Grow and let grow!
|
|
SHODHAKA Life Sciences Pvt. Ltd. |
|
|
||
Several bibliographic software have undergone changes of different nature during the period (2006 to 2010) of this comparative study. ReleMed, Ali Baba, eTBLAST, which were used extensively in the study, have now stopped working completely (ReleMed) or partially (Ali Baba, eTBLAST) or is not consistent in its efficiency (BioAsk). HubMed, which was not working for a long period (hence excluded earlier), is now functioning (and is included). PubMed and GoPubMed have introduced new useful features. The results of Google Scholar always seem to be in flux and several details of its functioning remain unknown to public. Nevertheless, the overall reliability of Google Scholar has not changed.
Checking the relevance of results to the current period
The initial detailed studies were conducted between 2006 and 2008, the results of which are provided in this article. Hence, it was necessary to investigate if the results are applicable even after multiple changes in the search engines, as indicated in the above section. Thus, the following searches were repeated in 2009 (June to October):
Round I:
("alternative promoter" OR "alternative promoters") AND (mouse OR mice OR mus)
Round II:
NMR studies on the quadruplex nucleic acid structures in the context of HIV.
Full text search engine efficiency testing:
Search to find commonly used primer designing programs.
The overall identification of the top performers did not change when the first round study was repeated. The precision was slightly better in the recent results compared to earlier ones. HighWire Press, showed a lower PPS compared to earlier scores, even though its recall score remained the same.
In round II, the comparative performances of PubMed, Scopus, Google Scholar, HighWire Press, EBIMed and CiteXplore reproduced the same pattern as earlier results. The relative rankings of the tools, as per the results from 2009 results, were identical to the results obtained in 2007/8. Hence, we concluded that the results from earlier studies are applicable even now, despite an increase in the number of publications as well as the changes in search engines.
Scirus performed much better in 2009 results of the first round search than its earlier (2007) performance. However, the tool was frequently inconsistent. Therefore, it was not directly included as one of the top performers. Nevertheless, it was re-assessed with a single query set used for qDNA, in the second round (June- Sept 2009). The results of the analysis further confirmed that Scirus did not contribute any advantage over any of the other selected tools for second round. In addition, we found that Scirus continued to show a lower performance compared to that of Google Scholar and HighWire Press in the 2009 full-text analysis (almost akin to the trend seen in the 2007 studies), on primer designing tools.
Round I:
("alternative promoter" OR "alternative promoters") AND (mouse OR mice OR mus)
Round II:
NMR studies on the quadruplex nucleic acid structures in the context of HIV.
Full text search engine efficiency testing:
Search to find commonly used primer designing programs.
The overall identification of the top performers did not change when the first round study was repeated. The precision was slightly better in the recent results compared to earlier ones. HighWire Press, showed a lower PPS compared to earlier scores, even though its recall score remained the same.
In round II, the comparative performances of PubMed, Scopus, Google Scholar, HighWire Press, EBIMed and CiteXplore reproduced the same pattern as earlier results. The relative rankings of the tools, as per the results from 2009 results, were identical to the results obtained in 2007/8. Hence, we concluded that the results from earlier studies are applicable even now, despite an increase in the number of publications as well as the changes in search engines.
Scirus performed much better in 2009 results of the first round search than its earlier (2007) performance. However, the tool was frequently inconsistent. Therefore, it was not directly included as one of the top performers. Nevertheless, it was re-assessed with a single query set used for qDNA, in the second round (June- Sept 2009). The results of the analysis further confirmed that Scirus did not contribute any advantage over any of the other selected tools for second round. In addition, we found that Scirus continued to show a lower performance compared to that of Google Scholar and HighWire Press in the 2009 full-text analysis (almost akin to the trend seen in the 2007 studies), on primer designing tools.
Checking other features of the search engines:
In addition, a new survey of the search engines was carried out in February 2010 to ensure that no major changes have occurred. The results were found to be consistent, despite minor changes in different tools. For example, Google Scholar has now enabled an additional category of ‘legal opinions and journals’ and has also increased the number of languages to search and retrieve citations (from 8 to10). Although the outlook of PubMed and HighWire Press has changed, the important findings of the study remained unaltered.
copyright © shodhaka life sciences private limited; all rights reserved