The strongest user research combines mixed methods. Research with more than one type of method adds an extra layer of investigation to your project, in order to answer multiple questions at once. It’s convenient, more in-depth, and will potentially save you money by preventing errors and wrong assumptions.
How to conduct effective mixed methods research
Think of mixed method research as the intersection between quantitative and qualitative — when human thoughts and feelings merge successfully with statistically relevant data to give a clearer picture of what you’re trying to find.
But what’s the best way to take the multi-faceted approach of mixed methods research?
Here’s everything you need to know, including how triangulation in mixed methods research can super-strengthen your conclusions.
What is mixed methods research?
Put simply, mixed methods research is the use of two types of research – specifically the combination of quantitative and qualitative research to get the best of both worlds. Think of it as a blended hybrid form of user research, allowing you to find results to a question from two different data sets.
Let’s imagine we’re researching how customer service might influence people’s online shopping behaviour. Mixed methods research could include a one-to-one interview to understand customer attitudes and perceptions (which is qualitative) and the use of a customer satisfaction score (the quantitative, numerical side of things).
Or say we want to find out why users are dropping out of the services page of a website. It could be that the listing isn’t what they need, or that the page has a wrong/misleading link. In order to answer this question, a researcher could run one-to-one usability tests (asking the user to perform tasks and describe their experience) before delving into web analytics to understand the problem.
Qualitative vs. quantitative…
The mixed method approach combines both types of quantitative and qualitative data, with one compensating for the weaknesses of the other when it comes to your analysis (known as triangulation…but we’ll get to that in a bit).
For example, if some participants answer your qualitative interview questions half-heartedly, we could then choose to combine the findings with a survey to test them.
Before we go any further, it would be useful just to define qualitative and quantitative research…
Usually in the form of conversational-based data.
Helps researchers understand people’s feelings, motivations, expectations, and pain points when experiencing a product or service.
A flexible mix of methods including one-to-one interviews, focus groups, ethnographic studies, and contextual inquiries.
Derives numerical findings from aggregated data and statistics.
Creates an organised analysis of a question or issue with measurable answers.
When used in conjunction with qualitative research, it can create a set of holistic findings.
Challenges and benefits of mixed methods research
The mixed methods research design works marvellously for blending questions, approaches, and different tools together to gain deeper insights. But that doesn’t mean it’s without its setbacks.
The benefits of mixed methods research are pretty self-explanatory, but let’s weigh up the challenges too.
Obtain richer, more comprehensive data to expand your evidence and improve the validity of your findings.
Generate new insights and thought processes to potentially help you with problem-solving other projects across the board.
More time-consuming and requires specialist skills in both qualitative and quantitative methodologies.
Often more costly due to the to-dos of recruiting people and facilitating/moderating the tests.
Conflicting and unwanted results might happen, with quantitative data telling you one thing and qualitative data telling you another (which means you run the risk of creating more work for yourself to get the results you can trust).
When to use a mixed method approach?
As with all user research, it depends on the research objective at hand…
If you can answer the research question sufficiently with either type of data, we recommend saving yourself some time (and money) by just opting for one because there’s ultimately no need to use the mixed method approach. For example, if you simply need to decide between design options, a quantitative preference test with a few follow-up questions would suffice. There’d be no reason to dive into an in-depth one-to-one interview.
A mixed methods research design should be used when a decision needs to be made based on both qualitative and quantitative data. Quantitative data finds a problem, then qualitative data explores the problem more in-depth – validating the numeral findings (or the other way around, but more on that later).
In other words, after you’ve obtained a data set from users that says one certain thing, you then carry out interviews, focus groups, and conversations to gain insights into the reasons why.
Let’s go back to our online shopping and customer service example…if we’re only interested in customer satisfaction levels, quantifying the results with an NPS score for example would definitely do. However, if we want to know the reasons why, i.e. what specific factors influenced their customer experience, then qualitative data can find that out. Makes sense, right?
Carrying out mixed methods research
Although there’s no clear-cut approach to mixed methods research, there are two core sequential types:
Explanatory sequential design
In a nutshell: you collect quantitative data and follow it up with qualitative data to contextualise and explain the data (hence the name ‘explanatory’).
You might collect data that indicates a low customer satisfaction score, and then try to understand why by conducting focus groups or checking out what reviews/feedback has been left.
Exploratory sequential design
This is basically the opposite way around. In a nutshell: you collect the qualitative data to develop an initial understanding and then follow it up with quantitative data to measure your findings numerically.
We might develop a hypothesis about the feedback customers are providing about their customer experience. And then you can carry out a quantitative study based on your findings to confirm/reject them.
Analysis and triangulation in mixed methods research
Although you’ll collect results separately, what’s most important is that you need to find the correlations and connections between the two sets of data. And remember, it may differ depending on whether you’ve carried out exploratory or explanatory research.
For instance, we carried out interviews with employees to understand what makes them happy at work. Once we collected the qualitative data, we then carried out a thematic analysis to identify the common patterns (or themes) between them.
Let’s say you discover that closeness with your team is a common factor in workplace happiness. Afterwards, you can carry out a quantitative study (perhaps a survey) with a question like: ‘In order of importance, how do these workplace factors affect your happiness?’. From these results, you can then validate or invalidate your finding with quantitative data to back it up.
There’s an official term for this too, triangulation…
Triangulation in mixed methods research is using different sources of data to arrive at the same conclusion. Essentially it’s a way of enhancing the validity and reliability of the findings by compensating the weaknesses of your methods with another. Minimising the risk of research biases.
Say you carry out focus groups, one concern might be that the generalised findings won’t represent the population and are only relevant to the individuals in the focus group itself. To counterbalance this, you can triangulate the findings by using a larger sample size in quantitative research and identifying whether the original findings are still present. If they are, then bingo — they’re valid and you were right all along!
The mixed methods research design is simple. It’s about getting to the very heart of a complex question or problem…because you’re using more than one type of research to find a solution.
Think of mixed method research as the intersection between quantitative and qualitative: when human thoughts and feelings merge successfully with numerical data to give a clearer picture of what you’re trying to find.