Recently, several students experienced difficulties in their research process, which raises many doubts and potentially negative feelings toward research. Furthermore, research is a challenging skill to master, and it is very different from many of the other skills. In my experience, students in HCI research typically come from three types of backgrounds: 1) computer science or computer engineering undergraduate students who have worked on some projects before, but without deep exposure to research 2) students with design background who have learned the design methodologies, but less trained in scientific disciplines 3) former employees from a company who are well trained in engineering disciplines, but have not worked on research before. Students from these backgrounds may all face significant challenges to master HCI research skills, although with different characteristics. The purpose of this article is to provide an overview of the characteristics of (HCI) research and how students from different backgrounds can better overcome the barrier of entry for HCI research.
The HCI research process
So what is it like to work on an (HCI) research problem? If I can make an analogy, it is like exploring an unknown land (see figure below), and finding a new optimal path that’s previously unexplored in that piece of land.
This unknown land exhibits the following characteristics:
- Its boundary is unknown. You only know the rough direction/location of it, but its exact position, scope are not specified. In fact, it is your job to define its scope and location. This by itself can be a challenge to many students. Given the enormous world of knowledge, how can one pick the most interesting and feasible unknown land to explore?
- Not all unknown lands are interesting or solvable. Some unknown lands are solved (existing paths are already optimal, thus no need for new paths). Some contain dead-end paths, which can be frustrating after you have spent a lot of time exploring it and find there is no way out. There is no guarantee that a new optimal path can be identified in any given unknown land. When such situations happen, one needs to either adjust the path, adjust the boundary (scope) of the unknown land, or find an alternative unknown land to explore. Such adjustment of path, boundary happens very frequently, and in some cases, one needs to abandon the earlier exploration and find a new unknown land to explore.
- An unknown land can (typically) only be explored progressively. When exploring the land at the ground level piece by piece, one can encounter challenges and difficulties in each of the sub-areas of the unknown land. During this process, if without a bird-eye view, one can often get lost and feel frustrated. The experience is like, after finally overcoming very difficult challenges in one sub-area of the unknown land, new challenges suddenly emerge from the next sub-area, and this process may seem endless. Without a deeper understanding of the nature of research, and the ability to see the bird-eye overview, students can easily feel frustrated, discouraged during this process.
- Not all segments of the optimal path are equal. Although the goal is to identify this new optimal path, not all segments of the path have equal importance. Some parts of the path are on the critical path, which needs to be identified first. Other parts can be secondary or complementary. During exploration, one needs to prioritize the critical elements of the path first before moving to the secondary ones.
Given the characteristics, how can we minimize the pain in the exploration process?
- Limit the size of the unknown island. With limited resources, the size (scope) of the unknown land a student can explore is limited. Thus, it is essential to choose a piece of unknown land of adequate size for exploration. If the size of the unknown land is too big, exploration under a time constraint (typically maximum 1 year for a paper submission) will be extremely challenging. Choosing an unknown land with the right size is the first skill a student needs to learn and potentially master.
- Define the boundary/scope. Once a land of a certain size is chosen, the next important step is to define its exact boundary. How to define the boundary of an unknown land? One effective approach is to write an abstract and outline. Students are often told the importance of writing abstracts and outlines before any prototypes are developed and any results are collected. They may be puzzled: what’s the point of writing them when we don’t have a prototype and know nothing about the results? The answer is tightly associated with the first characteristic of the exploration process. The writing of the abstract and outline is to help you to define the boundary of the unknown land. A well defined outline helps to avoid one common mistake: during literature review, many related, interesting ideas will surface, and they can easily distract you to spend a lot of time exploring ideas outside the scope of your unknown land, which can be counterproductive.
- Distinguish critical vs. secondary paths. To explore a path more effectively, one needs to be clear about which part of the path is critical, which needs to be explored first, and which part of the path is secondary & complementary, which can be explored at a later stage. Note that this distinction is hierarchical. Even in a critical path, certain exploration is more critical than others. This may feel abstract, so let me provide an example: A paper has 3 studies: study 1 is to motivate the investigation (people need Bla), study 2 is to show the advantage of Bla (Bla is better than a number of alternatives including the state of art), and study 3 is to generalize the finding of study 2 in real world settings (ecological validity). Among these three studies, which one is the on the critical path? The answer is study 2. Remember, we are looking for a new optimal path, and the optimal means it is the best path. Among the three studies, study 2 establishes this claim, thus it needs to be worked out first. However, it doesn’t mean everything related to study 2 is critical. For example, to report a formal study in a paper, we may need 18 participants, but to know the result, typically 9 participants are enough. Thus the first 9 participants are on the critical part of the critical path while the remaining 9 participants are less important. When performing this study, one can prioritize the first 9 participants while leaving the remaining 9 participants to a later time or for someone else to work on in order to concentrate on the most important tasks.
- Detects dead-ends or failed paths early and always be prepared to adjust the boundary of the unknown land. This can be difficult for students coming from the industry and engineering background. Engineering emphasizes preciseness and clear requirements from the beginning, yet research, especially in the beginning stage, requires extreme flexibility and an open mind. In research, we probe the unknown land to check for the possibility of the existence of a new optimal path first, and always ready to change to a new unknown land if the result of the probe is negative. Thus, it is important to design the most efficient probes to understand the situation and get ready to adjust exploration based on the results. It can be particularly painful, if one spends a lot of time and effort exploring a piece of land, and finding it to be a dead end in the end.
- Seek help and work with others. Research is hard. Developing a birds-eye view is difficult for beginners. There are many challenges and it can appear endless at times. Thus, don’t be shy to seek advice, help from others. Manage your resources strategically. Balance your life and prepare for a sustainable effort. If you can successfully overcome these difficulties, you will be able to taste the excitement and joy of research, as it is truly a beautiful thing.
Article by Dr. Shengdong Zhao