigger picture so it's the actual detail of actual events as she said the seemingly trivial and mundane that on closer examination is anything but so nothing that means everything I love that Alexa Hepburn studies Burps and Stuart Reeves as you heard there considers how we talk to Alexa not Hepburn that's got confusing and Anne Weatherall who I think is is here somewhere hi Anne and has listened out for that sharp intake of breath that a GP might do as they examine you to me this is psychology at its best I sometimes think we psychologists are the wambles of science so if you're not a child of the 70s you probably won't have any idea of what I'm talking about there but basically in the words of the theme tune we're making good use of the things that we find things that the everyday folks leave behind and Liz put that much better when she said conversation analysis addresses the male biomass to problem of psychology and I think thirdly and finally I'm drawn to contrast and contradiction so I like happy music with words about sad things or vice versa posh burgers overthinking love island which is something that Liz and I have in common so what has kept me interested in stories of psychology I think for so long is that we're constantly encountering hard data that turns out to be rather soft soft data that perhaps we should consider hard truths that might be myths and heroes who become villains quests that have dead ends and wrong turns and these this is all the stuff of storytelling and conversation analysis has some ripping yarns so to conclude in terms of a method in search of a problem in in life in psychology there are always problems and I hope that you like me have been convinced by the power of conversation analysis in finding real people living real lives and giving them a voice to tell their stories and actually doing something about those problems thank you very very much I think it was a really brilliant and very engaging discussion fantastic questions from the audience both online and face to face so I'm really thanking you very very much everyone for coming and for being taking part online as well so thank you very much