This summer, I was hired as an undergraduate researcher for the “Peer Mentoring” project at the University of Washington’s Information School. What started as a valuable internship to supplement my studies in Human Centered Design and Engineering has evolved into a long term commitment to my growth as a user researcher. This year, I will be continuing my work conducting user studies and analyzing results. My early work and experiences on the project are captured in my internship logs below:  

 

Project Overview

This quarter I have carried out a summer research internship with the Information School at the University of Washington. Specifically, I have been working on a peer mentoring project, which aims to connect cancer patients on online communities. For the purposes of the project, a peer mentor is defined as a member of an online community, usually a patient or a caregiver, that other users could turn to for help or advice. As the project has continued, this definition has narrowed to peers who can play a mentorship role in regards to the cancer experience. For example, in a cancer community, a peer mentor might be able to tell a new user what to expect when dealing with cancer, explain various treatment options, give tips on managing side effects and help with any number of other things that affect everyday life.

Before I came on as part of the research team, the project had been underway for a few years. Primarily, this time was spent perfecting text analysis techniques for extracting content such as health interests from online health communities. A small design study was also conducted to determine what interaction styles and content densities were preferred by users when reading through digital profiles.

Upon my introduction, the research agenda at hand had shifted to determining exactly what types of information members of the cancer community deem most important when identifying and reaching out to mentors. I was hired to help create and conduct a user study that is intended to answer this question.

Overall, the experience has been an extremely positive one. I have both honed the skills that I have been learning in my HCDE courses and gained entirely new insights into the world of user research. For the remainder of this paper I will work to explain these valuable lessons and provide a deeper understanding of my work and accomplishments over the course of my internship.

 

My Role

While my formal job title for this position has remained “undergraduate researcher,” I found that I began to wear more than one hat as I gained my bearings.  Because my tasks varied, in turn I found myself playing the role of student, teacher, coworker, friend, researcher, programmer and human-centered designer. The nature of my work allowed me much freedom in how I wished to grow over the course of my internship. Each week I attended two meetings to present any progress that was made with the user study and to ask for team input when needed. At these meetings I was likewise kept up to date on concurrent projects, conference proceedings and upcoming events. The rest of the time I worked relatively independently. Of course, I periodically checked in with coworkers but, for the most part, my progress came from my own curiosity and drive rather than from following a list of tasks that had already been set in stone.

Week One

During my first week on the job I spent most of my time getting up to date on the project. I studiously read up on papers related to online communities, health communities, language processing, interaction design and matchmaking. I also completed my Human Subjects training to prepare for conducting user studies.

Week Two

Week two was focused on confirming the effectiveness and efficiency of the text analysis tools that had been developed for the user study. By this time our web developer was able to extract health interests from samples of written text. User data acquired from a partnering online community was processed in this manner as well as run through Linguistic Inquiry and Word Count (LIWC) to extract personal style terms. I spent my time searching for, reading and compiling a spreadsheet of cancer stories and metadata to be processed and matched using previously established algorithms. From this information, I then identified a list of the top five user stories which were used as a model for directing study participants in writing their own personal stories.  

Week Three

By week three I had become familiar with several profile interaction mockups that had been made by design students for a previous study within the project. I became excited about data visualizations and conducted my own research on the subject. In the end, I created a collection of different visualization examples along with notes on each overarching category. I then drew from these examples as I began sketching my own ideas for how we might visualize user statistics and personality information on both an interactive digital profile and the static paper profiles that would eventually be developed for the user study. During this time I also took it upon myself to begin learning how to use Axure prototyping software for mocking up my interaction ideas.

Week Four

Week four began the long journey to finalizing the user study procedure as well as finalizing the profile template that would be used for the study. In short, participants would be asked to provide a 1-2 page story about their cancer experience which would be processed and matched to potential peer mentors from our database. Paper Profiles would then be created for mentors who had a specific combination of high and low matching scores when compared to the participant. At a follow-up session, participants would be asked to rank the mentor profiles based on their preferences as well as answer a series of related questions. In the end this would allow us to identify what profile content was most helpful to participants when choosing a mentor.

I decided to take the lead on mocking up potential profiles using sample data from the database. In my first iteration content areas included demographics, health interests, and personality. However, this would change countless times before finalization.

Week Five

By week five it was confirmed that the text analysis tools being used could not be directly tied to personality. As a result, I spent my time creating several new profile templates and visualizations to be narrowed down by a team discussion and vote. The final consensus was to display demographics, health interests, sample posts and LIWC terms all in the form of plane text. Data visualizations and interaction design would then come into play after the current study parsed out user content preferences.

Week Six

Week six was pilot testing week. This was the first time that I walked through the user study with someone other than my roommate or a coworker. Everything went surprisingly smoothly and with a few minor tweaks to the protocol and profile templates I was ready to begin recruiting participants.

Week Seven

With the profile template finalized, our undergraduate web developer began automating the template making process. Unfortunately, this was not far enough along by the time we brought in our first participants so I ended up spending a ridiculous amount of time filling profile templates manually from spreadsheets that were produced by our lead developer. Nonetheless, I was able to complete the task and the user study continued to progress rather smoothly.  I learned  how to handle sensitive documents and  upload files to a secure server. I also began keeping detailed spreadsheets of participant data, metadata, profile rankings, short answer responses and, of course, the notes that I had taken during my interviews.

Week Eight

More of the same, I continued my user study work by scheduling another participant, and helping our web developer work out the kinks in his program for generating profiles. It was a slow going process but eventually all but the personal style section of the profile could be easily entered into the templates. The rest, I unfortunately still had to copy and paste by hand.

Week Nine

Week nine marked my last user study of the summer. The participant was an absolute sweetheart. We got on so well over phone and email that she brought me a little package of London decorated potpourri as a bon voyage for my study abroad trip. This was also the week that our lead researcher made her transition over to another position. Accordingly, I picked up a few more sections of the user study workflow mostly involving the process of selecting mentors based on various matching requirements.

Week Ten

This is my last week on the job this summer, as I leave for an early fall study abroad trip on Thursday. I am excited to have been offered the opportunity to extend my position, so I will be returning to my research work at the beginning of autumn quarter. My primary objective for this week is to make sure my fellow undergraduate coworker is ready to take over for me in the user study while I am away. I have been putting together collections of user study documents, profile template filling tutorials and general words of encouragement to leave him with before I go.

 

What I Learned

1. Quality Research Takes Time

As expected, this internship allowed me to develop my skills as a user researcher in a real-life, professional environment. Perhaps the most dramatic lesson I learned is that quality research takes a lot of time. The project itself had been going for years and my user study ended up taking several weeks longer to finalize than anticipated. In the process, I went through a small forest of profile template drafts before a consensus was reached on the final design. Thankfully, in the end, the numerous iterations seemed worth it. In practice, the study protocol came off smoothly, making participants feel comfortable and engaged throughout their correspondence.

2. Building Rapport is Vital for User Research

As a researcher, I also discovered the importance of empathy and composure during highly personal and emotional interactions. I knew going into the project that I would be working with cancer patients and that the potential for tears during an interview would always be lingering. Even so, I never could have imagined how it feels to be sitting next to the most cheery participant you could ever hope to have and then watch a wave of grief wash over them as they choke into tears in the middle of a story about their deceased husband. The care that must be taken for research in the cancer community is tremendous and I take it very seriously. The amount of calculated effort that a researcher puts into building rapport with their subjects shows through in the quality and emotional depth of the data that they are able to collect.

3. Emotions Can Get the Better of You

I suppose the lesson I have learned while conducting interviews is not so much how to be compassionate to a participant who confides personal information. I have always considered my tenderness towards others a strong suit of mine. What I learned about myself through these emotional periods is that sometimes I may bring a little too much of myself into my work. To explain, there was one participant in particular whose caregiver story greatly mirrored my aunt’s passing from cancer earlier this year. I read the story so that I could ground myself in the participant’s perspective for the interview. In doing so, I ended up weeping quietly to myself at my desk. To be clear, feeling deeply moved by a story is not the issue here. My concern came when the participant began to choke up during her interview. For the first time, I almost lost my composure as my eyes began to water, but I realized that, as a professional, what I really needed to do was to be there for her from a position of strength and help guide her through to the end of her thought process. I swallowed hard and maintained a supportive expression. When it was all over I decided that if I wanted to keep my cool during interviews I would need to try and draw a mental line between building rapport and my personal mission to fix every broken heart that passes by my door. Professionalism in user research often requires walking that fine line. 

4. A Smile Goes a Long Way

On a happier note, my consideration for the emotions of others has helped me bond with other members of my team. The other undergrad on the project and I quickly became friends and allies when working on the user study. This helps meetings and collaboration go smoothly. I am also extremely appreciative of the fact that I am treated as a peer rather than as a subordinate intern by other members of the project. I feel that this trust in my abilities is part of what allowed me to grow so much over the course of my internship. Mostly, I think the respect I have earned is because I am competent at my job. But I think that showing up to meetings with a smile, a can do attitude and the occasional box of cookies certainly helped my cause. 

5. Phone Interviews Are Not That Bad

For some reason I have always had an irrational fear of talking to people over the phone. I think it might have something to do with being cut off from facial expressions and other body language ques. Having to do initial phone interviews with participants certainly helped alleviate whatever anxiety I once had. Luckily, members of the cancer community are notoriously kind and understanding. While I still like to write out some notes on what I will say beforehand, I now feel much more comfortable striking up conversations with strangers and recruiting them as participants over the phone.

6. I Can Do This!

Finally, a lesson that I learned is that I am far more capable than I ever considered myself to be. I was able to take on and excel in a role that I had never really taken on before. Moreover, I was competent enough to interact with participants up to three times my age and still earn their respect. The chance to feel like an educated professional is empowering and only adds to the excitement of moving forward with my degree. I may not pursue a career directly involving health informatics but I think that the experience I have gained will serve as universally useful for my future work in user research.

 

Qualitative Analysis – Affinity Diagramming 

    

 

 

Data Template (with placeholder text)