Discussions on the art and craft of research

Tag: Search strategy

3rd Annual Library Prize Recipient – Dominic Lentini

For the last three years, Shain Library has been awarding a Library Research Prize. Each student applicant must submit detailed information on their research process, and get faculty support for the submission. This is the essay/application for this year’s winner, Dominic Lentini. Dominic is a senior, and a double-major in International Relations and French. His paper was entitled: Media Framing, Violent Protest, and Race: A Comparative Analysis of The Wall Street Journal and the New York Times’ coverage of the Ferguson Protests.

Dominic Lentini ’18
Research Statement
Spring 2018

Describe how you came to choose your topic, specifically noting any pre-research that you did. What sources did you use in this pre-research? To what extent did you consult with librarians, faculty, or others? How did this pre-research lead you to your topic?

The process through which I arrived at my final topic was very time consuming. The first proposal I wrote was about protest repression, media coverage, and the police. For this, I first consulted the textbooks as well as other assigned readings for our class and I critically examined their bibliographies to help guide me in the direction of appropriate and related literature. This search involved exploring both theoretical literature to establish a framework for my analysis, as well as information on potential case studies and primary sources that could be used to take the existing research in a new direction. Using those sources as a springboard, I compiled a large list of peer reviewed articles and books on protest policing, policy, and organization, as well as on media coverage of protests.

However, as I began to read through those texts, consult with my professor, and meet with research librarians, I realized that what I had proposed could be three separate papers. While my research clearly started with a very large scope and a lot of energy was used to research topics that I did not write about, over roughly a month and a half of reading and evaluating sources, I eventually guided and narrowed my initial interests into a topic that was appropriate for the course: Media Framing, Violent Protest, and Race: A Comparative Analysis of The Wall Street Journal and the New York Times’ coverage of the Ferguson Protests.

Describe your process of finding information for your project. Note specifically the tools you used to undertake your research, as well as the specific search strategies you used within these tools. (Note: “Ebsco,” being an umbrella vendor, is not a specific enough response when identifying tools; listing the “library database” is also an unacceptably vague answer.

Specific tools include JSTOR, America:History & Life, Web of Science, etc., along with OneSearch, the new library system.)

As I described in question one, the first step in my process was exploring the sources used in the bibliographies of sources provided for my course. Following that, however, I used the library’s OneSearch, JSTOR, Political Science Complete, Google Scholar, and Lexis Nexis as the main tools for finding new articles and books. Within all of the databases I would do advanced searches with a variety of different search terms such as “framing,” “media framing,” “framing violence,” “framing race,” “framing protest” and many combinations within those terms. I would then read the abstracts to gauge potential relevance, and save every potential article to RefWorks so that I could later examine their
methodology, data, and conclusions. For anything I could not access through those databases, I used both the CTW network, WorldCat, and Inter Library Loan to access them. Additionally, within JSTOR I explored the utility of their text analyzer, which is in its beta mode.

For collecting my primary source newspaper articles, I initially used ProQuest Newspapers. I even contacted them, with the help of Andrew Lopez, to learn about how they code and sort their articles. For my data collection process, I used ProQuest Newspapers to search all articles published in certain date ranges based on set search terms in order to create frequency tables of article publication, and then to do content analysis of a selection of those articles. I realized, however, that some articles were coded inconsistently, and consequently double counted, which threw off all of the article counts. Thus, I did not end up using that particular database. Instead, I used the website search function for both The New York Times’ and the Wall Street Journal’s sites. Using the search functions within each newspaper required more manual work, and I even called the WSJ to get information about how their search feature functioned, but it ultimately provided me with the data I needed.

Describe your process of evaluating the resources you found. How did you make decisions about which resources you would use, and which you wouldn’t? What kinds of questions did you ask yourself about resources in order to determine whether they were worthy of inclusion?

I went through several different steps to evaluate my sources. Firstly, for my literature review, I only considered peer reviewed articles and books. Within sources that met that requirement, I would examine their research methodology as well as their bibliography in order to gauge the soundness and scope of their argument and conclusions. This process, however, still left me with more articles than I could use. Consequently, I made my final selection with the intention of laying a base to the framing literature, then additions and variations to that literature, and finally critiques to it. The ultimate goal was to paint a well-rounded picture of the literature.

The process for selecting background pieces for my case study was more challenging. For one, due to the slow process of academic publication, there does not exist a huge body of peer reviewed literature on the Ferguson protests. Consequently, most of the information on what transpired had to be gathered from newspaper and magazine sources. This, however, left me in a paradoxical situation because I was being pushed to use newspapers as the background for a paper in which I was arguing that newspapers paint a “framed” version of what transpired during the Ferguson protests. To try and mitigate this issue, I used a wide range of newspapers and magazines, as well as any quality academic literature I could find, in order to cross reference and evaluate the validity of my sources. While this did not totally eliminate the issue, it definitely reduced its severity.

Research and the Information Process

What is the “information creation process,” and what does it have to do with scholarly research? Short answer: a lot.

Longer answer (if you’re still with me!): In a previous post, I wrote about the first of the threshold concepts developed by the Association of College & Research Libraries’ (ACRL) in its “Framework for Information Literacy for Higher Education.” Threshold concepts are key points that an information-literate student or researcher needs to be able to grasp and utilize. The ACRL’s first threshold concept, “authority is constructed and contextual,” seemed tailor made for a political moment rife with discussions and anxieties about fake news, post truth and alternative facts, and indeed we librarians engaged in several robust discussions with faculty about how to approach this topic in the curriculum.

At first glance, the second threshold concept, “Information creation as a process,” might not resonate quite as strongly. But perhaps it should. Because an understanding of the process of how information is created is necessary for discerning how various books, articles, etc., might be useful or not useful — or for figuring out how authoritative they might be.

First, what does that mean, “information creation as a process”? Information objects — resources that are in various containers, including books, newspaper articles, scholarly journal articles, magazine articles, websites, blog posts, tweets, scholarly proceedings and yes, Facebook posts of dubious origin — all are informed by, and in turn inform, other information objects. In other words, the information contained in these various objects is used to create other objects; in turn, as these objects are read and shared, they may themselves serve to generate tweets, news articles, scholarly articles and books. The process is not merely circular, but infinitely weblike in the way that new information shapes, and is shaped by, already existing sources.

What’s more, the kinds of information objects that are available on any given topic depend on several factors: how long the topic has been in discussion, the extent of the discourse surrounding that topic, and the information objects that have already been created on that topic.

Let’s take, for example a story that’s very much been in the news: Donald Trump’s controversial executive order on immigration from Muslim-majority countries. Although the news of the order hit newspaper websites very quickly following the announcement, even before those stories appeared there were innumerable tweets, Facebook posts and other social media commentary offering even quicker takes.

How useful are these quick takes on the issue? Certainly, they serve as a record of the issue’s explosiveness — its vast potential for altering domestic and global social affairs, politics and business. If one’s project is to document the proliferation of information objects about the issue, then gathering the tweets and quick news pieces would be essential. Similarly, a researcher who sought to document the speculation about the effects of the travel bans would need to look at these early sources.

But the usefulness of any particular information source depends very much on the nature of the project at hand — and that source’s place in the information cycle. What if, instead of looking at the speculation about the executive order, one’s project was about examining its effects? For that, one needs something more analytical — one that views the immigration orders at a distance. One might first look at any newspaper articles that appeared since Trump’s announcement — but again, depending on when the research was being conducted (six months after an event? a year? five years?), more detailed, rigorously conducted scholarly sources might be available.

Let’s look at how this plays out in various searches. Searching Google for “Donald Trump executive order immigration” in April — roughly three months after the initial announcement — yielded nearly 3 million results. Closer to home, entering the same keywords into the library’s CrossSearch tool (which searches books as well as numerous article databases) predictably yields fewer results. (Still quite a few at more than 2,000, but not quite the 3 million that Google unearths.)

Screen Shot 2017-04-19 at 11.35.12 AM

Of these results, most are news articles:

Screen Shot 2017-04-19 at 11.35.41 AM

Only a handful appeared in academic journals, and most of these are quick takes of only a page or two, certainly not the detailed, rigorous, analytical pieces one might expect to find in a scholarly publication. To find such an article, it would be necessary to wait until more time had passed — until scholars had been able to perform studies that involved gathering and interpreting survey data, and that carefully surveyed the available literature on the subject.

Screen Shot 2017-04-19 at 11.36.11 AM

Note, too, that the search doesn’t yield any books at all. That’s because books take even more time than scholarly journal articles to put together. Books need to be proposed, written, edited, rewritten and then published before they can arrive on the scene. The advantage of books, whether they appear in print or electronic form, is that they typically represent some of the most considered, most rigorous thinking on a particular topic. But they take time, and usually depend on the fact that previous news articles, scholarly articles and other materials have already appeared on the topic.

And so a hypothetical researcher seeking to examine the effects of immigration bans would certainly want to look at detailed scholarly journals and books, if they were available. But to determine the kind of questions that are even possible, it’s necessary to understand what materials might be in the information-creation process, and thus how helpful or authoritative they might be in answering the question one poses.

So when you ask a research question, it’s a good idea to think about where a topic might be in the overall information-creation process. The answers will help not only in guiding you to the best possible sources to answer your question, but also in figuring out what might be available in the first place.

— Fred Folmer

Google Searches: Mastering Your Domain

As we all know, Google searches typically return huge amounts of results — numbering well into the millions — for practically any initial search. As frustrating as this can be, it’s especially vexing when adding additional search terms not only fails to limit the number of results in any significant way, it also may screen out lots of possibly useful items.

But it turns out that there’s a way to limit your results that doesn’t involve adding search terms — and that can help you retrieve useful items much faster. This method involves adding a domain search along with your search terms. And while this may sound daunting, it’s actually quite easy.

First, you can search in a wide domain, such as .edu, .org or .gov. The latter can be especially useful in finding an array of public materials, such as government reports, statistics and court cases. Suppose we’re looking for statistics on gun violence. If we search for “gun violence statistics” (without using quotes in the search), a Google search turns up 4,550,000 results — a hodgepodge of news stories, government reports, nonprofit organizations (including pro-gun advocacy groups like the NRA) and Wikipedia entries. But if we enter the following string into a Google search, we’re now at 30,700 results:

gun violence statistics site:gov

See how you do that? To whatever search you’re doing, you add the word “site,” followed by a colon, followed by the domain. In this search — which now only looks at sites with the “.gov” domain — the very first link that appears is a report from the Bureau of Justice Statistics documenting firearm violence statistics from 1993 through 2011.

Screen Shot gov search
What’s more, once we know that the U.S. Bureau of Justice Statistics ( may be a useful site in which to look, we can limit our searches to only sites within the Bureau’s domain, that is: “gun violence statistics”. To wit:

Screen Shot bjs search

We now have only 425 results to comb through. Moreover, we can easily see other categories, such as “Weapon Use,” on which the BJS offers information, and adjust our search terms accordingly to pinpoint the items that might be most useful to our project.

This technique can be used to similar effect to search .org or .edu domains — whether we’re looking within entire domains or particular sites. It’s especially useful when searching organizations big, complex web structures, such as the United Nations ( There are, of course, pros and cons to searching any one of these domains, and this may be the subject of a future blog post. In any case, it’s worth remembering that limiting to a wide or narrow domain may help you find useful items quicker, and at the same time lessen one’s experience of information overload.

— Fred Folmer

© 2018 ResearchScapes

Theme by Anders NorenUp ↑