Evaluation and organization of search results

Arrangement of search results

Before you start examining the documents you have found, pay some attention to the way they are arranged:

  • chronologically (the newest documents are displayed first);
  • alphabetically;
  • according to their relevance (the system arranges the documents according to the frequency of appearance of search words, considering also the location of search words in the document – in the title, abstract or elsewhere).

Usually, it is possible to choose among different arrangements of documents.

Evaluation of the relevance of results

Relevance of the document is its value regarding the search profile.

The relevance of a retrieved document can be found by examining its title, key words, subject headings and abstract. If the document lacks an abstract, it is quite difficult to assess its relevance to your subject. Sometimes, relevance can be assessed only when the full text has been retrieved.

Naturally, you wish to get only relevant hits, but you will, unavoidably, get unnecessary results too. The precision of the results is expressed by the ratio between relevant hits and the total of hits.

At the same time, the aim of the search is to find all documents on a certain subject contained in the database so that none of them remains unfound. The term used in information science is recall, expressed by the ratio between relevant hits and the total of relevant documents in the database.

If your search retrieves all documents on your subject contained in the database, you may also get many irrelevant results that have to be manually weeded out.

If almost all retrieved documents are relevant, then, probably, the database contains even more relevant documents that you did not find. In building a search, it is essential to find the balance between these two possibilities.

If your query is too narrow, you will miss a number of relevant documents, but if your query is too wide, you will get a number of irrelevant hits too.

You have to consider the aim of your search. If you need to analyse your subject more thoroughly, the recall is important and you need to plan a wider search to find all important documents.

Too few (relevant) hits

  • Revise your search words to make sure you have considered all options.
  • Try to find additional search terms from the subject index of the database and use the key words index, if it exists.
  • Truncation of some words can be of help.
  • Check the subject headings added to relevant documents;
  • Important new search words may also be found in the abstracts or fulltexts of the documents.
  • Check the references added to relevant documents. Citation databases let you search for articles where some other article is cited. 
  • Think about other databases you might want to use.
  • If the number of hits is too small, try to drop some search words to make the search wider. E.g., if your subject is the influence of media violence on Estonian young people, then, dropping the word Estonia from your search may result in a much larger number of still relevant hits.

Too many hits

  • Use the proximity operator instead of AND.
  • Use quotation marks for phrase search.
  • Limit your search to the field of the title or abstract. This will yield more relevant results, as the important words that describe the content should be found already in the title or abstract.
  • Limit your search by publication dates.
  • Shift the truncation symbol farther to the right.
  • Use narrower terms.

Databases offer a link (RefineEditModify, etc.) for changing the query.

Exporting of search results from the database

You can use the following ways:

  • saving;
  • e-mailing;
  • printouts
  • export to reference managers, for example to RefWorks, EndNote Web, Zotero etc.

Saving of searches

Most databases allow the saving of your search in order to return to it later. This option can usually be found under Advanced Search. Find the link Search History and choose the search you want to save.

Alert service

Many databases offer the Alert service, meaning offering regular information about documents added to the database on the subject of your saved search. You have to choose this service when saving your search, specify how often and for how long a period you wish to get the information and add your e-mail address. You will receive e-mails telling you about new documents on your subject in the database. This way, you can easily keep up with new information and do not have to repeat your searches again and again.

Similar information can be subscribed for in the form of RSS information feed from many databases.

back