CSA Internet Database Services (CSA-IDS) has become (for several reasons) the most innovative of the online information services which host commercial databases. Its efforts  earned my kudos earlier (as well as some of my criticism for the very modest sorting and index browsing capabilities). CSA-IDS  brings out the best of those databases which are getting enhanced by cited references, such as PsycINFO, Sociological Abstracts, or the Worldwide Political Science Abstracts.

Including cited references is the newest  type of value-added information (except for the ISI citation databases which were born with this feature) provided by  indexing and abstracting services  – and a precious one.

If some of the cited references have an abstract in the database and they are hot-linked, then this feature is even more valuable as  users can find more information about the cited references by clicking on their links. The best online  services (such as Ovid with the CINAHL database) have been offering this feature for some years, but CSA-IDS has added a new dimension to it in 2003.

What is novel and laudable  in the CSA-IDS software is that it adds one more piece of information to an increasing number  of  cited references listed in the record, the so called “cited by” value. It can be very informative as the more often an item which appears among the cited references is cited, the more relevant it is likely to be for the topic being searched. (There are, of course exceptions, such as books about general foundation, principles and/or research methodologies and tools which are the most cited, but topically are not closely related to the subject of the article). 

It is apparent at first glance that cited item #4 is likely to be a promising  article not only because it has the highest absolute “cited by” value among the first five cited references but also because it received those 21 citations in a relatively short time  since it publication in 1996. Its citations in  the past 8 years results  in a much higher adjusted score per year than citation #6 which received its 33 citations since its publication in 1954. 

Clicking on the “cited by” hotlink will bring up the records which cited the first item on the cited reference list, the article about The Media and the Gulf War. The records of the citing articles in turn will show their own cited reference(s), which of course, include(s) the one you clicked on.

An other extra feature in CSA-IDS is that it counts and shows the records of the citing articles from some other citation enhanced databases in its stable (but not from PsycINFO which regretfully seems to prefer its not so splendid isolation). This may significantly extend the scope of cited and citing references, and is one of the reasons for the good “adjusted cited by” score of  item #4 in the original cited list of references, as it is often  cited by articles included in the SAGE Full-Text Collection of Politics and International Relations.

CSA-IDS  displays the “cited by” values prominently also on the short result list on the search result lists, again, facilitating the selection of the most cited (thus presumably the most relevant articles). This is not a big deal yet in databases which have only a small proportion of the records enhanced by references, because their “cited by” values are low at present.

However, in a database, such as PsycINFO where more than 210,000 records are enhanced by nearly 10 million citations, the “cited by” values appearing in the rightmost part of the record separators provide significantly different indicators for the citedness of articles. For fairness, ISI database records have this information but they are displayed only in the full records, not in the short one in the result list.

 In case of a larger result set, it would be useful to be able to sort the results by decreasing order of the “cited by” values. Although these values are not part of the records and may keep changing whenever new records are added to the database about articles which cite documents already in the database, sorting by these dynamically generated values are not much different than sorting by standard bibliographic data elements of the records.

A much bigger deal would be to sort by citedness value the cited references (if citations reach a level of, say, 50 or 100). Thee are  listed in alphabetic author order within the records, but re-sorting them would help users in selecting the most cited (and possibly most valuable) references. Below is the original, author alphabetic sequence of cited references in plain vanilla spreadsheet format for the prototype.

Below is the same list of cited references within a record, re-sorted by decreasing “cited by” values.

The ultimate option could be – let me  fantasize a bit here, after all once you get a hammer everything looks like a nail - to sort the result list and the list of cited references on the basis of adjusted “cited by” score. It is the quotient of  the number of citations an item received and the age of the cited item.  While the top two cited references in terms of absolute number of citations are still among the top three on the basis of the citedness quotient,  the second ranked book by Vrij with a 11.67 quotient may escape the users’ attention because it appears at the bottom in the original list of cited references sorted by author.  It is a secondary question how do you calculate the age of the document to minimize the handicap of those documents which have been published in the last month or quarter of the year, .

Of course, many users could not care less about cited references to start with. For others, the author, journal and/or currency of the documents may be more important but this great idea of CSA-IDS to display the “cited by” value of the source items, as well as of the cited items is worth to be deployed more. There will be more databases with cited references in the coming years (and many more cited and citing references),. Extending the measurement of citedness beyond the limits of a single database (as CSA-IDS has started to do it), just adds more value to the cited-by indicators, benefiting from the also growing interdisciplinary citation pattern –exactly as Eugene Garfield envisioned it.

Back to Cheers & Jeers 2003