Tips to Perform User Research on Marginalized Groups

For my group project, my team was developing a product to help the homeless. While this seems like a respectable goal, it soon became clear that we were facing one massive hurdle that threatened to undermine our entire development process - how could we effectively interview homeless people while minimizing any risk to the researcher, finding participants in an ethical manner with no incentives given to them and get answers that were relevant and meaningful and actually helped us better understand our user base?

From my experience doing this project, I came up with a list of tips to keep in mind that can very effectively increase the quality of the research and reduce the effort needed to come up with a group of pilot testers from your target audience.

1. Understand Who Your Userbase Is
   This may seem like a simple point, but it is crucial to have some understanding of your user base before actually starting the User Centered Design Process. When we first decided we wished to create interfaces to help the homeless, we neglected one very important point - this marginalized group is most likely going to be very difficult to conduct research on because this group of people are unlikely to have the patience or interest to sit down and do long, complicated interviews. We had to retroactively change what research script we used very late in the process to a questionnaire because we couldn't find willing participants who were willing to do interviews with us.
2. Know What Actions Can Alienate Them
   Since you are dealing with Marginalized Groups, it is important to remember that they are in the fringe of society and that there can often be psychological issues relating to some of the things they have to deal with. When we talked to one homeless person, we realized that he became very annoyed when we asked him about his finances, or when we asked him questions like 'what do the homeless do about .....?' It was clear that he was not doing well financially and it was frustrating for him to have to talk about this. Also, he was homeless and having to constantly hear this annoyed him. We realized we should rewrite our questions like 'What do you do about.....' instead of 'What do homeless do......'
3. Know Which Organizations Already Help Them
   If it is difficult to find user testers for your research, it may be helpful to contact organizations that understand this group and already have connections with them. They themselves can provide valuable information about this userbase and provide you access to their network of marginalized individuals.
   
4.  Keep the Safety of Researchers in Mind
   The unfortunate truth is that many marginalized groups are found in high-risk situations that can also endanger your researchers. One of our group members had previously done research with a marginalized group and she reported that one of her previous team members contracted a disease from not taking appropriate precautions when interacting with this marginalized group. It is important that we help these groups but it is absolutely crucial to not compromise the safety of your researchers. Take steps to minimize any risk that can come from conducting research.
   
5. Collaborate With Other Groups Performing Similar Research
   While this isn't always possible to do, our internet searching revealed many other groups that were doing somewhat similar research, While we didn't directly collaborate with them, we learned a lot about how the User Centred Design process applies to the homeless from reading their posts. If you are finding it very difficult to conduct research, consider teaming up with groups doing similar research and pool your resources - having to share data is less deal-breaking than having no data at all.

Microsoft's Vision of the Future

In 2009 (which was a whopping eight years ago), Microsoft released a video showing a vision of what they believed the future of Computer UIs would look like (possibly in 2020). I only recently came across this video, but to me, even now, nearly a decade later, some of the technologies seem completely highfalutin - things that are still in the realm of sci-fi and nowhere near ready for consumers.

The first thing that this reminded me of is a topic we discussed in class a couple of weeks ago - it takes decades for a tech idea to go from the furthest reaches of the mind's of researches or visionaries to the production chain to the pockets or desks of consumers (although with the current Internet-of-Things trend, tech goes further than just your pocket or your desk!) It takes quite a few more years for new technology to mature and become highly usable and functional. The first touchscreen phone came out in the '90s. Even now, after all these years, companies release new touch interactions that further enhance productivity and usability. While the layman may believe that technology changes quickly, the reality is that it takes years-and-years before a single compelling idea can become widespread in the consumer arena.

Now, going back to the video, I wondered - some of these things certainly look impressive, but how far away are we really before we have access to such things? It seems I am not the only person to have asked myself this question, because I found my answer in the comments

Roman Yoshioka provides the following breakdown of present-day approximations to some of the conceptual technologies seen in the video:

Just to give you all some hope, these are the things that are already possible: 
0:08 (Do you have a cat? - Real time translation): Skype translator  
0:38 (Shannon's stuff - Remote classroom): Google Classroom  
0:48 (Work Schedule - Digital calendar): Outlook, Google Calendar and every other calendar app 1:08 (Flight tracking) I can only recall Cortana and Google Assistant doing this. 
1:35 (Air gesture): Leap Motion  
2:09 (Holographic info): Microsoft HoloLens, though with a huge device on, and Google Glass 
1:44 (Foldable phones): [RUMOR] the Surface Phone 
2:55 (Screen hubs): Surface Hub 
3:28 (Voice assistants): Cortana, Google Assistant, Amazon Alexa, Hound by SoundHound, etc. 
3:51 (Fingerprint recognition): Has been around for quite a while 
4:29 (Smarthomes): Maybe a combination between Alexa and Surface Hub? Possible but not done 5:12 (Image recognition): Microsoft CaptionBot, Google Photos can do it too  
5:29 (Green roofs): Possible but barely implemented

In other words, it seems we are at that point in time where truly next-gen computer UI is in that experimental, still-developing stage just as touchscreen phones were still being refined in the 90's. When I first started taking this class - a single, worrying idea worked in the back of my head. Do we really need any new innovations in UI? The keyboard and mouse are decades and decades old yet they are still the preferred choice for interacting with PCs. Touchscreens are ideal for smartphones and tablets and it doesn't seem like there are any better alternatives in the horizon. What else do we really need? But then again, what if the first GUI designer asked why we needed GUIs when the command line works so well? Why do we need more RAM? As one tech visionary so foolishly asked. It is human nature to be complacent. But this video indicates that are still innumerable things that could become more functional had they sported better UIs. It is our job as scholars of UI to find those things, document them, experiment with them, work with them, refine them, create them. This 8-year-old video has given me the inspiration I needed to believe that there is still so much work to be done before humanity can truly stop and say yep, now our work is really done.

Bad Dpi Scaling: Bane of current day UX?

We live in a world where every company is forever in pursuit of increasing pixel density. There was a time when quad-HD (a resolution standard) was considered off limits to all screens except for very high end computer monitors used by content creation professionals. Now, any flagship phone released in the past 2-3 years will sport that particular resolution, with rumors of even higher resolutions in the pipeline.

In principle, it could be argued that this could possibly better the user experience. It is true that not being able to resolve individual pixels makes the experience of using a screen more pleasing for the end user. However, there is evidence that the human eye is unable to appreciate any improvements from 300 ppi onwards. This doesn't seem to have swayed companies from releasing ever more high density screens, however.

Take, for instance, this flagship monitor announced by Dell in CES 2017:

Now here is what the regular Windows 10 UI looks like:

Are there any differences that are immediately noticeable?

For one thing, the bottom screen seems to be consistent with Fitt's law. Icons are large, various menu elements are well separated, the text is easily resolvable by the human eye etc. The first picture, on the other hand, sports a UI that is tiny. Keep in mind that the monitor is 32 inches diagonally. So a person may be sitting a distance of 1.5-2.5 feet away from the monitor. Even from those distances, effetively manipulating the UI will become a massive hassle and will disrupt regular workflow.

The thing is, it doesn't have to be this way. There are many systems that incorporate perfect UI scaling such that low end, low res monitors and high end, high res monitors can both sport UIs that are optimal and are scaled consistently. There are algorithms that can granularly scale a UI instead of having to have a programmer "hard-wire" a couple of pre-set UI sizes for a couple of pre-set resolutions. I think this is one of those cases where we can appreciate a real world difference between 'coders' and 'computer science professionals' and 'hci professionals' because a coder unfortunately cannot appreciate the importance of UX in the same way someone familiar with the field can. They also may not be able to understand the idea that UI can inductively be scaled regardless of resolution, which is why they stick to "hard-wiring" only a couple of pre-set UI sizes.

I hope this disconnect between sub-optimal UI and ever-increasing resolution is addressed soon because otherwise, it is the end user who has to suffer.

First Post! Hello World!

Welcome to my first post on my CSC318 Blog!

I am a student of computer science in my final year of study. Over the past couple of years, I have learned various things that have deepened my knowledge on the mysterious black boxes that power so many facets of our lives in the 21st century. As I started to learn more and more, however, one thing became clear - much of what is learned in Computer Science revolves around the abstract. For example, as someone who has taken an advanced operating systems course, I can describe various esoteric algorithms on designing the scheduler of a kernel or how to prove the correctness of a concurrent algorithm on synchronizing threads or how many faulty aspects of a system can be reduced to an instance of the Byzantine General's and so forth.

But if I were asked to describe how the GUI in linux works, I would be hard pressed to go into any more detail than just stating some of the superficial aspects of the X windowing system. This has made me realise that I want to know more about GUIs - both the computations behind what make them possible as well as the more abstract and philosophical aspects of what constitutes good design. I want to be able to look at a user-oriented application and confidently identify whether or not it employs good design principals. I want to know more about the history and current state of the art of UX.

I am very confident that CSC318 will be a useful course as I enter the latter part of my degree. I hope this blog can bring some enjoyment to you, reader!

T.K.