@MASTERSTHESIS{pgi2018008,
	author = "N. Wells",
	supervisor = "M. Roper",
	title = "Exploratory Study into the Relationship between Statistical Measures of Keyword Quality and Citation Rates for Open Access Academic Papers",
	school = "Department of Computer and Information Sciences, University of Strathclyde",
	year = "2017",
	abstract = "An exploratory study investigating the relationship between two different versions of a particular statistical measure of metadata quality applied to Author Keywords (Ochoa and Duval's QtInfo) and citation rates for Open Access papers relating to diabetes, recorded in the Scopus citation index. 


17,451 articles were harvested from the CORE.ac.uk Open Access aggregator and  citation counts and keywords  for  3,588 of these were taken from the Scopus citation index and matched to the records from CORE, and analysed using R.


A variety of linear, quasi-Poisson and Negative Binomial regression models were applied, and most notably a highly statistically significant (p = 0.000109 ) and somewhat predictive (R2 = 0.01075) positive relationship was found between the Qtintokw  measure and citation rates, outperforming the related, previously investigated measure of Author Keyword count by an order of magnitude both in terms of significance and variance described. 
",
}