Tuesday, November 23, 2010

B - Positive-outcome bias in peer review

Emerson GB, Warme WJ, Wolf FM, et al. Testing for the presence of positive-outcome bias in peer review: a randomized controlled trial. Archives of Internal Medicine 2010;170(21):1934-1939. (doi:10.1001/archinternmed.2010.406)

Two versions of a randomized controlled trial that differed only in the way the main finding was described (positive finding or no difference) were peer reviewed by 210 reviewers of two journals (The Journal of Bone and Joint Surgery and Clinical Orthopaedics and Related Research). Three forms of positive-outcome were observed: reviewers were significantly more likely to recommend the positive version for publication; they detected more errors in the no-difference version; and they awarded higher methods scores to the positive version, even though the two versions had identical methods sections.

Monday, November 22, 2010

B - A CV of failures

Stefan M. A CV of failures. Nature 2010;468(467)
(doi: 10.1038/nj7322-467a)

The CV of a scientist does not mention his failed exams, unsuccessful fellowship applications, rejected projects or papers never accepted for publication. The author suggests to compile an "alternative" CV of failures, that could include every rejected application, project proposal and paper. Keeping it visible has two purposes: to remind each scientist of his own setbacks and to help other colleagues to shake off a rejection and start again.

Tuesday, November 16, 2010

B - Implementing open access

Armbruster C. Implementing open access: policy case studies. October 14, 2010. Social Science Research Network. Available at: http://ssrn.com/abstract=1685855

We are approaching the end of the first generation of open access implementation. This report evaluates progress by focusing on a small number of cases, including the University of Zurich, the Wellcome Trust, UK PubMedCentral, SCOAP3, the Howard Hughes Medical Institute, and the Austrian Science Fund. The impact of open access on digital scholarship is examined, with suggestions on what we can learn from the cases examined.

N - Preserving data

Scientific data can be lost because of the storage medium (fragile or obsolete) or because researchers weren't aware of its value. Nature News reports on an attempt to reduce the risk of data loss, launched to coincide with the biennial conference of the Committee on Data for Science and Technology in Stellenbosch, South Africa. The project is a global inventory of threatened data and was instigated by Elizabeth Griffin, an astronomer at the Herzberg Institute of Astrophysics in Victoria, Canada, and William Anderson, an information specialist at the University of Texas at Austin and an associate editor of the Data Science Journal. A task group hopes to catalogue data by the simple technique of asking scientists what they've got in their vaults.

Monday, November 15, 2010

B - Science publishing: whose intellectual property?

Salo D. Who owns our work? Serials 2010;23:191-195. (DOI: 10.1629/23191)
Intellectual property in scholarly communication is becoming increasingly complex, and research is becoming more collaborative and innovative. As a result, authorship and ownership criteria are being challenged, while institutions, funding bodies, and libraries are emerging as stakeholders in the publishing process. This article looks at where publishers fit in all this.

N - How much does peer review cost?

UK academics spend 2-3 million hours a year reviewing each other's work, according to a report by the Joint Information Systems Committee (JISC) Collections. The cost of this time to UK universities is estimated at £110-165 million a year, in addition to the £30 million spent on editors and editorial boards, and the £110 million spent on journal subscriptions. The report urges publishers to recognise the contribution that universities make to the publishing process and to see universities as partners rather than customers.

B - Authorship and industry support

Pollock RE, Ewer MS. The integrity of authorship: doing the right thing. Cancer 2010;116(17):3986-3987. (DOI: 10.1002/cncr.25268)

This article examines the balance between industry support and integrity of authorship. All articles submitted to peer-reviewed journals should be accompanied by full acknowledgement of industry-financed contributions, so that editors and readers can be clear about any relationship that could influence objectivity.

Sunday, November 14, 2010

B – Classification of biases in medical research

Chavalariasab D, Ioannidis JPA. Science mapping analysis characterizes 235 biases in biomedical research. Journal of Clinical Epidemiology 2010;63(11):1205-1215. (doi:10.1016/j.jclinepi.2009.12.011)

There are many different types of bias in medical research and publishing. This systematic mapping analysis of over 17 million articles from PubMed found 235 bias terms and 103 other terms used commonly in articles about bias. Forty terms were used in the title or abstract of more than 100 articles each. Clusters of terms were organized into macroscopic maps showing the distribution of bias types. Some bias tems (e.g. confounding, selection bias, response bias, publication bias) show increased use over time.