A systematic review of micro-level articles on second-order theory and European Parliament elections from 1980 to 2009

Abstract

This research consists of a systematic review of 54 articles published on the second-order elections theory and European Parliament (EP) elections between 1980 and 2009. The second-order elections theory was launched in 1980 by Reif and Schmitt (1980), one year after the first EP elections. This theory holds that second-order elections have lower turnouts, that governing parties usually perform worse, as do larger parties, while smaller or newer and/or radical parties are favoured. The theory assumes two main mechanisms to explain such voting patterns, both of which stem from the assumption that, for parties and voters, there is less at stake in these elections. Because there is less at stake, voters would either choose to vote sincerely – choosing smaller parties closer to their ideologies – or choose to punish or send a message to the governing parties – especially when second-order elections are held in the middle of the electoral cycle. Although these assumptions have been corroborated with aggregate-level data over the last 40 years, some authors have called attention to the scarcity and delay in the development of robust tests of the theory with individual-level data. This review found that tests with this type of data had been carried out since the 80s and 90s. However, most of these first articles with individual-level data did not perform robust statistical tests, as well as did not impact the literature, obtaining a low number of citations. Only at the turn of the 20th-21st century did this flaw begin gaining more prominence on the research agenda, especially between 2007 and 2009. Among other findings, this review also found that there is no publication bias on the studied topic - at least until 2009 - and that most of the authors on the subject are male and affiliated to German, Dutch, British and North American universities. Finally, it is expected that this review will help to better understand the second-order elections theory and to systematize the knowledge accumulated in these first 30 years of research on the subject. We also expect that this review will contribute as an incentive to carry out more systematic reviews in the larger Social Sciences field, demonstrating the importance of this method for the advancement of scientific knowledge and publishing in detail the step by step of our research design, inspired by Cooper (2016) and Figueiredo Filho et al (2014).