Steve Tippins outlines five mistakes in qualitative research: poor recruitment planning, confusing research questions with interview questions, misunderstanding data saturation, inadequate bias checks, and misaligned findings. He recommends planning recruitment early, pre-testing interview guides, stopping when you hit two consecutive similar responses, recording and transcribing interviews, and backing each theme with at least two participant quotes.
Chapters
Today I want to talk about some questions we get about qualitative research and challenges that may arise. And there's a number of them I've written down, so if I glance off a little bit to remind myself, that's just what's happening. One of the things, and this is a fairly common one, is difficulty recruiting participants, finding participants, figuring out who's going to be part of your study. Yeah, I mean, this is really important, uh, because you're typically only dealing with 10, 12, 8-something people. You want to find people that are valuable for your research. The dilemma is you have IRB, REB, maybe in Canada or other places you have institutional requirements that say you may not go out and find participants until you have approval, which means your proposal has been completed and your committee has approved it, and then the institution has said yes, the research is not going to hurt anybody, go ahead and start recruiting Participants, which means you're well along the way. You really, really, really should be thinking about recruiting participants from the beginning. And do I have access to people? And is it gonna be feasible? And is there a backup place to find people? Because participants are the guts of your research. You need them. So you wanna be able to know that they're out there. You can't ask people ahead, but you can say, I'm gonna be doing research and you think there are people who are interested. so that people are prepared to then be officially asked. But do this from the beginning instead of waiting until you get the approval to then finally start to go find people, because then there's a lot more pressure on you and you feel a lot more stressed. How do I conduct interviews to bring about or elicit meaningful data? Well, when you design your interview questions— and there's a difference between research questions and interview questions— your research questions are your overarching guides of what you're looking to answer. Your Interview questions are what you're gonna ask participants in each interview. You wanna be designing your interview questions so that the answers that will come about are gonna allow you to get to your research question, which means you don't just decide the day you're gonna do this, the day of an interview, that here are my research questions. You decide this well in advance, and your committee should be asking for them before they approve your proposal. And then you test them on friends, family, and see what kind of answers they come up with. In fact, your committee may even want you to pretest them with participants that are similar to what you're gonna be getting for your study because we wanna know ahead of time instead of after the fact that I'm gonna get the kind of information that I need. And a lot of these things that we're gonna be talking about is you're gonna notice that, yeah, think about this early. Oh, do that early in the process because we don't want surprises later on down the line when it can mean going back and redoing things. So take a little time to set it up better at the beginning. And I didn't say this so far today, but if you're here and have questions, speak up. I'm ready to take them. Now, data saturation is something that's confusing for people. You see people saying, "I'm going to interview 10 people or data saturation." Well, first of all, I'd rephrase that because if you say, "I'm going to interview 10 people or data saturation," your committee can take in a interpretation, okay, you interviewed 10, but you don't have data saturation, so keep going. If you say, I'm gonna interview 10 people or data saturation, whichever comes first, then you can stop at 10, even if you haven't had it. Now, what is data saturation? It's not a science. It's not, oh, if I reach here, it's automatic data saturation. It's, am I hearing the same answers over and over again? Have I gotten to, you know, 2 people in a row or 2 out of the last 3 have given me the same answers? I always suggest go one more. And if you still hear the same answers, you most likely have data saturation. Now, if you've told your committee, hey, I'm gonna interview 12 people, or data saturation, whichever has come first, and the first 2 people you interview give you similar answers, they're probably not gonna accept. But if you get to 7 or 8 and you've got similar answers all along, it's gonna be much more likely that they're gonna say, okay, you've made it, go ahead and start analyzing. If you got the first 2 people and they got same answers, the committee's not gonna be satisfied because You said you were gonna do 12 and there's 10 more opportunities for different answers. But if you've gotten 8 or 9 and everything is being similar, there's the odds on getting a lot of new information in the last 1 or 2 or whatever participants is much diminished. So there is a reasonability test here, but it's basically, again, not a science, but it's, yeah, I keep hearing the same thing over and over again and I'm close to the number I was gonna do anyways. Therefore, we'll stop there. Now, one of the other areas that comes in various forms is, hey, I've interviewed these people. I want to make sure I'm not putting any bias in and/or if I'm correctly, you know, representing what they said. If you are recording their interviews and then transcribing them, the opportunity for bias diminishes as opposed to just taking quick hand notes and trying to fill in from there. It can be easy to institute some minor biases. But another technique that's used is what a lot of people call member checking. That means, hey, I recorded this, I transcribed it, I'm going to send it back to each participant to see if this fairly represents what they said and allow them to make any corrections. So this way you're making sure that the data is as clean as possible before any analysis, you know, happens. Anytime we have people though doing the analysis, we can all come with own biases and we might get slightly different results coming out of it. Just know that if you are as clean as possible, trying to keep as much bias out, that should satisfy your committee that you're doing the research without bringing in bias or other things that can impact the results. Finally, and I've heard this one before, is, look, I'm having difficulty presenting what I found. Chapter 4 is, this is what I found. Here are the themes that came about, the codes and the themes. One of the things I see people and its variations of this is, here's a theme 1, and they say what the theme is, and then they go on to theme 2, but they don't provide support for theme 1. It's very common for people to, participant 1, or however they've pseudonym named the people, indicated this, participant 2 indicated this. You typically wanna have at least 2 participants' quotes to support a theme. And it gets a little strange if you keep seeing participant 3 in every single one of the themes. Dig deeper, try to find out if participant 7 has a similar quote or something. So you're not just relying on one person, and that's going to help you present your findings better. And you always go back to Chapter 3. What did I say I was going to do? Then do what you said you were going to do in Chapter 4. Qualitative research, great form of research if it matches your research questions and your problem and purpose, and it's the appropriate way to go forward. Great. Just think about things ahead of time so that when you get there, your data is good, it's clean, and you're ready to go. I'm open to questions. I look forward to interacting with any and all of you. I hope you have a wonderful and beautiful day and keep making progress on your dissertation. Bye-bye.
