Efficient study of a topic is a powerful skill for a wide variety of domains. However – despite having several classes on writing and research in High School and Undergraduate Studies – it wasn’t something that I truly appreciated and learned how to do well until Graduate School. I think this really boils down to lacking the kind of personal application that makes these efforts meaningful. However now that I have come to appreciate it, I want to highlight how I think how I study new topics, and how this applies to many topics outside of academic research.

I think its important to clarify that this post is not designed to present tactics but strategy about how to approach new topics. First, effective tactics vary widely between person to person. Some will prefer the use of electronic tools for certain tasks, while others will favor pen and paper. It is important to utilize the tools and techniques that work best for you. Second, effective tactics can vary widely by topic. Some topics have extensive free resources that cover a topic. Other require access to journals and conference papers to really approach a topic in depth.

Broadly speaking, I think effectively studying a new topic involves several steps:

1. Begin with a Research Question(s) – What are you going to research?
2. Build a Candidate Reading List – What places could have the information?
3. Read the Papers Intentionally – What do those places say?
4. Record, Organize, and Categorize – How can I remember what I’ve learned?
5. Focus and Refine Your Reading List, Update the Research Question(s) – What should I look at next?

If you are thinking this looks similar to the PQRST approach that I mentioned in the Learning to Learn Reading post, it does, but I think there are some different aspects to studying multiple works rather than just one that warrant studying this further. For example, in doing a comprehensive study, doing comparisons and contrasts are more important. Most works by a single author will present a single point of view on a topic – or at the very least favor a particular perspective. However by reading works by multiple authors, you can expose subtleties that wouldn’t appear if you just reviewed one topic. Secondly, with a comprehensive study, archival and retrieval becomes more important. This is mostly because of the shear increase in the volume of information which can make it difficult to keep it all in memory. With comprehensive studies it becomes essential to be able to find the knowledge that you already had quickly which is less important with reading a single work.

How Do I Craft a Research Question?

Your initial research questions should be design to answer the question, “Why am I looking into the topic?” The key tool in doing efficient research is maximizing the signal to noise ratio. You don’t gain anything from reading irrelevant documents. Starting with question, “Why am I looking at this?”, gives you a clear guide as to what is or isn’t relevant allowing you to ignore irrelevant sources faster to increase the signal to noise ratio.

In that vein, it is helpful to be as specific as possible! This too will help cut down on reading irrelevant documents. For example rather than “performance”, search instead for “network bandwidth” if you are really focused on the bandwidth of computer networks. You can always increase the scope of your research question later if you want to look at performance more broadly.

So what if you still don’t know where to start? I often start with a variant of one of these questions:

• What are the views on a topic? Understating how others view a topic has many benefits: it improves understanding and empathy, it can provide an concept of the prominence of an idea, and it can tell you what others view as fruitful areas of research.
• How does this topic apply to some other topic? One problem with academia is the apparent lack of application. Focusing on an application from the beginning can improve this. Many aspects of research involve applying some underlying skill or knowledge to a new topic.
• How can I solve X problem? This the fundamental engineering question. It may look like subset of the previous question, but has an important distinction: the previous question begins two or more candidate topics and looks for a candidate problem. The later begins with a problem and examines possible solutions. Both approaches can yield valuable results.
• What are the trade-offs for X? This is really the key scientific question. Understanding trade-offs can allow engineers and others to build upon your work to consider new topics and expose new tools.

• Try asking the questions above of the topic that you want to research. What aspects of your topic did you discover when thinking about it systematically?

How Do I Build a Reading List?

The next important question to ask is what to read and what to read first.

Its key to remember while building your reading list, you don’t have to read the entire document. I generally do a first pass on titles, then I filter more aggressively on abstracts, and finally I filter even more aggressively on skimming the relevant sections of an document. This is where tools like electronic search can be very helpful to look for key words and phrases and topics. I would suggest taking the time to read only the most information dense resources in full. When you go to cite or use a resource, then you should go back and get a more full context.

If you are completely new to a topic, it can be helpful to begin with some preliminary reading. This can take several forms. Here are some of the ones that I use most often:

• Begin with a survey or a book on the topic – I really don’t know anything about a topic the first thing that I try to find is a book on the subject. While most people have heard of books, some many not know about survey articles. Survey articles are documents that provide a comprehensive overview of a topic and tell you a wide variety of research going on or views on an area. So then why a book or survey article in the age of the Internet? First, Books and surveys can offer a holistic view of a topic and have fewer constrains to length or subject that a conference, journal, or article can have which allows them to be more introductory and provide more context. Second, books often are edited and reviewed which means that the document can be easier to read than that of some journals/articles.
• Search for components of the research question – Once you have a fairly precise research question, you can begin to search the Internet for resources on a topic. In additional to general search engines like Google or Google’s Scholar search engine, sometimes specialist search engines such as the ACM digital library search or the IEEE Explore in the field of computer science can provide more specific results for their topic area. A use of both can vastly improve the variety of results that you may find.
• Look for authors, keywords, and further/active questions. As you are reading, make note of authors, key words/phrases and active research questions that these papers identify. This can aid in the development of further and more precise research questions to review later.

Once you have an initial corpus of documents, you can expand your reading list to contain other possibly relevant materials. There are three key ways that I try to do this:

• Read books that cite/are cited by other works that were relevant – Academic work seldom occurs in a vacuum. In additional to understanding the topical context, it can be valuable to understand the temporal context as well. Why are the authors writing about this topic now? How did there work change over time and why? Understanding the answers to these questions can give you an idea of the barrier of entry to a useful contribution.
• Read more than just websites. As I said the subsection on preliminary reading, not all writings are written in the same style. Even within a given genre of non-fiction, there are substantial differences between journals, conferences, books, debates, tech reports, and websites. Understanding how each of these are used in your specific discipline can improve your ability find the kinds of information that you are interested in.
• Consider alternative perspectives. This is especially important in topics that are contentious or unclear. For these topics it can be enlightening to get past the caricatures often used by opposite sides of a contentious issues to understand their fundamental concerns. While you don’t have to agree with them, understanding their perspective is helpful.

If you’ve followed all this advice in generating a reading list, you likely have more reading material than you have time to read it. Thus it becomes important to cut the list of things you have to read. Remember, the goal is to maximize the signal to noise ratio. I often will mentally sort the resources by expected value versus the cost of time to read. Then I mentally budget how much time that I have to read and will cut off the list when I run out of budget. As I suggested earlier in this section, you can do this in multiple progressively time intensive phases.

How Can I Record, Organize and Categorize My Reading

I think the biggest distinction between reading for pleasure, reading a single document, or doing a systematic study is the importance of recording, organizing, and summarizing information. Again remember the goal is improve the ability to find and retain information so that you don’t have to read the entire document again in full. Every time that you need to review an entire document in full for the same information or have to spend a long time searching for a document that you already found, you often are not using time as efficiently as possible. This is not to say that you shouldn’t read documents twice, but you should strive to find new information each time you do.

This takes two key forms forms: minimizing time to find a document and minimizing time to find information within a document. These are in some senses the same problem: finding the source (or part of a source) that has the information you are looking for. Finding information within a document requires or uses similar kinds of techniques but done at a more granular level. I would describe these techniques as indexing, summarizing, and archiving.

Indexing is the problem of making it easy to find a large number of documents often times this looks like organizing information by topic or keywords, tagging information with keywords, or putting it into a system/database that records this informant. Picking the indexes to use can be challenging. You often what a combination of specific and general tags to help you find information for when you want to find a specific thing as well as a variety of things about a topic.

Summarizing is a similar problem, but slightly more verbose, it provides a good way to zero in on a specific resource on a topic of interest, but provides greater context. Doing this well is challenging, if a summary is written two distinctly, it can be hard to find similar documents, but a summary that is too generic can not serve as a meaningful index. Again, the goal is to minimize the time to find information you’ve previously ingested. Often times it is important to record not only what you found, but how and when you found it. The more tags that you can associate with a piece of informant, the easier it will be to find later. Additionally, this will often can enable you to find new information because often times information is found in clusters.

Lastly is archiving. Websites go down all the time, books go out of print or are damaged, and sometimes your ability to access certain resources changes over time. Having a good record of information is useless if you cannot retrieve the information later. It is often important to make a copy of the information that you found so that you can retrieve it later even if the source is hard to find. However there can be costs to this, everything you store requires time to search through when you want to approach a topic. It can often be important to have a tiered system of archival for documents that are more recently reverent versus those which are more less currently relevant.

• Think about how you index and achieve your findings right now. Can you quickly find resources that you identified before? Why or why not?

Closing Suggestions

In closing, I would encourage the reader to apply these techniques to a wide variety of research problems. I use this both for work, but also when I attempt to answer questions posed to me by friends, or for personal study of topics that are important to me such as my faith.

Happy learning!