RETRACTED ARTICLE: Predatory publishing in Scopus: evidence on cross-country differences

Abstract

Predatory publishing represents a major challenge to scholarly communication. This paper maps the infiltration of journals suspected of predatory practices into the citation database Scopus and examines cross-country differences in the propensity of scholars to publish in such journals. Using the names of "potential, possible, or probable" predatory journals and publishers on Beall's lists, we derived the ISSNs of 3,293 journals from Ulrichsweb and searched Scopus with them. 324 of journals that appear both in Beall's lists and Scopus with 164 thousand articles published over 2015-2017 were identified. Analysis of data for 172 countries in 4 fields of research indicates that there is a remarkable heterogeneity. In the most affected countries, including Kazakhstan and Indonesia, around 17% of articles fall into the predatory category, while some other countries have no predatory articles whatsoever. Countries with large research sectors at the medium level of economic development, especially in Asia and North Africa, tend to be most susceptible to predatory publishing. Arab, oil-rich and/or eastern countries also appear to be particularly vulnerable. Policymakers and stakeholders in these and other developing countries need to pay more attention to the quality of research evaluation.

Supplementary information: The online version contains supplementary material available at (10.1007/s11192-020-03852-4).

Keywords: Academic misconduct; Beall’s list; Open access; Predatory journal; Research policy; Research system.

Fig. 1

% of predatory journal articles in total articles, 2015–2017. Source: Scopus (2018a), author’s calculations

Fig. 2

Estimated effects of GDP per capita (upper figure) and size of the research sector (lower figure) on the propensity to predatory publishing (total excluding Frontiers), GLM with logit link for binomial family, 2015–2017. Based on results in Column 5 of Table 5. Predictive margins with 90% confidence intervals are displayed