This post puts together some informal thoughts on how to get the most out of an academic paper. I’m grateful to discussions with Pat Reed, Thorsten Wagener, and Klaus Keller through the years that have given me some of these ideas.
How to Find Good Papers
- Use Web of Knowledge and/or Google Scholar to search for the most relevant citations. You can even start with a general topic, such as “water supply planning”. There will be 1,000s of citations of course. But “sort by times cited” and you will likely find the most important benchmark papers everyone has read. Download these (at least) and read them (preferably). You’ll be expected to know these references!
- Make good use of review articles. Did you know that Nicklow et al. (2010) reviewed applications of evolutionary algorithms in water resources? (find it here: http://link.aip.org/link/doi/10.1061/(ASCE)WR.1943-5452.0000053) Review articles like this are great resources for learning a lot about a field. There are similar reviews for hydro-economic modeling (Harou et al., http://dx.doi.org/10.1016/j.jhydrol.2009.06.037) and multi-reservoir operations (Labadie, http://dx.doi.org/10.1061/(ASCE)0733-9496(2004)130:2(93)). I’m sure there are good examples for your field too.
- Look at group websites to collect more than one paper from the same author. Our group website is a great example of course.
- Use literature reviews in other papers and theses. A lot of times, other papers, dissertations, and theses do some of the work for you by reviewing the literature in a particular field. Use these resources and download the papers cited by these other authors. Of course, do not plagiarize their words. If you’re borrowing ideas from a list of literature from Smith (2012), you can even cite Smith (2012) by saying “As reviewed by Smith (2012)…”
How to Read a Paper
You’ve found some good papers to read. So you get yourself a cup of tea, print out a paper, and start out at page 1. That’s not really the best way to go about reading the paper! What if this paper isn’t one that you actually need to read? Let’s face it, you will probably have to cite 100 papers in your thesis and it is difficult to read every single one, especially in one sitting. What if the important info doesn’t start until page 15? The human attention span is not very long, and you could get yourself lost.
Instead, try this approach:
- Read the abstract. A good abstract will tell you what the paper aims to achieve, what methods the authors used to achieve those aims, and the implications of the results. A great abstract will also discuss the limitations of prior work in the field, and how the presented work could be expanded to other studies or other fields.
- Does the abstract seem relevant and interesting? Great. Now Look at the figures. What types of analysis are the authors presenting? Do the figures make sense, and do the captions explain what you’re supposed to look for? When you’re reading the full text later, you’ll want to use the figures as a roadmap. It’s helpful to know what’s coming so that you’ve seen it before you get there.
- Is the paper still keeping your interest? Wonderful. Time to read the conclusion. The conclusion should give the authors’ insight on what it is that they actually did. This should give you the take-home message that you should, well, take home when you read the work.
- Now you can Start at the beginning and read the paper. Pay particular attention to the methodology — if the paper talks about a basin in Malaysia, it probably uses a model or analysis technique that you could apply to your own basin. It’s not a good enough excuse to say “Oh, well the authors aren’t working on a problem that’s exactly like mine.” You should try to be familiar with papers that are from all sorts of different fields.
Remember that you can get a lot out of the first few steps of the process. So if you look at the abstract, the figures, and introduction, you may get enough out of it to save the paper for a more careful treatment later. It’s better to be familiar with a whole lot of references from many different authors and groups, in my opinion, than get tunnel vision on one paper. Especially since you will get more out of a paper if you revisit it later after you’ve learned more about the field.
Some Tasks to Try
A lot of people need to “do” something when reading to make sure they get the jist of the paper effectively. Here are some suggestions:
- Highlighting. This is pretty self explanatory, but try the features in Adobe Reader or the free FoxIt reader (see http://www.foxitsoftware.com/) Also question things that you don’t understand or don’t agree with in the margins of the paper (i.e., “What were they thinking?”). This really helps when you revisit the paper later.
- Write a one-sentence summary. This is harder that it would seem at first. How do you distill a 20 page paper down to a single sentence? This is a good habit to get into for every paper you read, especially since you will probably need to do it when you’re writing the literature review in your thesis. Most papers will put a sentence like this right in the abstract, so adapt it from there.
- Write a 500 word summary. Again, it’s harder than it initially seems. This page gives some helpful hints on writing summaries. Always do this without plagiarizing the original material. Writing a succinct summary of something can be a valuable skill, especially when adapting your own work in different venues.
Download Reed and Minsker (2004) “Striking the Balance: Long-Term Groundwater Monitoring Design for Conflicting Objectives” here: http://link.aip.org/link/doi/10.1061/(ASCE)0733-9496(2004)130:2(140). It’s a foundational paper for our field, since it’s one of the first applications of a many-objective (4 or more) optimization problem in water. Fulfill the following tasks:
- Write a one-sentence summary of the paper.
- Write a 500-word summary of the paper, making sure you hit the most important results presented there.
- Provide a brief critique, including one thing the paper did well and one thing it did poorly or you want to see expanded.
- List the most important 3 references cited in the paper and discuss their relevance to the study. Does the current study expand or improve on these references?
As always feel free to add comments or questions below!