4. Basic Mixed Methods Research Designs | John W. Creswell | University of Michigan

Summary

Highlights

• John W. Creswell introduces mixed methods research designs for a comprehensive understanding. 🎓
• Learn about convergent design, which combines qualitative and quantitative data for thorough analysis. 🔄
• The explanatory sequential design helps clarify and explain quantitative results through qualitative research. 💡
• Explore how exploratory sequential design starts qualitatively to develop robust quantitative tools. 🛠️
• Gain insights into integrating different data forms for richer research outcomes. 🔗

Key Takeaways

• Discover the three fundamental mixed methods designs: convergent, explanatory sequential, and exploratory sequential. 🌟
• Convergent design merges two databases to capture different perspectives. 🌀
• Explanatory sequential design uses qualitative data to explain quantitative findings. 🔍
• Exploratory sequential design begins with qualitative data to build quantitative instruments. 🔧
• Understanding these core designs improves research flexibility and depth. 📚

Overview

Chapters

• 00:00 - 00:31: Introduction to Mixed Methods Research Designs The chapter introduces the concept of mixed methods research designs, presented by John Creswell from the University of Nebraska-Lincoln. Creswell emphasizes the diversity of mixed methods designs, noting the numerous types and classifications available.
• 00:31 - 01:42: Basic Designs in Mixed Methods Research The chapter focuses on introducing basic designs in mixed methods research. It emphasizes the simplicity of these designs, urging researchers to identify the underlying basic design of their study, which can evolve during the research process. The chapter offers guidance on how to determine these basic designs, acknowledging that they are typically emerging rather than fixed.
• 01:43 - 05:56: Convergent Design In 'Convergent Design', the focus is on altering a design, but the process remains consistent within its framework. The chapter delves into the methodology backing each design choice, highlighting 'validity threads' that scrutinize the design's execution. The concept of 'convergent design' is emphasized, suggesting a deliberate approach where various elements come together purposefully. This systematic approach aims to ensure a robust foundational structure that enhances the design's overall validity.
• 05:57 - 17:55: Explanatory Sequential Design The chapter explains the concept of explanatory sequential design, where two databases are merged—one qualitative and one quantitative. The process involves sequentially connecting these databases, allowing the second, quantitative database to explain the findings of the first, qualitative database. This method is contrasted with exploratory sequential design, which starts with qualitative data and builds upon it with quantitative data.
• 17:56 - 26:16: Exploratory Sequential Design The chapter discusses the exploratory sequential design in research methodologies. It begins with an explanation of the convergent design, which involves collecting and analyzing both qualitative and quantitative data separately before merging the findings for interpretation. The goal of the design is to capture multiple perspectives on a problem using different types of data, such as survey or rating forms.
• 26:17 - 27:15: Conclusion The conclusion chapter outlines the methodology used in the research, detailing a mixed-methods approach. This approach involves a questionnaire for quantitative data collection and individual interviews for qualitative insights. The chapter emphasizes the importance of integrating these two perspectives for a comprehensive analysis. Visual representations of the research design can vary, being either vertical or horizontal, but this variation doesn't affect the overall research process. The chapter concludes with an overview of steps involved in conducting this methodological conversion, highlighting how qualitative and quantitative data complement each other to enrich the research findings.

4. Basic Mixed Methods Research Designs | John W. Creswell | University of Michigan Transcription

• 00:00 - 00:30 good afternoon this is john creswell from the university of nebraska lincoln i'm back again to talk about mixed methods research in this video series that i put together today i want to visit with you a little bit about basic mixed methods research designs the first place to start is that there are many different types of mixed methods designs out there and there are many classifications the thing about doing a
• 00:30 - 01:00 design using the design and mix methods is to keep it very simple and what i'm going to focus in on what i call the basic designs at the heart of every mixed method study is a very simple basic design and your task as a researcher is to figure out what that design is and i can help you through this slide set we also know the designs are emerging often emerging rather than fixed so that during a project we might
• 01:00 - 01:30 change the design and that's certainly fine also within each design basic design there are certain steps i'll be going through i'm going to show you some diagrams of what the design looks like and i'm going to talk also about what we call the validity threads or methodological issues and actually conducting these designs so the three basic designs are a convergent design and these names are very intentional convergent means we're going to be
• 01:30 - 02:00 merging two databases explanatory sequential meaning we're going to sequentially connect the qualitative and quantitative database and have the second database help explain the first explanatory and then exploratory which is just the reverse of explanatory we're going to start qualitatively and build something quantitatively through the second database so we call that an exploratory sequential design
• 02:00 - 02:30 so let's first start with this convergent design and this is a very simple graphic of how the design sets up we're going to collect qualitative data and analyze it and quantitative and analyze it and then compare the two results we're going to merge these two databases and then make an interpretation the whole intent of this design is to capture two different perspectives on a problem one based on a rating form or a survey or
• 02:30 - 03:00 questionnaire that people might fill out and the second on individual interviews and so those personal views are going to be compared with the more statistical quantitative views and we're going to bring those two views together now this design can be drawn vertically as in that last slide or more horizontally in this slide it doesn't make a lot of difference i've seen them drawn both ways the steps in conducting a conversion
• 03:00 - 03:30 design they're just five of them first we're going to collect both quantitative and qualitative data typically at roughly the same time we're going to analyze both databases independently and then we're going to bring the results together we're going to compare the quantitative and qualitative results now when we compare these two results we're then going to compare them to see if they the two pictures of our problem actually converge or they might diverge
• 03:30 - 04:00 now if they diverge we need to explain that divergence and so i'll go into some of the explanations as to how we might help to explain that divergence if it occurs i've drawn a kind of a process diagram here starting with a comparison question and then both forms of data collecting both forms of data and then merging it and then comparing the results and then
• 04:00 - 04:30 resolving discrepancies or convergence or divergence so this is this follows those steps i've just mentioned now i put something else into this diagram and that is that there are certain points that you need to pay attention to if you're using this this design these are potential validity threats in other words they might be points at which your results and interpretation may be an act inaccurate if you don't follow some of these procedures the
• 04:30 - 05:00 first one is i would set up parallel constructs variables or questions between the quantitative and qualities they need to be so if i'm asking uh open-ended questions qualitatively about self-esteem then i need to measure quantitatively self-esteem so so that's what i mean by parallel questions looking at the same concept both quantitative and qualitatively and the reason for that is if i'm going to merge these two databases i've got to
• 05:00 - 05:30 have some common ground for merging the two databases the second point here to pay attention to would be the sample size typically people gather more quantitative data than they and they have a larger end than the qualitative end but i've seen many convergent designs where people gather both the quantitative and qualitative data using the same sample size usually these results are presented
• 05:30 - 06:00 separately and one of the biggest challenges in using this design is how do you actually merge two databases how do you merge stories for example with numbers and so i'm going to talk about three strategies for merging that are part of this design these three uh i call the side by side comparison of the data two databases transforming data
• 06:00 - 06:30 and then creating a joint display so these are three techniques that have emerged when people are are comparing the quantitative and qualitative data in a convergent design this is a side-by-side comparison this is a passage right out of an article by classen looking at older driver safety and when you go into the results of this article you see that they've arrayed the findings such as the finding on previous motor vehicle
• 06:30 - 07:00 convictions first of all they talk about the quantitative results on previous motor vehicle convictions and then they talk about the qualitative results and then they compare the two so they are actually putting side by side in a results section typically this is found in the discussion section of both the quantitative and qualitative results so you can actually see if they converge or diverge
• 07:00 - 07:30 a second is to transform data now this is a study where the the individuals gathered quantitative and qualitative data and analyzed both forms of data but they took the qualitative results and transformed it into numbers they counted how many themes or you could count how many codes and so then that quantitative data is merged into the other quantitative database and
• 07:30 - 08:00 so the data are the qualitative data are transformed into quantitative forms of data so that's called data transformation this is a newer area to emerge in convergent designs it's called a joint display and what you do in this case is come up with the table such as i'm showing you here where you're arraying on different topics in your study the qualitative results
• 08:00 - 08:30 typically themes and then the quantitative results in the statistical analysis and then you might have a third third column comparing those actual results so what i'm doing here is i am presenting in one table both the quantitative and qualitative data results so that i can compare the two so that's the third approach well the issue often arises in a convergent design what if the two
• 08:30 - 09:00 databases diverge or they're contradictory or they're very different perspectives begin to emerge how does a researcher go about addressing this well here are some of the strategies that have emerged in the literature first of all researchers often go back to the original databases and re-examine them they may have missed some themes they may have not looked at some of the quantitative results accurately
• 09:00 - 09:30 another strategy is to go out and gather more data to resolve that discrepancy that can be expensive it's maybe not is used as frequently as it perhaps should be often people go back to to look at those questions to see if they're parallel between the quantitative and qualitative database now strategies that are probably not recommended would be at the bottom of my list here one is some researchers side with either the
• 09:30 - 10:00 quantitative or the qualitative data and they give it more value and they say perhaps the quality really tells the true story or perhaps the reading skills on this instrument tell us a better story often doctoral students will look at the discrepancy and state it as a limitation in their project it's not a practice i'd recommend for publication but you do see this in convergent designs when you choose this type of design
• 10:00 - 10:30 often people use a convergent design when they're out in the field and they've got to collect both forms of data at roughly the same time so it's an efficient efficient design the basic idea is that you're trying to merge these two forms of data for to develop a more complete understanding of your problem so if you feel that your problem can best be explained by both quantitative and qualitative data this is a good design
• 10:30 - 11:00 and people use this when they have a good understanding of some of these merging strategies such as a joint display ways in which they want to portray the data side by side how do you actually spot a convergent design in a journal article well you start by looking to see if the authors called it a conversion design at the beginning but then you look at the intent of the study are they actually trying to merge the two databases to have a more
• 11:00 - 11:30 complete understanding or a better understanding of their problem you can then look into the results section or the discussion section and see if they brought the both both of the databases together they actually merge the two databases so those techniques you can easily spot a convergent design i would have to say because of the challenge of trying to merge the data setting up the project properly with parallel questioning
• 11:30 - 12:00 the convergent design is a fairly challenging design to use and often people come to it first and mix methods because they think that the only way i can do mixed methods research is actually to merge the two databases there's an easier way here's the second design explanatory sequential design and i think this is a design that's probably the most popular among doctoral students graduate students across the united states at least
• 12:00 - 12:30 i call this a two-phase design where they're first collecting quantitative data such as doing a survey project and then following up qualitatively to explain those quantitative results in more detail when you actually talk to people in the second phase qualitatively you can understand your survey results a little bit better so the intent of this design is to use the quantitative data to help explain uh use the qualitative data to help explain the quantitative results
• 12:30 - 13:00 so the steps are fairly straightforward here the first phase is collect quantitative and and then analyze it and then look closely at those results to determine what needs to be explained further are there surprising explanations are there extreme cases that need to be looked at by their significant predictors then you go out and collect the qualitative data to help explain some of those quantitative results you analyze
• 13:00 - 13:30 it and then end this project by talking how about how the qualitative data helps to explain the quantitative results well the diagram of this might look something like this in this picture i have in the the large boxes the major steps first quantitative data collection analysis then qualitative and then some interpretation and then i've also listed given some examples of procedures at each stage
• 13:30 - 14:00 and then some products that a researcher might put together well here's another diagram kind of the step-by-step process where you're starting with a question you're starting with quality quantitative research questions hypotheses collecting data analyzing and moving on to qualitative what i did in this diagram was i put in some of the challenges the validity
• 14:00 - 14:30 threats that are likely to arise so one challenge is once you've got the quantitative data analyzed you have your statistical results what results do you follow up on and thinking through the possibilities of what results need further explanation that's a challenge and then also what participants qualitative participants can help you provide more information about those results help
• 14:30 - 15:00 explain those results so the sample size is another issue in this design if you're going to follow up qualitatively to explain a quantitative database you need to choose qualitative participants in phase two that are from the same sample that filled out the instruments in phase one so you have to pay attention to sample
• 15:00 - 15:30 size here and then also at the interpretation phase interpreting the quantitative results based on your qualitative data is definitely a step that needs to be taken it's not always taken in these designs this challenge of deciding what quantitative results to follow up on or to build on in your qualitative phase you could look at demographics you could
• 15:30 - 16:00 look at outliers extreme cases important significant results maybe insignificant results that you thought might have been important as you did your quantitative analysis you could follow up qualitatively in these four areas you can also set up a joint display where you can array your quantitative and qualitative data together for example in this example the quantitative results show no
• 16:00 - 16:30 significant difference but the qualitative follow-up in the next column did show that there were some significant points that were being made this is a study about caregiver givers and then the third column over addresses how the qualitative findings help to explain the quantitative results in more detail so what i've done here is i've arrayed in a single table both quantitative and qualitative results
• 16:30 - 17:00 and i've shown specifically how the quality helps to explain the quantity of data when you choose this type of design well you can see if it's in two phases quantitative followed by qualitative it takes a long time so you've got to have the time it's not as efficient as a convergent design where you gather both forms of data when you're out to the field once people that use this design are often those that are very quantitative in their orientation because the entire
• 17:00 - 17:30 project the phase one of this design starts quantitatively so you're really building on your quantitative perspective that emerges through the first phase i think people also use this because it's a very elegant simple design it makes simply make sense that if you collect quantitative data it's helpful to follow up qualitatively to explain some of those statistical results and so it's seen as a fairly
• 17:30 - 18:00 straightforward elegant design how do you spot one of these designs in a journal article well of course the first point is you see if the authors called it an explanatory sequential design assuming that they're familiar with some of these aspects of this design we've talked about but i would look to see if there's qualitative data that actually is tied to the quantitative results you know helps to explain the quantitative
• 18:00 - 18:30 results do the authors make that explicit and typically the qualitative should follow the quantitative in any discussion so look in terms of whether the results are built sequentially from the quantitative to the quality over time now what i'm going to do is reverse these two phases of the explanatory sequential design and rather than starting quantitatively
• 18:30 - 19:00 i'm going to start with qualitative data collection and this actually is a design that i call a three-phase design i collect qualitative data and analyze it and then i build something quantitatively and then i test out this quantitative instrument or intervention or typology whatever i built in my second phase i it out quantitatively so you can see now we've moved from a convergent design which is a single
• 19:00 - 19:30 phase to an explanatory design which is a two-phase to now an exploratory design which is three-phase so we're building into a more rigorous type of mixed methods project here the intent is to use this design when you first need to explore so the overall intent of an exploratory sequential design is really to explore first to build a better quantitative phase the steps in this design are as follows
• 19:30 - 20:00 first collecting quantitative qualitative data analyzing it designing then something quantitatively a quantitative strand based on what is learned from the qualitative results for example it could be to develop a new instrument because there's not an existing instrument that exists that is good to study this particular sample that you're looking at another might be to modify an existing instrument to develop an intervention that might
• 20:00 - 20:30 actually work with a group of people another reason would be to develop some type of a classification typology taxonomy that then would be tested quantitatively so there's a number of quantitative types of products that could be developed in this second phase and then in the final stage collecting the quantitative data to test out this insta i'll use an instrument here for an example as an example and then analyze the data
• 20:30 - 21:00 and then explain how the quantitative results really help to understand not only the qualitative themes at the beginning but also provide a better instrument a more useful intervention approach a more useful typology so again i've drawn a picture here of the process of research starting with a mixed methods question gathering data qualitatively moving to the second phase of developing
• 21:00 - 21:30 something in this picture i talk about developing a good psychometric instrument and then phase three where you're actually testing out that instrument and administering it quantitatively so those are the three phases now in this picture i've also put some of the validity threats or the challenges using this design one of the key challenges is deciding what qualitative results to use and how
• 21:30 - 22:00 to use them for example if you've gathered qualitative data and you have qualitative themes for your analysis how do those themes then relate to developing an instrument or what qualitative themes do you look at to help develop this quantitative phase also if you are developing an instrument good skill psychometric scale construction is an important issue so now what we have is to use this design
• 22:00 - 22:30 you need to know about qualitative research quantitative research measurement how to develop a good instrument as well as mixed methods research so you can see there's a number of skills that are required to really utilize effectively this design and then towards the end of this design explaining how the quantitative really helps to build a better understanding from the qualitative results
• 22:30 - 23:00 here's a picture i put together of this type of a design that starts qualitatively you analyze collect and analyze the data come up with results then develop an instrument and then test out the instrument so one of those challenges how best to move from a quality from qualitative themes to developing items that go on to an instrument here's a little
• 23:00 - 23:30 schema that might help out first of all you take quotes and turn them into specific questions or items in your instrument then the codes that you use from your qualitative analysis become your variables and the themes the broader perspective that you're developing from your qualitative analysis becomes the scales in the instrument so there's a way that you can move from your qualitative data into
• 23:30 - 24:00 developing an instrument and of course coming up with with good steps for constructing a psychometrically sound instrument is also a very rigorous process and there are books out there on scale development that help you build some of these steps but these might be some typical examples conceptualizing the constructs developing items based on the literature and the experts pilot testing them
• 24:00 - 24:30 gathering more extensive information on a larger sample and looking at exploratory factor analysis confirmatory factor analysis and then administering it to a larger sample yet where you check for reliability and validity so the steps involved in putting together a good psychometric instrument that would have good scores for validity and reliability is a very rigorous time consuming process so when do you choose this type of
• 24:30 - 25:00 rigorous design first of all you need at least time for two phases if not three phases here it's very attractive to qualitatively oriented scholars and students because you're starting qualitatively and then you build into the quantitative direction it's also helpful in designing instruments that might work with a special sample or population
• 25:00 - 25:30 where existing instruments are currently not available coming up with intervention procedures for an experiment where some of those procedures may not be known in advance or even coming up with typologies so there are some very specific applications of this and there there are some great mixed method studies in the literature on uh the exploratory sequential design that take you through very rigorous procedure these three phases of procedures
• 25:30 - 26:00 how do you spot this exploratory sequential design and journal article well you first look to see if authors are calling it an exploratory design next you might look to see if it starts qualitatively in other words it's a qualitative phase precede a quantitative phase and then you look to see in that quantity phase what's actually being developed as an instrument being developed our activities being developed for an intervention as a new typology
• 26:00 - 26:30 being developed you look for that quantitative component that's being developed and then whether it's tested out so i've gone through three types of designs the convergent explanatory sequential design and exploratory sequential design and if you look closely at every mixed method study within their design should be one of these three and then in some cases they build onto
• 26:30 - 27:00 these basic designs other features which i call my advanced mixed methods designs and in the next video i'll talk about some of these advanced designs thank you