or, The Hitchhiker’s Guide to Fear and Loathing at a Public Library Reference Desk



Reference Question of the Week – 5/22/11

   May 28th, 2011

I was all proud of myself for ultimately finding the answer to this question - but afterward I discovered the answer wasn't nearly as hard to find as I had thought. Oh well.

A patron came up to the desk and pushed this newspaper clipping towards me (click to read it):

WSJ clipping

As she did this she said,

This from the Wall Street Journal page A12, but I can't remember date. I looked through all the issues you have, but it's not in any of them. You need to find out when is this article from because I want to read the rest of it. I'll leave this with you and go back to my computer, so just bring it over when you find it.

We keep the last three months of the WSJ in print, and since she said she looked at every issue, that ruled out anything between now and March 2011. I asked her if she had any idea when she photocopied it, and she said she thought it was in March, but it could have been a little earlier, so I decided to focus my search between January and March 2011.

Unfortunately, we don't have subscription database access to the WSJ, so I went to their website to see what kind of archive search they had. Their search did allow limiting to a date range, so I combined that with what seemed like the most important keywords from the article (Victim Funds, Columbine High School, Virginia Tech, Von Mour, etc) - and came up empty.

I usually don't have much luck with newspaper searches, so I quickly switched to Google - sometimes their cached results contain a better record of what has been published (or at least a temporary view through paywalls), and I'd be happy if I could even just get a citation. I searched on various iterations of those same keywords, and included the "site:wsj.com" limiter, but still had no luck.

I thought I was on the right track, but just using the wrong keywords, so I reread what I could of the article, looking for something unique. Towards the top of the page is a photo caption listing peoples' names, so I tried another Google search for "victim funds" "dan smolnik" site:wsj.com and got exactly one hit.

Clicking into the article and skimming it, I saw the same "Victim Funds" table, and also did a Ctrl+F for the phrase "As many as 4.62 points," which appears at the bottom of the clipping, so I knew this was the right article.

So yay, that made me happy. I scrolled back up to the top to find the date: March 28, 2010. Wow, the patron had the right month, but the wrong year.

I went over to the patron's computer and pulled up the article for her. At first she was skeptical because of the year, but when I showed her the table and the same paragraphs from the clipping, she agreed it must be the same one.

After getting back to the desk, I felt pretty proud of myself for being able to unearth this based on such a fragment of a clipping - no title, no author, no date. But I was curious if the search on the WSJ website would have found the man's name. I tried it, and it didn't - until I remembered to expand the time frame to 2 years, and then it did.

I also found success searching on the phrases "As many as 4.62 points" and "token of support from the community" which were in the article. At this point, my pride dissipated, as I realized I had just picked all the wrong keywords from the start - making what should have been a 1-2 minute search unnecessarily long. Luckily it didn't matter in this case, as the patron was still around - but next time I'll just start with random phrases as keywords and see how it works.




Tags: , , , , , , , , ,


6 Responses to “Reference Question of the Week – 5/22/11”

  1. Kniffler Says:

    If you have a sample of the text you’re looking for, searching for sample low-probability phrases is the best thing ever. It’s why the ability to do a full-text search is exponentially better than just a keyword search.

    It’s especially useful for sourcing things where the thing they will be indexed by is the thing you’re trying to identify: mystery song lyrics, unsourced screenshots, unattributed quotations, etc.

  2. Brandy Stillman Says:

    Kniffler I agree. I got the idea from Amazon’s “statistically improbable phrases.” I have found many a song from the unique bits of lyrics, and caught several plagiarizers too.

  3. Brian Herzog Says:

    @Kniffler & @Brandy: I agree, and I’m actually a bit worried that this is going to turn into some personal challenge: how obscure of a phrase can I use and still find the right result? Hmm.

    But I guess this is one drawback to Google Books and the idea of everything-online: the bigger the pool gets, the fewer unique phrases there will be – especially with all the historical texts, and the millions of modern articles and blog posts and tweets, etc. Maybe “statistically improbable phrases” is just a window we can enjoy right now, but which will be replaced with something else once the improbability of any given phrase levels out due to sheer volume.

  4. sli Says:

    wow, was the patron really THAT rude?

  5. sfer Says:

    … which goes to show that sometimes my husband’s theory (he’s also a librarian) is right: never trust patrons. They will CATEGORICALLY AFFIRM that they ARE 100% SURE that the author’s name is such and such, or the title such and such, or the publisher such and such, or the date such and such. No matter how sure they are, they might be making a mistake… and you might be losing your time just because you believe blindly on what they say. Dr. House is right: people lie all the time, even when they don’t know they are!!

  6. Brian Herzog Says:

    @sli: unfortunately, yes. Worse still, she’s one of our regulars. She’s usually very appreciative after you find whatever she’s looking for, but fairly brusque with the initial question.

    @sfer: very true – one thing I learned repeatedly in library school is that you need to listen to what the patron means, despite what they’re saying.