Peer Evaluation in Parliament – Assessing the Parliamentary Performance of Belgian MPs

While there has been an increasing interest to study individual representatives’ behaviour in parliamentary democracies, most studies analysing their performance focus on the amount of formal parliamentary activities (e.g. Akirav 2016; Däubler et al. 2016; Papp 2018). However, parliamentary performance consists of more than the number of formal activities. Taking that into account it has for example become a tradition of Belgian newspapers to publish assessments of the work of MPs before elections. While some of these assessments aim to go beyond assessing the quantity of parliamentary work, they lack transparency and a broader scientific foundation. How could thus a broader concept of parliamentary performance (including e.g. the quality and the impact of parliamentary activities) be assessed? This paper aims at contributing to that debate by analysing implicitly underlying narratives regarding parliamentary performance in the political science literature and previous measurement approaches. Therefore, a theoretical scheme is developed to categorise narratives of individual parliamentarians’ performance. Moreover, a more encompassing framework to assess parliamentary performance of MPs is proposed looking not only at the amount but also at the quality and effectiveness of parliamentary work. Based on several theoretical dimensions of parliamentary performance a peer evaluation survey has been designed and forwarded to MPs of three Belgian parliaments (Federal, Flemish and Walloon). This article presents the potential benefits and pitfalls of the use of a peer evaluation survey to assess the parliamentary performance of individual MPs. The author argues that the methodology is particularly suitable for dimensions of parliamentary performance where a high amount of expert knowledge is required (e.g. assessing the relevance of parliamentary questions of individual MPs). Practical issues when conducting a MP peer evaluation survey are discussed and potential sources of non-response and item non-response are analysed. Furthermore, the peer evaluation survey data is aggregated with the help of Bayesian models controlling for several rater effects (e.g. party bias, gender bias, rater severity). Subsequently, the article analyses the relation between the aggregated peer evaluations with self-evaluations on several dimensions of parliamentary performance as well as potential sources of a positive bias for self-evaluations. Finally, the association of individual factors (e.g. gender, seniority, own performance of rater) with rater precision are examined.