Below you will find answers to some of the most commonly asked questions when using the new federated search service – SuperSearch.
FAQs last updated 26th March 2008
If you have any other questions about using SuperSearch please email me.
Will databases which are not currently cross-searchable (e.g. Swetswise) be included in future? If so, when will this happen?
Serials Solutions merged with another federated search company, called WebFeat, in February. WebFeat have their own library of connectors, and it has recently been confirmed that this will be made available to Serials Solutions customers. Our customer services representative cannot tell us exactly when it will be available but an announcement is due in the next few weeks to provide more information.
The two connectors which we had previously commissioned Serials Solutions to ‘fast-track’ for development were British Education Index on Dialog and Swetswise. Both of these should become available via WebFeat’s connector library.
How are results returned from the selected databases? How does SuperSearch choose which records to display?
Rather that retrieving all the results from all the databases at the same time, which would slow the system down, SuperSearch initially retrieves between 20 and 50 results from each resource being searched, which it combines into a single set and sorts into reverse date order.
Which 20-50 results are retrieved from each resource will depend on how that resource returns its results by default. For example, if the native interface displays results in reverse date order as a default, SuperSearch will retrieve most recent. If the native interface displays results in “relevance” order, then SuperSearch retrieves the most “relevant” etc.
The speed with which each resource returns its results doesn’t affect whether its results are included in SuperSearch’s initial result set, as it waits for each resource to respond before combining, deduplicating and sorting the results into date order. A resource would only be excluded if it returns zero results, if it times out completely, or if the search fails due to a system error.
This has an impact on the way SuperSearch balances small and large databases. Imagine Resource A is a core database for the subject and has 1000 results for a search, and Resource B is marginal to the subject and only has 5 results. Only some of the results from Resource A will be included in the first set whereas all results from Resource B will be displayed, even if they are less “relevant” than the results from Resource A which haven’t been included. You may want to bear this in mind when deciding whether to include marginal databases in your subject categories. On the plus side, this also means that large numbers of results from big databases like Business Source Premier doesn’t lead to the exclusion of results from smaller databases.
How much can we customise the interface? What can’t we change?
We can customise:
-
the header of the SuperSearch pages (above the Basic/Advanced Search tabs) and the footer of the pages (beneath the blue line)
-
the names of the subject categories and the databases included in each category
-
the database names and descriptions
-
the text behind the “How to Search” link, and what the link is called
-
whether SuperSearch defaults to Basic or Advanced Search, and to Search by Subject or Search by Database
-
which search types are displayed (eg title, author, full text, keyword) and which is the default
-
the wording of error messages within SuperSearch
We cannot customise:
-
the search functionality itself
-
the clustering functionality
-
the format of the results list. There are some options which can be switched on or off, for example we can choose whether to display or hide the summary of results by database by default. However, we do not have completely free choice as these elements are core to the system.
Serials Solutions also offer an XML API which would allow additional programmatic control over SuperSearch. However, this requires the involvement of software developers. The WOLF team are currently using the API to develop a SuperSearch tool for WOLF module.
Is there any option to display/hide all abstracts?
Unfortunately at this time the option to display/hide all abstracts isn’t available. However this request has been forwarded to the Serial Solutions development team to look at as an option.
Why can’t I export to EndNote?
Although SuperSearch does have the option to export to EndNote via RIS format files, references are imported into EndNote as a generic reference type, as the database connectors do not provide type or genre information. Two consequences are:
- Journal title goes into the “Alternative Title” field
- Publisher details are not exported with book
The EndNote option has therefore been removed at the request of some librarians.
Serials Solutions are developing one-touch export direct into EndNote Web, which will make export to EndNote easier although won’t necessarily help with the data accuracy. However, one-touch export could be included in Article Linker, so single citations could be exported more reliably. If the University experiments with RefWorks, we can also test Serials Solutions’ export to RefWorks, which is better developed than EndNote.
Is it possible to limit searches to full-text?
SuperSearch would only know if a result was full-text if there was some kind of flag in the metadata returned by the source database to indicate it. Since source databases do not reliably indicate full-text in this way, SuperSearch is unable to filter results in order to limit them to full-text. With some resources, such as Wiley Interscience, it isn’t possible to limit to full-text even in the native interface, and federated search engines can’t do things, which the native interface can’t do.
However, one option would be to include the phrase “Full text” in the database description where relevant, in the same way that we currently indicate, “Not cross searchable”. The user could then choose to search only databases that are majority full text. One possible disadvantage would be that, by choosing to search only full text resources, users may neglect abstracting databases or databases containing a mixture of abstracts and full-text. On the other hand, it would alert users to resources most likely to provide full text results.
The Article Linker buttons and links in both SuperSearch itself and the native interfaces will also help users navigate through to full text where it is available through any of our subscriptions, and directs them to search the catalogue for a print copy if no electronic copy is available.
Can I search all of the cross-searchable databases in one search form?
Yes, you can select all the cross-searchable databases by selecting ‘Select All’. But this is not recommended as it may take some time before the first set of results is returned.
If you already know the best databases to search for the subject, you can select them from the ‘Search by database’ list, otherwise it is recommended that you narrow your search by subject first. This will automatically narrow the databases searched to those considered by the subject librarian to be the most relevant to the subject.
Why use SuperSearch?
If search results are not as comprehensive as results obtained through the individual databases, why should I recommend Supersearch to students/staff?
- Supersearch is a good place to find material quickly, however it’s not meant to replace more complex searching of the individual databases.
- You may wish to use Supersearch to help users determine which of the databases will produce the most relevant results for in-depth searching through the native interface.
-
It could be marketed to students as direct academic replacement for Google searching.
-
For LC staff, SuperSearch can be used for checking of availability of Inter Library loan requests. (searching OPAC, Scholar Google and electronic databases in one search.
How does the ‘Clustered Results’ work?
Clustering in SuperSearch dynamically scans the results of each search as they are displayed, building separate linked lists of identified authors, journal titles, subjects and data ranges, which in turn quickly identify underlying patterns, themes and trends. Suggesting further refinements and new ideas for the user’s search query.
Email and Export options are greyed out on the results page.
You may not have selected any results to export yet – click on check box by the side of results or tick “select all records” at the top of your search results. 
Export/Email options should now be available
Items selected for export will stay selected while navigating or downloading new pages of results. However you will need to export your selected items before refining or starting a new search.
Why do only some records display “More full text options”?
“More full text options” only appears when the reference contains a title or ISSN, plus volume and issue. If this metadata is not returned by the source database, the system does not have enough information to generate an OpenURL and so the Article Linker screens cannot be generated either.
Can I search for items published during a specified range of dates e.g. last 10 years?
Unfortunately at present you can only specify a single year in the advanced search form. Not all databases offer the ability to search by year range, and in those cases the request is simply ignored by the vendor, resulting in any relevant articles being returned regardless of published date.
However once your results list is displayed you can group the results by date. In the ‘Clustered Results’ panel (left) Click on the ‘Date’ tab, and your results will be grouped by year. Clicking on one of these categories will show only results from the chosen year(s).
When I type a phrase the AND operator is added between keywords
When searching for a phrase enclose search terms with quotation marks; otherwise words will be treated as separate search terms.
Both the basic and advanced search fields allow Boolean and phrase searching.
My search results show that there are thousands of results returned; yet only a small number of those result pages are listed.
Because of the large number of results retrieved, results are only returned in small batches to increase response times. When you reach the final page of results, you can click on the link 
to see the next batch of results retrieved. (top right hand side of results) The new set will be appended to the original results set or you have the option to merge the next set of results with your original results list. You will need to follow this procedure for each new download of results.
