Thursday, September 14, 2006

UMass/Globe LexisNexis numbers: I refine the search

I was intrigued enough by the debate over the Boston Globe's coverage of UMass Boston vs. other local colleges -- and the interpretation of LexisNexis search results that might indicate bias on the Globe's part -- that I decided to conduct my own refined LexisNexis search, using some of the lessons I have learned in my own thesis research. It took less than 20 minutes, but the results clear away some of the lingering doubts concerning passing references in full-text searches in LexisNexis. I submitted the results to the Universal Hub posting "Why Doesn't The Boston Globe Care About UMass Boston," where someone questioned the original LexisNexis full-text search conducted by The Concerned Members of the UMass Boston College of Public and Community Service. I am posting the results here as well, at the bottom of this entry.

Before I describe the results, let me first say that the problem observed by commenter tblade on Universal Hub -- full-text searches turning up too many passing references to alumni, sports, etc. that naturally favor the big private universities over UMass Boston. This issue is I have been aware of since starting my own LexisNexis research on Xinhua/NCNA content relating to Vietnam. Full-text searches in 1970s Xinhua content turns up a lot of passing references to Vietnam, i.e., the ambassador from Vietnam being mentioned in passing, in the last paragraph about a banquet celebrating China's National Day. My solution in the second-stage data collection (Yoshioder sampling): Restrict searches to the headline, which gets articles that are really focused on the issue, rather than mentioning it in passing in the fourth paragraph. So, for the UMass LexisNexis searches, I restricted the results to headline references only (with one exception, described toward the bottom of this post).

There was also another potential problem with the original LexisNexis research conducted by the UMass group complaining about Boston Globe's coverage of their school: The methodology is described as follows:
full text Lexis-Nexis search on the names of each of the 4 private colleges for articles of all types appearing in the Boston Globe
Does that mean the group only entered one name for each college/university in the LexisNexis search field? If the answer is yes, the searches will be flawed owing to the fact that the schools often have more than one name (i.e., formal and nickname) and headline writers try to use the short version whereever possible. In other words, if you conducted a search for "Boston University," but not BU, you'd miss almost every reference to Boston University, because headline writers for the Globe use "BU" most of the time. So, in my searches, I used multiple terms -- "Boston University" or "BU".

My results were naturally much smaller, mainly because of the headline restriction, but I think they support the UMass group's initial observation. There is an undeniable disparity in the Boston Globe's coverage of local colleges and universities. That doesn't mean that the other UMass student (tblade) who questioned their use of LexisNexis is wrong, however. In his two comments on the Universal Hub thread, he did have some valid points about the nature of what makes news, and why the private schools will get more coverage from the Globe.

OK, without further ado, my results, which I have also submitted to Universal Hub:
Good point, tblade, but I think the UMass student group is onto something here. I used LexisNexis Academic to restrict the articles to those that mentioned the schools in the headline, which screens out most of the passing references to alumni, sports, etc., and instead concentrates on legitimate coverage of college/university-related issues. Here's the ranking, for 1/1/2006 to 9/10/2006:

"Boston College" or "BC" in headline - 187

"Harvard" in headline - 170

"Boston University" or "BU" in headline - 62

"Northeastern University" or "NU" or "Northeastern" in headline - 47

MIT or "Massachusetts Institute of technology" in headline - 47

"University of Massachusetts at Boston" or "UMass Boston" in headline - 4

UMass in headline, and "University of Massachusetts at Boston" or "UMass Boston" in the full text, but not Dartmouth or Lowell or Amherst in full text (this allows for the short version of UMass in headline, which copy editors and editors prefer, but the article is about UMass Boston, as opposed to UMass Amherst, UMass Lowell, and UMass Dartmouth) - 11

"Framingham State" in headline - 2
One important note about these LexisNexis searches: I did not restrict by section or by the Globe's two higher education reporters. This resulted in an especially large number of "hits" for BC, which has strong football and basketball teams that are frequently mentioned in the Globe's sports section.

Nevertheless, I'd be very interested in finding out what the Boston Globe and its two higher-ed reporters, Marcella Bombardieri and Sarah Schweitzer, have to say about the results of the LexisNexis searches, and the complaint raised by The Concerned Members of the UMass Boston College of Public and Community Service.


Anonymous said...

This is just a direct cut and paste of my Universal Hub comment:

Well Done.

I support the group's argument, but I was upset at the authors' laziness. As a student, I know that UMB and CPCS are under represented in media; but I fear that if someone acheives the same conclusions as I did, they may dismiss the argument and continue to ignore UMB.

To take on The Globe, writers need to have their ducks in a row. A more iron-clad analysis, one that doesn't take significant additional time (20 mins by your count), excludes the letter from certain criticisms. You succesfully argued the group's point using more realistic data. I also like how you pushed it a step further and included institutions such as Framingham State and used additional school monikers (Boston College + BC).

At any rate, this thread and your two posts over at gave me some food for thought on differnt uses for LexisNexis and how to improve my own search techniques. God bless hte interweb.

Good luck on the thesis.

I Lamont said...

Thanks tblade. I wouldn't have taken this step if it hadn't been for your comments on Universal Hub questioning the results of the original LexisNexis research conducted by the Concerned Members of the UMass Boston College of Public and Community Service.

Anonymous said...

I thought it was common knowledge that the Globe, particularly the Sports Dept., is biased in favor of BC. The Sports page is practically an extension of BC's pr and recruiting program, with frequent hyperbole and exaggerated claims about BC's supposed emphasis on academic standards and under-reporting on scandals at BC, particularly if there is criminal conduct.