Speaker 1: Hi, Dr. Eli Lee over here of DEDUCE. My training is as a quantitative psychologist and I've been working in the field for over 20 years. But over the last 15, I really dove into doing mixed methods research. We found out a long time ago that there really weren't any tools to help us do the kinds of work that we wanted to do when we were using methods from psychology, anthropology, sociology, marketing, research. So we came up with a tool. It's evolved over the last 10 years and now we're really happy to show you what DEDUCE can do for qualitative research and mixed methods research. So what are mixed methods? Well, most people are pretty familiar with quantitative research methods. Quantitative methods are great for learning about what kinds of things there are, what kinds of people are they, and how many. So we might ask questions about people's demographics, their gender, age, income level. We might get scores. We might have them rate on a 1 to 10 scale how important to you is education, religion, family. And we might have them fill out other kinds of scales like depression. How happy are you on a scale of 1 to 100 or a bunch of items that we fold together to make scale scores. And so what do we do with these things? We can analyze them as individual variables like age group, 1, 2, 3, how many, what percentage of our population sits in each of those categories. We can do bivariate relationships looking at age by sex, for example. So here's the age group here, but when we break it out by males and females, we see there's a different percentage in each of those groups. And we can do multivariate types of analysis as well. So here's an average rating of how important family is to you broken out by the age group and by sex. There's all kinds of great and powerful things that we can do with quantitative data analysis, and we're not here to talk about that. Qualitative methods seek to understand human behavior and the reasons for those behaviors. A lot of people are less familiar with qualitative research methods than they are with quantitative. But basically qualitative data and the methods that generate them are seeking answers for why people do the things that they do and how they do them. What motivates them? What feelings drive the kinds of decisions that people make? What values? What things about their cultural background are important in understanding why they act the way that they do? So as a simple illustration, let's say we're interested in understanding why people make certain decisions about hotels. We go out to the field and have people just tell us stories about their most recent hotel decision. Typical approach to these data from qualitative research perspective would be to go through those texts and look for the themes that people talk about in a consistent way. So let's say for example after we look at a whole bunch of these stories, we hear people talking about luxury, sophistication, and intimacy on a pretty regular basis. The next thing that we do is go through those stories looking for sections of text where people talk about one or more of these things. We can also apply a weighting system. Here we've got an example of a 1 to 10 weight scale based on how important it is. So let's say in this story here's a piece of text where they're talking about luxury and intimacy. So we'll say 1 and 3 are the codes and they really thought luxury was the most important factor. So we'll give that a weight of 10 and intimacy was moderately important, we'll give that a 7. So we continue going through those texts looking for all the sections where people talk about one or more of these texts. So here's another one. They're talking about sophistication. Sophistication they say, well you know people talk about that but it's really not that important to me. So we'll give that a weight of 2. Here's one more. They're talking about luxury and sophistication. Again sophistication is really low, luxury remains really high. But wouldn't it be cool if we could put all these methods together? Some famous researchers have talked about the fact that all research methods have flaws. So by mixing methods together we seek to capitalize on the strengths of each and avoid some of the weaknesses. Deduce is great just for qualitative data analysis but when people are trying to bring it all together in a true mixed method design that's where deduce really shines. That's why we built it and that's what it was designed to accomplish as efficiently as possible. These studies can be really complex and people are looking for answers as efficiently as they can. With today's technologies we were able to put together a tool that accomplishes all of this and gets you to your answers very quickly. So come check it out. Alright welcome to my computer. I'm going to just log in to deduce but remember as a web application anywhere I have a computer and internet access I can log in to my project. So here's our hotel project. Our resources are here. Here are the tags we've already created. Here's some excerpts that we've already created and tagged. Our descriptor data are over here. Here's the fields that we've defined and here are the data themselves. And right here on our dashboard we get a glimpse of the many data visualizations that deduce produces automatically. Most of them can be modified and all are dynamically linked to the qualitative data that they represent. So if we're interested in drilling down to those excerpts from the 50 plus age group that were tagged with cost, simply click on that bar, pulls those excerpts up and we can go ahead and exam them further. Also deduce is very transparent and allows you to move seamlessly throughout your database. So if we pull up excerpts that we're interested in and we want to drill in a little bit deeper simply clicking on that excerpt takes us to the excerpt itself in the context of its original source. So let's go back home. It's also important to point out that all the charts, graphs, list of excerpts, descriptor data can all be exported from deduce with a simple click to be popped right into presentations or imported into other software. So let me show you one of the other important set up activities, getting your documents or resources into deduce. I'm going to go ahead and just import a document, prompts me to find the resource on my computer. I'll go ahead and give it a title, submit that to the database. You can see that deduce supports images in virtually any language. So here's some example text from the hotel study. So I'm going to go ahead and just block a piece of text to create an excerpt. I'll right click, create excerpt. Now I can go ahead and attach tags to it. This person talks about sexiness factor but that's not particularly important for them. Also sophistication is reasonably important. We'll give that seven. Go ahead and create another excerpt here. Service is high on their list. Let's go ahead and create an excerpt there. Give service a ten. Another thing that we need to do when we put resources into a project is attach it to the appropriate individual so that it's linked to the descriptor data. Let's go ahead and attach that person here. Okay, so we can see that one descriptor has been attached now. Now the really fun part. Let's get over to the analysis center. There are a wide variety of charts, graphs, plots and tables in deduce. There's really too much to show in this brief introduction so let's go after answers to just a couple of questions. You see that there are lots of charts, tables and plots that are available in deduce. Let's go ahead and look at something based on our tagging activity. We'll look at tag co-occurrence for example. What we see here is code by code matrix. This shows where two tags have been used on the same excerpt. The numbers, the tagging activity and all the descriptor data expose the patterns in our data. The qualitative data that sits behind those images is what really gives us the richness in this mixed methods research. We see here that 13 times luxury and cost were used on the same excerpt. If it's meaningful to our research that people talk about cost and luxury at the same time, we might want to go and look at those excerpts. Here we can pull them up. We can explore them. Again we can jump back to the resource and export these as well. Let's go ahead and close that. In terms of exposing patterns, bubble plots are really illuminating. This plot gives us information about the average weights or importance or sentiments that were associated with our tagging activity. I'm going to go ahead and break things out by age group. Let's use the tags luxury, sophistication and intimacy. This plot actually gives us four dimensions to look at our data. The plots themselves can communicate a great deal, can be exported and popped into presentations. We can also learn a lot more about the pattern by drilling down into the excerpts themselves. This bubble here represents the age group 50 plus and this group was relatively low in the importance of luxury but relatively high in sophistication and intimacy in their hotel decisions. If we're putting together a marketing message, we want to come up with something that's really going to resonate with this particular subgroup. We go ahead and open up those excerpts and we can really get a feel for how people are talking about these things and understand the reasons for why they feel the particular characteristics are more or less important when they're making their decisions. There's so much more I'd love to show you about deduce but I think we're out of time. Thanks for checking in. We've got a number of other videos on our website that give you step-by-step instructions about how to set projects up and how to get the best out of deducing your work.