This is the odds ratio and, very briefly, if this number is close to 1, the variable has no effect t

1
EU Developmental Referendums Knowledge, Turnout
Vote Choice
Prof. Richard Sinnott (University College
Dublin), Prof. Palle Svensson (Århus University)
Stephen Quinlan (University College Dublin)
FP6 CivicActive

• This is the odds ratio and, very briefly, if this
  number is close to 1, the variable has no effect
  the more the ExpB number is above one, the
  greater the effect of the variable on increasing
  abstention the more it is below 1, the greater
  the effect on reducing the level of abstention.
  The right-hand column in Table 3 shows the level
  of statistical significance (this should normally
  be below 0.05). The point of the present analysis
  is that, controlling for the effects of
  socio-demographic, attitudinal and political
  factors, an important range of communication,
  campaign and knowledge variables had substantial
  effects on turnout/abstention in the second Nice
  referendum.

Introduction EU developmental referendums deal
with treaty- based changes in European
integration. They are quite distinct from
accession referendums. A key difference is that
the decision in a developmental referendum
directly affects all member states and the
process of integration itself. Accession
referendums directly affect only the state
concerned. The purpose of this paper is to
examine aspects of peoples behaviour in
developmental referendums in the four member
states that have said no in such a referendum.
Because Ireland is the only member state holding
a referendum on the Treaty of Lisbon, particular
attention will be paid to that case.
Comparisons The four no
outcomes in developmental referendums occurred in
quite different public-opinion contexts and with
quite different underlying trajectories and
levels of support for membership (see Fig.1).
Because Denmark and Ireland have each held EU
developmental referendums on more occasions that
any other member state, they form a particularly
appropriate pair for comparison. The differences
in abstention and in the volatility of the yes
vote are striking (see Fig. 2 and 3).
Knowledge Knowledge plays a crucial role in
referendum behaviour. Judging by responses to
open-ended questions, lack of knowledge
contributed substantially to the very high level
of abstention in the first Nice referendum in
Ireland (see Table 1). It also contributed to the
no vote in that same referendum (Table 2) and
to the no vote in the Maastricht referendum in
Denmark in 1992 (Fig. 4).
Response to the
campaign What makes for an informed and
self-confident electorate? A comparison of the
two Nice referendums in Ireland suggests that the
campaign in the media and interpersonal
discussion are crucial. As Fig. 5 shows, there
was substantial improvement in both respects
between the two referendums. There was also
increased satisfaction with some official sources
of information (the governments White Paper and
the information disseminated by the Referendum
Commission). Multivariate analysis This allows
us the measure the effect of an individual
variable (for example, knowledge of the EU) while
holding the effects of all the other variables
constant. The essential information is contained
in the ExpB column (see Table 3).