STUDY 3: By-Election Research (PHASE 1)
Abstract
One notable gap in relation to this series is monitoring dynamics and developing a more nuanced understanding of context in relation to municipal by-elections. To this end, this short concept note proposes pulling together three strands of information, two of which already exist and the third which is suggested as possible new research undertaking by the Electoral Commission. This has been conceived as being undertaken in three distinct phases.
Phase 1: Development of a by-election database using voter participation (Zip-Zip machine) data
The voter???s roll in conjunction with the voter participation (Zip-Zip machine) data, which is generated on Election Day, is an invaluable source from which key analytical information can be derived regarding the basic characteristics of by-elections and how these are changing over time. This initial phase of research is intended to develop a database based on information that is derived from the Electoral Commission???s administrative data. Each record in the By-Election Database would contain basic attributes about that by-election, including the following:
??? Basic characteristics about the municipal ward council seat that has become vacant: name and number of ward, province, date of by-election, etc.
??? Number or registered voters, the number that participated in the by-election and the calculated turnout rate based on these figures.
??? The same registration, participation and turnout figures as above disaggregated by gender;
??? The same registration, participation and turnout figures as above disaggregated by age group / age cohort;
??? By-election outcome: reported in a non-partisan manner, this would ideally include two variables, namely (i) whether the municipal ward council seat is retained or lost to another party candidate; (ii) the margin of victory (percentage point difference in the share of the vote received between the winning party and the runner-up)
??? Other variables deemed of relevance once a review has been undertaken of the voter participation (Zip-Zip machine) data, as well as other specific contextual variables of interest that could be drawn from official statistics (area-level unemployment rates, education levels, etc.).
In developing the architecture of this database, it is proposed that the focus for this initial phase should be exclusively confined to those by-elections that have occurred since the 2016 Municipal Election. This is likely to be in the range of 30-50 by-elections. This would allow us to produce and refine the structure and content of the By-Election Database based on a manageable number of by-elections. It would also enable us to prepare a demographic profile of recent by-elections by examining turnout patterns in the context of by-elections and discern whether there are distinct variations in these patterns based on age, gender and geographic location. The by-election outcomes variables would be used to determine whether there are distinct turnout patterns in cases where a seat changed party hands or won/lost based on a significant or small margin.
Should it be approved by the Electoral Commission, this phase of research could be accommodated within the current 2022/23 financial year.
