2
2004). Higher proportionality should also enhance the election of candidates from
minority or under-represented groups.
District magnitude is normally defined as ‘the number of seats filled at an election
in a district’ (Taagepera and Shugart 1989, 19; Farrell 2001, 6). But this definition does
not cover the case of multi-member NP systems where voters cast as many votes as there
are seats to be filled. To cover this contingency, a more precise definition should read
‘the number of seats filled at an election in a district divided by the number of candidates
electors may vote for’. By definition district magnitude is 1 in NP systems that either
elect a single member or elect as many members as electors have votes.
The existence of multi-member NP districts poses an awkward question: how do
people distinguish between a preferential vote in, say, a five-member STV district, and an
NP vote in a five member district with five votes to cast? In each case, electors can
reasonably expect to elect five candidates. Are they aware that STV means they are less
likely to waste votes by voting on an ordinal basis? If not, the number of candidates they
can vote for in a plurality election for may give them the impression that their votes will
be more effective than they are. In this case, the number of candidates that can be elected
per district could enhance turnout. However, because multi-member NP districts will not
deliver proportionality, one would expect lesser representation of minorities or under-
represented groups. District magnitude, in these terms, is an interaction of number of
candidates elected per district and the STV ordinal ballot structure. Standard electoral
system theory would predict no turnout or representational effects for number of
candidates, but significant effects for district magnitude. Representational effects,
however, may be more open to question. Higher women’s representation, in particular, is
Convention