Five Elements of U/X and Research Design

What can researchers and scientists learn from U/X designers? Are the basic principles of User Experience design the same as the basic principles of designing a research project?

Consider the story of two (fictitious) experts in the field of U/X design and engineering research.

Amiyah Vu is a 32-year-old full-stack web developer. She has been working for a regional health authority for 7 years, championing the implementation of their new electronic health record. She has been very successful, and her health record implementation is currently in use by dozens of hospitals and health clinics.  It is early 2021 and COVID-19 vaccinations are being distributed to priority groups in her region. The initial rollout of vaccinations was handled with a telephone roster - and was a disaster.  The region lacked the ability to change appointments as vaccine supply shifted. Staff at the call centers were constantly overwhelmed with calls. On a Friday afternoon at 15:00, the chief medical information officer calls her, and in a desperate voice asks, "can you build us a website for the vaccinations?"

Stephen Simpson is a chemical engineer. He has been working  for Suprevia Petrochemicals for 12 years, and is now the lead engineer. His team has been developing several iterations of a novel spray-on coating that has been designed to reduce the rate of corrosion of industrial piping while being more durable than existing coatings. However, retooling the existing facilities at Suprevia will be expensive. Previous retooling for the last chemical innovation nearly drained the company of its finances, and in the end the product had been unsuccessful. The CEO and strategic management board are asking Stephan to provide robust evidence before retooling. Stephen receives an urgent email, asking him to present a research protocol to the shareholders at tomorrow's meeting.

Clearly both of our protagonists are in challenging situations. Both will need to develop solutions from the ground up. Where can they start?

1. Strategy

As described by Jesse James Garret, strategy is the starting point for any great user experience. Strategy answers the question of why to build the website at all. Why should it exist? Who will be willing to use it? What are the business objectives? This involves two parts: site objectives, and user needs. The site objectives are the business, creative, or internally created goals for the site. These are best assessed by careful consultation with the company stakeholders. User needs, conversely, are best discovered by careful customer discovery.

For Amiyah to build an effective website for vaccinations, she will need to start with asking the health authority what the goals of the website are. Is it education? Appointment booking? Is the website supposed to save money for the health authority?  In addition, she will need to know about the user needs. Why would the users go to the site? What are they expecting to find? What are they expecting to do? Amiyah knows that she cannot continue her project until these questions are answered. She meets with the chief information officer and the CEO of the health authority. They confirm that the site objectives are to provide an effective way to manage vaccine delivery to ensure optimal vaccination rates and avoid vaccine waste. She meets with several small improvised focus groups to assess the user needs. The focus group show that the primary user need is to book an appointment quickly online and be able to change it easily. Amiyah feels she is ready to move on the next step in her project.

When designing a research project, thinking first about strategy is also important. I divide this into two questions that I ask myself and the team on any project I do. First: "What are you trying to prove?" I find this to be the most valuable question a researcher can ask. Much like the site objectives for U/X design, without a clear answer to this question it is nearly impossible to create a coherent research study. The second question is "who cares?" Similar to the user needs section in the U/X design, it is critical to define in advance who will benefit from the research.

To create a sound research study to evaluate the new petrochemical coating, Christopher will need additional information from the CEO. What are they trying to prove? That the new additive gives industrial pipe a longer lifespan? That the additive saves money? The coating  can replace a more toxic additive? Christopher also needs to know the answer to "who cares?" Who is the target audience? Is it the shareholders? The customers? The environmental protection agency? Stephen meets with the CEO and the chair of the strategic board. They confirm that the primary purpose of the study is to prove that the new additive will increase the effective lifespan of industrial piping. They would like to create a report to present to the shareholders - proving the additive is effective and encouraging them to retool the factory for production. Stephen feels he is adequately informed to move on to the next step of the project design: the scope layer.

2. Scope

For U/X design, the scope layer builds on the strategy layer below it. Scope includes both the functional specifications and the content requirements. Functional specifications are the comprehensive list of all the features of the website. These specifications include all the features that must be present to make the site meet the business objectives and also the user needs. Content requirements include all the ingredients needed to provide the information users will use to complete their tasks. This may include text, audio, video, external content, or database records.

When Amiyah is designing her vaccination website, she will build on the strategy layer to develop the scope. She will need input on the functional expectations. What will the users want to do on the site? Book an appointment? Cancel an appointment? What content will they expect? Instructions on how to get to the vaccination center? A live chat function to speak with a public health nurse? A downloadable information sheet about the vaccine?  It is pointless for her to go ahead with the next part of the design until these questions are answered. Amiyah forms a small working group to discuss the scope. They decide that the site will initially contain only the functions of booking and cancelling an appointment. The site will also feature links to already published content from the World Health Organization on vaccine safety.

Carefully considering the scope of research projects is equally important. Researchers often use the "PICO" mnemonic to consider scope:

• P - Population: Who is the subject of the research? This could be a patient group (such as gender, age, or medical condition) or the experimental unit such as an app-user,  lab mouse, or steel beam.
• I - Intervention: What does the researcher plan on doing during the experiment? This could be a pharmaceutical treatment, psychological intervention, an app interface, or changing the shape of a steel beam.
• C - Comparison: What will the experimental group be compared to? Examples could be placebo, another pharmaceutical treatment, a different app interface, or a different shape of steel beam.
• O - Outcome: What will the researcher measure? Depending on the experiment, a researcher might measure blood pressure, weight loss, time to perform a transaction on an app, or strength of a steel  beam.

Setting down the PICO is vital to the next steps, And, should take the strategy into account to ensure that the choice of scope meets the goals of the strategy.

For Stephen, he will want to set down the PICO for his additive experiment. Stephen decides that for his experiment the population will be commercial industrial piping used for carrying hot water in warehouse buildings. The intervention will be coating of the pipes with his experimental additive. The comparison group will be similar pipes without the additive treatment. Finally, for the outcome, Stephen chooses the results of the GMW14872 cyclic corrosion test. Again he reviews his research methodology with the CEO to ensure that the scope aligns with the strategy before moving on to the next layer: structure.

3. Structure

When speaking of U/X design, structure includes the description of how the site will behave. It includes interaction design and information architecture. Interaction design specifies how the user will interact with the site as far as task flow and functionality. Information architecture describes how the information will be arranged structurally. This can include at the front-end (what the user sees) and in the back-end (database organization).  Referring back to the scope at this time is mandatory. The best structure will be the one that most easily allows the user to perform the tasks of the functional specifications and the find the content of the content requirements.

The next step for Amiyah in her vaccine app will be to carefully describe how users will interact with her site. Often this can be done by sketching out a typical user workflow on a whiteboard. This can involve fictitious "personae" of typical web-users, or can involve getting actual users to describe or demonstrate how they would use the site.  Having a good user experience means making the structure of the website intuitive and predictable. Amiyah decides to interview a few actual users to see how they would want to interact with the site. She quickly discovers that many of the users are elderly, and many are not daily web users. She realizes the that website must have a basic interface, and that the users are strongly favoring a simple workflow to book an appointment that involves only a few clicks. They also want to be able to easily go back to the site to confirm their appointment without having to create a user account.

For research, the structure layer outlines the overall methods of the project. This involves setting down the null and alternative hypotheses, choosing a study design, and choosing the test statistic. Researchers will need to consider multiple factors in choosing this design. Foremost, research is guided by the evidence-based pyramid that clearly outlines that all study designs are not created equal:  well-designed randomized controlled trials provide superior evidence. However, the level of evidence must be balanced against practicality, ethics, and cost. Setting the study structure is the most technically challenging aspect of research project design. Choosing the wrong study structure can be a disaster: either by failing to meet the needs of the strategy and scope below, or by ruining the skeleton and surface above. Researchers should always seek assistance from a statistician if there are any doubts in their design. Always seek assistance in high-stake, expensive, or important research.

Stephen enlists the help of a statistician to design the structure of his experiment. They decide on a simple parallel groups randomized design. The null hypothesis of no difference between pipes treated with the additive and those that are not treated will be tested against the alternative hypothesis that the pipes are different. As the results of the GMW14872 are a decimal number, they decide to use a two-sample t-test as the test statistic. A sample size calculation is performed by the statistician in consultation with an expert in corrosion resistance. They calculate a sample size of 40 pipes will be sufficient. Twenty pipes will be randomly sampled from a large shipment and chosen for the additive group. Twenty pipes will be randomly chosen for the control group. Only now can Stephen go on to the next layer: Skeleton.

4. Skeleton

In U/X design, the skeleton layer includes the visual form, presentation, and arrangement of the site. This includes three parts: interface design, navigation design, and information design. Interface design - what we commonly consider the UI (user interface) - is the visual arrangement of the interface on the screen, including the location of the elements, buttons, links, and images. Navigation design is the arrangement of the interface elements in a manner that is intuitive and functional for the users. Finally, information design is the visual arrangement of the information on the site. This includes where the information is located, how the user accesses it, and how the information is displayed. A very efficient way to develop the skeleton is through wireframes. The designers look back to the structure layer, and, taking into consideration the interaction design, draw a black-and-white schematic of how the site works. Usually, the wireframes include placeholder images and unformatted text - the goal is to demonstrate the usability of the site without being distracted by the cosmetics.

To design the vaccination app's skeleton layer, Amiyah again looks to her colleagues and several test users. They sit together as a group in front of a whiteboard and start to draw out in block form what the site should look like. They decide that users should immediately see a welcome page with two big buttons: book an appointment, and view/edit my appointment. Likewise, they also want a toolbar at the bottom with links to further information, the telephone hotline, and the emergency information number. Amiyah draws up wireframes, and presents them to the Chief Information Officer and CEO. They review the wireframes together to ensure that all the requirements specified in the structure level are indeed met  by the skeleton.

For researchers, the skeleton is the basic outline of the research report. Is this an academic report? If so, the researcher should consult the author's instructions in the target journal for guidance. Is this an informal or internal report? Perhaps the six-sigma format with its explicit executive summary would be superior. What plots or figures will be needed? The structure of the project should be carefully considered ensuring that the skeleton comes naturally as an extension of the research design. For instance, decisions on the type of plots are largely dependent upon how the outcome was measured in the structure layer. Then too, specific designs may mandate specific plots: regression diagnostic plots for linear regression for example.

Stephen decides that since the pipe coating experiment is designed for internal use only, there is little need to develop a formal academic report. He decides to base the skeleton of the research on the standard six-sigma report design. He will include a table describing each of the two experimental groups to confirm that no important differences between the pipes in the two groups existed at the time they were sampled. Furthermore, he will also display the results of the outcome as box plots. A short slide presentation will be developed based on the six-sigma report. With this skeleton in mind, he is ready to move on to the surface layer.

5. Surface

Surface, in U/X terms, is the visual design that the user actually sees. This includes colors, fonts, images, sounds, videos, and branding. The graphic design of the site. What makes a good graphical design? Looking back to layers below is the best way to determine the surface design.

Amiyah admits that graphical design has never been her strongest point. She elects to get help from a local firm that has a great reputation in creating beautiful and effective web designs for many regional businesses. They meet to review the strategy, scope, and structure. Together, they go through each wireframe and decide on the best graphical layout to make the site accessible and effective. For instance, as they know that the site will be used by many elderly users, they elect for a clear, large, readable font. Also, they choose graphics and color in line with the specifications of the regional health authority: both because this is a requirement specified by the health authority, and, to provide confidence for the users that they are indeed on the correct site and can trust the interface with their personal information. The design firm draws up a branding guideline document for the site, and Amiyah reviews it with the Chief information officer before proceeding.

For researchers, the surface layer encloses the often tiresome tasks of editing. This includes choosing the correct point of view for the discussion - traditionally the third person past passive has been used, although some writers and journals prefer the second-person plural, such as, "we treated 48 patients with placebo." The tone and mood are also important. Academic journals requiring a far more formal tone than field reports for example. Checking, rechecking, and then re-rechecking of grammar is important. So is language structure. The surface layer is the one that the readers will see, and errors here hurt the validity of the findings. This can be especially difficult to writers who are writing in their second (or third, or fourth) languages - getting a native language speaker to review your text is always worthwhile. Finally, the surface layer involves formatting the report for the settings. For example, academic journals will often have firm specification of what level of subtitles can be used, how references must be formatted, and how the title page should appear.

The surface layer of Stephen's pipe experiment begins with Stephen writing a first draft of his report and distributing it to the rest of the team. They work in an iterative fashion to build a clear and concise report. They recheck the grammar and spelling by using an automated grammar checker. Looking back on the skeleton, structure, scope, and strategy layers below helps guide them on the most appropriate presentation. For instance, since they know that the shareholders of the company may want to read the report, they ensure that technical language is appropriate for these readers. They have a colleague from the structural engineering department, who was not part of the study group, read the paper to ensure the results are clear to non-experts in chemical engineering.

Putting the Five Layers into Action

The five layers of U/X design has become a standard method for developing and visualizing effective user experience for over a decade. In our example, it was easy to see how Amiyah developed her vaccination website stepping layer-by-layer through strategy, scope, structure, skeleton, and surface. Each layer was developed with stakeholder engagement to ensure that sufficient information was clarified before proceeding to the next layer.

Researchers could benefit greatly by following a similar five-layer design for their projects. Again, stepping through strategy, scope, structure, skeleton, and surface in an orderly fashion, and ensuring that the end product answers the two most important research questions: "what am I trying to prove?" and "who cares?"

Unfortunately, many researchers fail to carefully develop their designs before proceeding to the data collection phase. In many ways, for research projects careful and deliberate development is even more important than it is for software design. Often software designers benefit from a fluid and agile development cycle where changes can be made easily after the initial implementation. This is usually not true for research experiments. When designing an experiment, decisions made at each layer are often irreversible. Mistakes at any layer can be costly. As a statistician, I have been in the unfortunate situation of seeing far too many researchers discover errors far too late. In many cases, following the five layer procedure for design could have saved them time and money and have led to a far more convincing result.

Are you ready to put the five layers of research design into action for your next project? Get the infographic summary. to get started. And, please, tell us about your experience here on our blog or on any of our social media channels.