Time series analysis of publication counts of a university: what are the implications?

Abstract

Yearly publication counts of research institutions and universities continue to be a widely-used parameter to assess their research productivity, and such evaluations have been successfully used to analyze the influence of research support policies at various levels. This study was designed to analyze the yearly number of articles having an Akdeniz University address and that appeared in the Web of Science databases from 1996 to 2009. Time series analysis of the number of published articles was used to determine the impact of alterations in the number of faculty members and research funding as well as changes in the institutional and country-wide research support policies and encouragement mechanisms. It was observed that alterations in both the number of faculty members who are active in research and the total amount of research funding each year may explain the general pattern published articles. However, there is a period with significant deviations from the trend predicted by these relationships. This period, corresponding to the years 2002---2008, is discussed in terms of the effects of policy changes which may have positive and negative contributions to the predicted pattern. Mathematical analysis of publication time series, together with parameters expected to affect research output, may provide valuable insight into the effectiveness of research support mechanisms.