9. Basic Statistical Undeterminacy Problem
Whites
Blacks
Pi
Ai
Bi
1 – Pi
Di – Ai
(1 – Di) – Bi
Total
1
Di
1 – Di
Pi = AiDi + Bi(1 – Di)
Voters
Non-Voters
10. How to solve an unsolvable problem?
• Surveys – sometimes unavaliable
–
Lack of resources
–
Unreliable responces
–
Ex post research (I do not see dead people)
• Statistical models
12. Model assumptions
• No changes in population
• (No spatial autocorrelation)
• A distribution of unobserved variables
corresponds to an a priori chosen distribution
14. Bayesian models
• Utilizing information on the distribution of
observed variables to better assume a
distribution of unobserved vatiables
15. What is it good for?
• When dealing with aggregate data
• Filling cells in a contingancy table
A
B
C
?
?
?
α
?
?
?
β
?
?
?
γ
16. Examples in political science
• Social, economic and demographic
characteristics of a population (e.g. an
electorate of a political party)
• Voter transitions between elections
• Ticket-splitting
18. Research question
• What the hell is going on?
• The process of party system structuralization
in Central Europe in 1990s and 2000s
• Central European party systems appear to be
unstructuralized but there might be a
„hidden“ structure
• „Parties come and go but the structure
remains the same (?)“
19. Research question
• Is there a stable underlying structure of
electoral behaviour in Central Europe?
• The question is whether voters continually
vote for political parties that occupy relatively
similar position in a political space
20. The thesis
• V4 countries
• Parliamentary elections since early 1990s
• Voter transition rates between political parties
in pairs of consequential elections
22. A conceptualization of political space
An old leftist
party
An old leftist
party
An old rightist
party
An old rightist
party
A new
leftist
party
A new
leftist
party
An old
rightist
party
An old
rightist
party
23. Model A
An old leftist
party
An old leftist
party
An old rightist
party
An old rightist
party
A new
leftist
party
A new
leftist
party
An old
rightist
party
An old
rightist
party
24. Model B
An old leftist
party
An old leftist
party
An old rightist
party
A new
leftist
party
A new
leftist
party
An old
rightist
party
1
An old rightist
party
An old
rightist
party
25. Model C
An old leftist
party
An old leftist
party
An old rightist
party
An old rightist
party
A new
leftist
party
A new
leftist
party
An old
rightist
party
An old
rightist
party
26. Model D
An old leftist
party
An old leftist
party
An old rightist
party
An old rightist
party
A new
leftist
party
A new
leftist
party
An old
rightist
party
An old
rightist
party
27. Operationalization
• How to operationalize ideology?
• How to operationalize „new“ and „old“?
• Are these two concepts sufficient to describe
a political space?
28. Possible results
• „Crazy 1990s“ and „boring 2000s“?
– Lack of structure in the 1990s and stability in
2000s
• Why?
– Relative unexperience of voters
– „Crystalization“ of parties‘ ideological profiles