Response to Watson R (2011) Commentary on Oermann MH, Shaw‐Kokot J, Knafl GJ & Dowell J (2010) Dissemination of research into clinical nursing literature. Journal of Clinical Nursing19, 3435–3442. Journal of Clinical Nursing20, 595–596

Abstract

The authors would like to thank Watson (2011) for his insightful remarks on our paper (Oermann et al. 2010). Watson (2011) highlights two relevant issues from our research. The first issue is that of the tension between citation analysis and author impact. Of the two, citation analysis has the longest history and appears more widely accepted. Author impact, the h-index, is still developing. The second issue that Watson identifies involves citation patterns, which is getting increased attention in the literature. Our citation analysis examined the assimilation of selected nursing research articles into clinical literature. The articles selected for analysis were published between 1990–2000 and tracked from publication to 2009. During this period, journals moved from print only to print and electronic format. A 2010 study examined how electronic journal collections were reflected in citation patterns of authors in dentistry, nursing and pharmacy at a large urban university (De Groote & Barrett 2010). From 1996–2008, the number of citations to articles in print only journals in nursing decreased while the number of citations to papers available online increased. The authors suggested improved linking of full text articles directly into electronic databases possibly contributed to the increase. Based on De Groote and Barrett’s study, nursing faculty are citing papers published online more than those published only in print. In response to Watson (2011), we reanalysed the citations in our study to determine whether there was an increase in the number of citations after the journals began publishing in both print and electronic formats. The number of citations and time in years over the periods when the journals were available in print only and both print and electronic formats were analysed. Poisson regression was used to model numbers of citations as they are counts while converting them to citation rates per year (using the appropriate offset variable). Generalised estimating equations (GEE) techniques were used to model the correlation for an article’s citation rate before and after its journal began publishing papers electronically as well as in print. The GEE estimated intra-article correlation was high at 0·65, indicating that an article’s print citation rate was highly correlated with its print plus electronic citation rate. The mean citation rate increased significantly (p = 0·011) from its estimated print only value of 1·5 citations per year to its print plus electronic citation rate of 2·4 citations per year. Another factor influencing electronic access is the linking of electronic articles to databases. Some database vendors negotiate directly with publishers to create links. For example, in 1995, the Cumulative Index to Nursing and Allied Health Literature (CINAHL) began providing direct full text links to selected journals. There are currently over 700 linked journals. Another option, software such as Serials Solutions OpenURL linking, goes beyond a single database to include linking in all databases subscribed to by the library or institution. While there are many reasons authors cite other studies, we suggest that electronic access to articles influences citation patterns. This area deserves more analysis.