Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 9

MOEHLER, DEVRA AND STAFFAN LIBDBERG. 2009.

“NARROWING THE LEGITIMACY GAP: TURNOVERS


AS A CAUSE OF DEMOCRATIC CONSOLIDATION” THE
JOURNAL OF POLITICS, 71(4): 1448-1466.
Replication by Marco A. F. Araujo
Course: Multilevel Modeling (POS6933)
INTRODUCTION

 Purpose of the paper: investigate whether democratic electoral qualities can


reduce the difference between winners` and losers` trust in governmental
institutions.

 They use data from 3 rounds of Afrobarometer surveys, including 11 countries in


round 1, 15 countries in round 2, and 18 countries in round 3, in a total of 44
surveys.
MODELING STRATEGY

 This is a two-stage linear hierarchical model

 In the first stage, the dependent variable is regressed on independent variables of


level-1

 In the second stage, the intercepts and slopes of the main level-1 independent
variable is regressed on independent variables of level-2.
MODELING STRATEGY

 The first stage dependent variable: trust in governmental institutions


 The first stage independent variables: winner-loser status, gender, age, education,
urban residence, poverty, media exposure, political participation, and political
interest

 The second stage dependent variables: weighted intercept from the first stage
model, and weighted slope of winner-loser status.
 The second stage independent variables (democratic electoral qualities): elections
that produce turnovers (i.e. incumbents are electorally defeated by the opposition
candidates), elections that are peaceful, elections results that are accepted by the
political parties, and elections that are perceived as free and fair
FIRST STAGE REPLICATION

First stage replication was successful!


SECOND STAGE REPLICATION

TABLE 3: Estimates of the Second-step


Estimates Model A Model B
0.017 -0.035
Turnover of
(0.132) (0.019)
Power
0.899 0.089
0.244 0.006
Peaceful Process (0.189) (0.037)
0.215 0.867
0.097 -0.027
Opposition
(0.110) (0.043)
Acceptance
0.390 0.541
-0.086 0.019
Free and Fair (0.142) (0.073)
0.551 0.797
1.374 0.140
Constant (0.186) (0.088)
<0.01 0.131
R2 0.166 0.123
Prob>F 0.328 0.140
N 41 41

Second stage replication: turnover coefficient is the only one statistically significant as in Moehler and Lindberg (2009), however
at the 10% level. All the coefficient magnitudes differ. Data for 3 surveys are missing in the dataset.
EXTENSION

 The extended model looked for whether any of the four democratic electoral
qualities could reduce a “gender gap”.

 Gender gap is measured as the difference between women`s and men`s trust in
governmental institutions.
EXTENSION RESULTS
Using the Stata command “.lincom” to calculate the marginal
effects, after the main model:
TABLE 6: Marginal Effects of Gender*Turnover of Power
Gender Coefficient 95% Confidence Interval
Women (=0) 0.004 [ -0.147 ; 0.155 ]
(=1) 0.034 [ -0.115 ; 0.183 ]
(*) p≤ .1; (**) p≤ .05; (***) p≤ .01.

TABLE 7: Marginal Effects of Gender*Peaceful Process


Gender Coefficient 95% Confidence Interval
Women (=0) 0.127 [ -0.115 ; 0.370 ]
(=1) 0.157 [ -0.087 ; 0.402 ]
(*) p≤ .1; (**) p≤ .05; (***) p≤ .01.

TABLE 8: Marginal Effects of Gender*Opposition Acceptance


Gender Coefficient 95% Confidence Interval
Women (=0) 0.056 [ -0.136 ; 0.249 ]
(=1) 0.086 [ -0.109 ; 0.282 ]
(*) p≤ .1; (**) p≤ .05; (***) p≤ .01.

TABLE 9: Marginal Effects of Gender*Free and Fair Elections


Gender Coefficient 95% Confidence Interval
Women (=0) 0.012 [ -0.258 ; 0.283 ]
(=1) 0.042 [ -0.228 ; 0.313 ]
(*) p≤ .1; (**) p≤ .05; (***) p≤ .01.
CONCLUSIONS

 The replication results do not strongly support Moehler and Lindberg`s (2009)
findings

 The replication show that none of the four democratic electoral qualities
moderate the “gender gap”

You might also like