 Hello, everyone. Today, what we're going to do is look at qualitative coding, especially qualitative frameworks. And what happens when you get into a very common situation where you have a very large number of codes or themes that you're using for the analysis of your project and when that gets a little bit too much to handle. Now, the stuff we're going to talk about is quite specific to Quercos, but most of it's quite generic, actually, to qualitative software and any software package that you're using. So hopefully it'll make some kind of sense. So what I've done here is I've created a coding framework. Now, this is a fictional one that I've created. And I think you'll see that it's quite a mess. This is the kind of project that I get sent a lot, helping out people with Quercos. And they say, help. I've been doing ground theory. I've been going through my project. I've created an extremely large number of codes. Now, on the screen at the moment, we have about 100 codes. That's probably about a typical size for the projects I see. I think that's a fairly common number of codes in qualitative analysis projects. Obviously, this can vary greatly. You may just have a few things that you're trying to pull out. So you could have less than a dozen codes. You may have, I've seen projects with many as 500 or 600 codes. Now, depending on the project, that might be suitable for you. But you will need some way of working with those codes and managing them. So in Quercos, you create these codes. Just by pressing on this plus button here, we can create a new code. We're going to call this one in the center. You'll see why in a minute. And codes are created with a random color at a random place on the canvas. And then you can click and drag to move them around. So it's very easy to make codes as you go along. Now, you can also drag and drop codes. So drag and drop text onto the plus button here to create a new code which already has that text in it. So we call this LSO from the London Symphony Orchestra. We've now got a new theme called the LSO, which is just down here. And it's in this yellow color here. And if we put our mouse over it, and the bottom left of the screen here, it will show us what the name of this theme is LSO. So it's very easy, perhaps too easy, to create new themes in your qualitative coding. One of the things you can do to make this a little easier, and I'm just going to do this now, make things a little easier for you, is you can zoom in with this button here to zoom into the canvas. So you can see if you've got lots of small bubbles like this, you can zoom in and see one particular area. So one of the first ways that you could start to sort, group these codes is kind of geographically. So with this canvas area you have on the screen, as I showed you when I dragged these codes to the Barbican Center and the LSO ones around, you can actually kind of put them in a particular place on the canvas, and this can help put things of a similar topic, similar themes together. So for example here, you'll see that I've got these themes here, which I've put together kind of closely together. We've got seats here. These are the different types of seats in a theater. So you may be getting a kind of sense as you look at some of these codes that I'm talking about, music, theater, plays, basically anything that might happen in a concert hall. The Barbican Center and the LSO will be out of clues as to what this project might be about, but I'm not gonna reveal everything just yet. But one of the things that I'm interested in looking at is how people in my research project are talking about the seats, the quality of the seats. And how I've done that is by having these codes here. So I've got back front, restricted view and upper circle. So those are the different kinds of categories of seats that I'm interested in looking at. So what I would do is I went through as somebody was talking here, let's say this section here is about, we imagine this is about the upper circle. We drag and drop some text onto the upper circle bubble. Now we've got some text on the upper circle. And if we want to see everything that we've coded once we've gone through all our sources to see what's on upper circle, we don't click. And then it'll show us all of the code on the right. Home button on the top will always take us back home. So we can create some kind of, make this a little bit easier to manage just by having this kind of geological crust. So just have some bubbles which are close together because they're related. So that's the first thing we can do. But we can also go a step further and we can actually create hierarchies. We can create a hierarchy of codes. So all of these bubbles here, back, restricted view, upper circle, front, they are types of seats. So where we have different types, different subsets, we can create a hierarchy of them. And we do that just by dragging and dropping bubbles on top of each other. So if I take the restrictive view bubble here, I can drag and drop that onto the seats bubble. And now restrictive view is a subcategory of seats. So it's a type of seat. And we can drag back onto there. Back could be about many things and so could front. But in this case, it's about seats. So drag and drop those onto there. Great. So now you can see when we hover the mouse over the seats button, the subcategories pop up. So we can still work with them, but they are kind of clustered together in this kind of petal fashion. So they take up a little bit of space. We've got a hierarchy. So now you can see we've made our canvas already a little bit neater. Excuse me. So what you might notice here is that there are probably a lot of themes like this that we do want to bring together in groups like this. So for example, if we can find, we've got a bunch of different things here. We've got Thomas Middleton. We have Sartre, we have Chekov. And here we've got playwrights. Okay. So Tom Stoppard, Brecht. So you can see now in the coding, there's a little bit of a theme emerging here around different types of playwrights. And again, the same way we can drag and drop bubbles onto here and group our playwrights together. And that means that if we wanted to see everything we've written about different playwrights, everything we've coded to that, we could just double click on the playwrights category. We'll now see our hierarchies expanded here. So we can see the subcategories here, but we can also see any texts that's associated with them. So now we've created these two little groups. I wonder what this project would look like if we use this grouping category to put them all together. Well, I've done this already for this project. So let's imagine you went through and you started putting those into different categories. So choose opening existing project. And I'll choose messy canvas tie-in lead. I think that's, no, choose that one. Yeah, here we go. So now we've put everything into groups. We can see our project has become a lot more manageable. So we actually have only three, four, five, six high level codes, which keep together everything that we want to be looking at in the project. So one of these is people and that's not specific persons. That's different roles that people have connected to the theater. So we've got choreographers, we've got the audience, we've got prompts, performers, actors, conductors, the lead and understudy, producer, critic. So those are the different people involved, but different aspects of what could be used to critique the performance. We've got choreography, reviews, the press, opening night, intubation, voice, timbre, projection, lots of things that might affect the popularity of the play there. It's gonna be a little bit, so you can see it's a bit better. We've also got composers. So composers of musicals here, an opera. So those are all neatly put together. And you can also see, I did something else. I've actually created a color theme. So all my composers are different kind of shades of bluey, pinky, red. You can see my playwrights, you may have noticed earlier, all different types of yellow, green, yellows. And again, this can help a great deal, not just to kind of find the theme you want, but also when you're looking at the coded extracts here and you're seeing the highlights. If you're seeing pink stripes, you might not remember that particular shade of pink means handle, but you'll know that it's a composer. So you can start to see these things pop out from your data as you go through and code them. Now there's also one other thing we've got here now is we actually have under theater, this is in the wrong place actually, we have Shakespeare. Okay. And there are sub-sub categories here. And Shakespeare's not in the right place because Shakespeare should be in playwrights. So we can pull that out from theater. We can put that into playwrights, great. So now we have Shakespeare's with the playwrights and we've got separate codes here for a bunch of his plays. So we can have sub-subcategories. So we talk about playwrights, we talk about Shakespeare, we can talk about specific plays. So this is a way, you could imagine we could have hundreds if we had all the different plays of Tom Stoppard, if we had all the different plays of Mamet and Thomas Middleton and Brecht, then we would be talking about a lot of different themes here and managing them in this sub-subcategories way will keep things neat. But obviously it's a logical order. So if we're selecting a piece of text that we want to code, we drag a drop and put it onto playwrights and then Shakespeare and then MVA. So that logical kind of grouping also helps us speed through the coding process. So that's one thing that we've done. But what if there's something else? What if there's some kind of category that we want to pull across that doesn't have this kind of hierarchical structure? Well, what if we've got themes which want to go across this hierarchical structure? Well, there's a different way to do that in quirks which is called the levels or grouping systems. It will be in the new release in a couple of weeks. This basically allows any number of themes, any number of your quirks, bubbles or codes to belong to any number of other groups. So for example, what's a good example here? So for example, we've got here theater sound engineer. So that could be in people, but it's also been related to theater because it's quite a technical role. So we may want to have something that goes across it. We have composers here, obviously attached to music, but some of them may have written plays as well or playwrights may have had their plays turned into opera or into other forms of entertainment. So let's try here creating a new category. So we choose quite properties. That's where we can change things like the color. We can put along a description. We can change the title. We've got this box here for levels. So let's click on manage levels. Create our first level and call it music. Create another one, which is called issues. We've got two very different categories, the ones which are actually going to go across all of the different codes that we have. So composers, I'm going to associate that with music, but not with technical issues. So I'll say that. And Philip Glass, okay, so he's a minimalist composer. So we'll put him into the music categories as well. No technical issues associated with him, unless we're trying to play one in a very complicated hostility. Now the playwrights category, there's nothing here that we want to have associated music, but there probably is here for the people. So for example, we've got the conductor. Now that's going to be something we want to associate with being about music. So you can see here we're going through and we're creating another way to group these codes together, but they don't have to be in the hierarchy. They can go across these. So any others here? So the current, no, I don't think any of those other words will probably be specifically about music. I don't think any of these will either. They might be though. So for example, voice. Let's put that as associated with music, but that might also be associated with plays. So we've created another one called plays. And obviously the voice is going to be very important when we're talking about performance. So it may be a baritone, it may be countertenor or something like that, but it also may be important in the play how someone projects their voice if they're not singing. So that way the subcategory voice here we know is something that has an element about it which is about plays and is about music. Is there anything else here? So genre, okay, right. So this will be a good one. So we've got musical here. So that's definitely going to be about music, ballet. Okay, so that's an interesting one. So this is probably going to be music as well, but it's something else, opera, great. That's going to go into music. And then for technical issues we may have things like this sound engineer property here. Let's not get too much into technical issues. So what would happen here is using the query view here, now you can see we can choose the level here. We have our levels, music, technical issues, plays. If we choose music, we'll get results here from all of the things that we associated with being about music. Now we didn't do any coding in this project. There's nothing coded to any of these themes that's not showing us any results. But this would show us only results which are about from the themes which we tagged about being about music. And we could likewise do the same for plays or we could do both. So we could do things which are about plays and music. So that's where the voice category would come up. If we did a query there, if we had some results voice would be the only thing giving us results here. So that's another way that we can kind of create these groups and categories. But there may come a stage where even with all of this it's just still a little bit too much to handle. There's a couple of other things in Quercos that might help. So one of them is on this view button here, there's lots of different ways that you can arrange this so we can put it into ascending order. In fact, let's just go back to the messy project for a second. So this would be very difficult for us to navigate, but if we put it into alphabetical order so this now puts it into a grid on alphabetical order could be a lot more easy to find actors adaptation as you like it audience by a beta. So they're not in any type of theme here because we undid the grouping that we did but now alphabetical order that might help. The other option that we have here is there's a list view and this is a lot more familiar if you've got something like review something like Quercos or a Vivo or QDA before. This is the kind of standard thing. It works in the same way as the bubbles. You can just drag and drop text onto them but because it gives you a list like this it can be a little bit more difficult to easy to navigate even when you've got these kind of subcategories and subcategories because everything is expanded out here. So where's my Shakespeare? There's my Shakespeare, turn with the shoe. So you can see all the subcategories here are expanded but let's go back to the bubble view here. There may be a stage that we turn off the alphabetical order. So there may be a stage that we decide that we actually don't need this many themes in the project. Now, when you go through and do the coding it's such a laborious process with qualitative analysis that it kind of makes sense to use as many codes as you think you might need. And the reason for that is it's a lot easier to merge themes together at the end than it is to break them apart. I'm gonna show you doing both of those things but take my word for it. Merging is easy, splitting apart codes is difficult because you have to say for every code in your theme which of the new things you will just put it up to. So for example, if we have this theme here oratorios, what if we wanted to split it up into oratorios written before 1800 and oratorios written after 1800? Now to do that you'd have to go through oratorios, everything you coded about in oratorios and you'd have to put those texts into separate categories. So separate categories for before 1800, after 1800. So essentially we'd have to go through and recode everything in that theme. But what if we've been kind of pedantically detailed all the way along? And we've done all of these things. We've created so many themes that they're gonna be very useful for us. Okay, so oops. So you see here, I've got two different categories here. So this is a slightly tongue-in-cheek one here. One for the Scottish play and one for Macbeth. I don't know if you know the old joke but there's a superstition about actors saying the word Macbeth as it's bad luck. And so it's referred to as the Scottish play. So what if I went through, I started coding Scottish play and Macbeth together and I decided that that's absolutely ridiculous. People aren't saying anything different about the ways they say that. Let's merge these two themes together. It's actually very easy. So if we right click on the Scottish play choose merge quirk into, that'll put everything that we coded into as the Scottish play into the Macbeth theme which is above it. So we'll choose merge quirk into and then we'll choose Macbeth. Anyone's called it Macbeth merged and they've changed but we'll just call it Macbeth. So now we got rid of the Scottish play but the contents were all subsumed into Macbeth. So we didn't lose anything. So that shows how easy it is to merge things together when you decide later on that you don't need that level of detail. It's much easier than breaking it out afterwards. So I think it's, a lot of students get to a stage where they have a canvas that looks like this. They have hundreds or hundreds codes and they panic that it's just too much to manage. It's too much detail. It's fine. You just need to have a separate stage of kind of coding your codes, bringing them together, merging them possibly and also putting things like this so you can merge together, group together. And so you may get to a stage where you've got a project that looks a lot more like this. It's grouped together. It's much easier for you to work with. It's also much easier for other people to work with. So if you're trying to share your coding with someone else, supervisor, someone else working on the project, collaborator, obviously having a logical structure like this can be really helpful. Now, the messy kind of canvas we had earlier, for a lot of people, you may get so much muscle memory, you know where things are, you know what you've called things, you're mean by those, but it makes a lot of sense like that and it's easier to have that messy canvas for you. When you start to want to share and communicate that project can be a challenge. So all of these different things can help you to do that. Okay. So that's the basics that we were going to talk about today about managing your coding framework. Remember, whether you're doing grounded theory, where you can create things on the fly, where you're creating a framework at the beginning and coding very strictly to that, both of these problems and issues can occur. So it's something that you should always bear in mind as an unfinished project. And in fact, one of the things that I would argue is most people have multiple layers, multiple levels of their coding. They will go through it multiple times and they will do other things. So for example, that's one way you might use the levels. You may have level one be my first coding attempt, level two, my second coding. Well, why might you do that? Well, if you see the kind of coding that we've done here, it's very basic what's called descriptive coding. So it's very literal analysis of what's in the text. Literally someone is talking about open night. Literally someone is talking about voice. Now we've had to do some interpretation to decide what goes into those themes, but they're not very high level themes. They probably don't connect very well with our research questions. They probably don't help us get to an answer about what we're trying to find in the project. At this stage, once you've been through coding like this, most people will find that it's time to do a second level of coding. So to keep these codes in coding, but to have a higher level of analysis. So for example, we may have here, there's leading to the demise of opera. So maybe one of the things we're talking about is why opera became so much less popular than it was a couple of hundred years ago. And when we go through in code, there may be specific things that our participants are saying that might illuminate that. You can also see that some of these codes might help. So maybe this is about reviews. Maybe it was about the popularity of musicals. So some of these themes here you might kind of pull out and start putting in this kind of like higher level category. So maybe musicals is a thing. Maybe we didn't talk about seats, but we don't have something about the cost. So maybe that was a big thing. So we can put in a little cost. So we may also get to an even kind of higher level analysis in culture, trends, fashions and things like that. And so we're trying to get to a kind of higher level. And then what we might say is that there's a research to question that's closer to our thesis. So if we're thinking that maybe access to recorded performances was one of the things which led to the demise of the opera, then that's something that we can pull out. And again, something that might come through as one of the very kind of basic things we've already coded. So maybe people were talking about first time they owned a CD, first time they owned a radio and so forth. That might explain why some of these factors have changed in the long run. So I would create a separate level for these kind of low level and these kind of high level themes. I'm gonna keep these levels that I have at the moment, but we'll create a new one. Oops, four. So I'll put that as one of my high level codes. And then later I can see, and you can merge things into there of course, some things that might be much closer to my thesis and my research questions. That's very difficult with this kind of abstract example to see how that will relate to your own research. But it's a very common thing that people will go through and do one to three times, reading through their data, coding through their data as a kind of a thesis emerges, an idea about what's the commonality and what's the interesting thing which is popping out across the data as a kind of a higher level theory. So these theory level codes are very good as putting it as a separate level. So that's pretty much all that we were going to do for today. Please let me know if you've got any questions, anything you would like me to cover in this session. I said today is slightly odd one. We're actually just talking about a very common issue, but one which lots of people talk about. So this is managing a large number of codes and what to deal with them. I'm just trying to show you how easy and colorful and fun that could be to do in Quarkus. So I don't have any requests from the viewers. So thanks for coming along today. As always, do go to our website, download the free trial, and you can also see that there are example projects that you can use there. There are blog post articles. We've also got video guides and tutorials which will help you go through and learn how to do this. Now the blog is also very useful sort of the information here because we have posted kind of a generic qualitative issue. So about qualitative observation, there's an article there. There's also a qualitative mass and monitoring evaluation, using qualitative literature reviews. But you can see there's one here about coding frameworks. So this one here talks about these kind of feedback loops I was talking about higher level codes. So there's a lot more detailed information on the blog if you want more on this specific issue. And that's just www.quarkus.com forward slash blog. Thanks so much for coming along today. Hope this has been useful. And do get in touch if you have any questions. Thanks very much for coming along today.