Calculating the total vote where the district magnitude is greater than one: A test of some algorithms using British local election data

Electoral analysis using aggregate data relies on the availability of accurate voting statistics. One vital piece of information often missing from official electoral returns, particularly British local government elections, is the total number of valid ballot papers. This figure is essential for the calculation of electoral turnout. When voters have a single vote and official information about the number of ballot papers issued is missing, a figure for the total vote can still be derived. However; local elections in Britain frequently use a system of multiple-member wards, where voters have as many votes as there are seats to be filled. in such cases, calculating the total vote and, hence, the turnout does present a real problem. It cannot be assumed that all voters win use their full quota of votes or that voters will cast a ballot in favour of a single party. This paper develops and tests different algorithms far calculating the total vote in such circumstances. We conclude that the accuracy of an algorithm is closely related to the structure of party competition. The findings of this paper have a number of important implications. First, the difficulties in calculating the turnout in multiple- member wards are identified. This will inform the debate about public participation in the local electoral process. Second, the method for deriving a figure for the total vote has an important bearing on a number of other statistics widely employed in electoral analysis.