Research Essentials

 

Who am I?

Are these notes available?

  • Yes - they are on http://researchclinic.co.uk

This session covers the essentials you'll need for effective

  • searching

  • provenance checking

  • time travel!

  • Googling: What you want vs. what you search for

     

    • Use the "logic of inclusion" - what words will / might be on the page?

    • Pick the strongest words. Look out for the pitfalls (e.g. swedes vs. Swedes)

    Example: how does the world perceive Sweden

    You could type "global perception of Sweden", "international view of Sweden" or "Swedish culture" into Google - with bad results.

Google only looks for pages containing your list of words. It doesn't understand the words and it doesn't care what you want.

For a search to work, you have to imagine which words would be on a foreign page about Sweden.

You have to visualise and make educated guesses (even if you do use popular stereotypes).....

 

  • Keep an eye on the results and the number of "hits". This will tell you if you've picked a bad word or misspelled something.

  • Linking words in a phrase

    • If you know that two or more words will appear next to each other on a web page (i.e. in a popular phrase), use quotation marks around them.

    A search for danish crime rate produces millions of results - most of which are irrelevant. Google will suggest any page with those three words on, in any order...eg:

    "Arsenal fans will feel robbed of the title, however no crime has been committed - unless you count Danish footballer Nicklas Bendtner's criminally low scoring rate..."

    Putting quotation marks around the phrase ensures that Google only suggests pages where the three words appear next to each other, in that order.

    However, as that only returns three results, it may be better to pick another phrase or strategy - eg. "crime in denmark"

  • Boolean logic - AND, OR & NOT

    Boolean logic allows you to specify what you want....

    • this AND that

    • this OR that

    • this but NOT that

    in combination....for my lunch, I'd like

    • a salad AND crisps OR peanuts AND coke OR lemonade, but NO meat

  • Google & logical "AND"

     

    • Google defaults to a logical AND. You don't need plus signs etc.

e.g. sweden culture abba snow meatballs ikea volvo

We use this Logical "AND" when it is important that all the words appear on the page - in this case to find results that mention popular Swedish icons alongside the word "culture".

  • Google & logical "OR"

    • The word "OR" (in capital letters) goes between optional words or phrases. One or more of which must be in the search.

If we wanted to search for articles in Norwegian or English about Norway's co2 emissions and their effect on global warming.

We could use a logical "AND" search for:

co2 "global oppvarming" "global warming" "climate change" Oslo Trondheim Bergen Stavanger

This gives us 3 pages of links to pages that have all the specified words. The down side of "focussed" is "limited".

Perhaps there are some great pages that don't have "global warming" in both languages or the names of four different cities.

If we want to introduce some flexibility into our search we can use the Logical "OR" and search for:

co2 "global oppvarming" OR "global warming"
Oslo OR Trondheim OR Bergen OR Stavanger

In this search we have insisted that search results conform to three requirements

1. co2
We insist that co2 is mentioned because that is a vital point to our article.

2. "global oppvarming" OR "global warming"
We also insist that either phrase is mentioned because co2 is not specific enough to climate change. We allow ourselves either English or Norwegian articles. We could also add in "OR climate change" if needed.

3. Oslo OR Trondheim OR Bergen OR Stavanger

To help tilt the story towards Norwegian domestic subjects, we ask that at least one of the named cities is mentioned in the article.

Note: leaving aside the specifics of this story, this gives you a pretty useful "reusable Google search". If you add a few more cities and store the result page in your favourites, you can tailor it to search for anything related to a country or its cities.

  • Google & logical "NOT"

    • To stop Google suggesting a pages containing a certain word, put a minus sign directly in front of that word.

    • If the Olympics are getting in the way of your Lillehammer search (or if Candice is messing with your Bergen), minus them out!

 

  • Search for file types

    • PDFs often contain glossy reports

    • Excel spreadsheets often carry figures and statistics

    • Powerpoint lsides have charts and neatly sumarise subjects

    • Press offices sometimes use Word for their releases

    • Google can be forced to limit its results to PDF, XLS, PPT & DOC

    • The syntax is filetype:pdf

    • Use the html version for speed and ease of use

    example

    Find PDFs mentioning statoilhydro, its plant at Melkøya & co2

  • Search by domain

    • You can tell a lot about a site from its domain name

    • French sites often use the country code .fr

    • French government sites use gouv.fr

    • Harvard uses harvard.edu

    • Statoilhydro use a .com for International familiarity

    • Their English language portal is at statoilhydro.com/en

    • Google lets you specify the domain type and beyond

    example

    Find PDFs from Statoilhydro that mention Melkoya

Note: even if a website has removed a PDF, DOC etc., it may still show up on a Google search. Choose "view as html" if the live link doesn't work.

 

  • Provenance checking

    • It's extremely easy to fake a website

    • The BBC got caught out by a top search result on Google

    • If you find a website and need to check its authenticity, see what others are saying about it by running the domain name through a search engine

    • Check the website's registration details using a WHOIS search

  • Time travel

    • If a webpage has recently disappeared, you can bring it back by using the search engines "cache". This is the copy it saved on its database the last time it visited the page.

    • The link to the cache copy appears after the description of the site in the search engine's search results.

    • The cached version is often faster, shows why the link was suggested, can be set to "text only" and highlights the search terms.

    • If you need to travel further back in time, use the Wayback machine. Norsk Hydro's website from 10th December 1997:

     

  • Questions & Answers

     

     

    My next presentation will be on Saturday and will cover more advanced techniques and specialist tools for investigations