Skip to contents

A data set on party mobilization; illustrates hierarchical models

Source: R/rainey2015.R
rainey2015.Rd

Rainey (2015) uses survey data to examine how party mobilization varies across district competitiveness and electoral rules.

Usage

rainey2015

Format

A data frame with 5,126 individual-level observations from five national legislative elections and 5 variables:

election_id

name for the election (character). Usually country name, except Portugal, which has two elections indicated by an appended year.

district_id

election and district label. The district numbers are arbitrary IDs.

contacted

self-reported party contact during the last campaign (factor with levels "Not Contacted", "Contacted").

district_competitiveness

district-level competitiveness measure in \([0,1]\) based on Grofman–Selb thresholds for d’Hondt/SMD systems (numeric).

electoral_rules

electoral formula for the district (factor with levels "PR" and "SMD").

For details on measurement, modeling choices, and case selection, see the article and online appendix.

Source

Replication data and code: Harvard Dataverse, doi:10.7910/DVN/27666 .

Details

In Rainey (2015), the key claim is that "the (positive) marginal effect of a district's competitiveness on mobilization is greater under SMD rules than under PR rules." Alternatively, disproportional rules strengthen parties’ incentives to mobilize in competitive districts.

There are three versions: rainey2015 includes all the variables and observations below; finland includes election_id and contacted for Finland only; and uk includes election_id, district_competitiveness, and contacted for Great Britain only. These smaller datasets are useful for simpler problems.

References

Rainey, Carlisle. 2015. "Strategic mobilization: Why proportional representation decreases voter mobilization." Electoral Studies 37: 86–98. doi:10.1016/j.electstud.2014.10.008 .

Grofman, Bernard, and Peter Selb. 2009. "A fully general index of political competition." Electoral Studies 28(2): 291–296. doi:10.1016/j.electstud.2009.01.009 .

Examples

# load data
rainey <- crdata::rainey2015

# table of observations per election
table(rainey$election_id)
#> 
#>        Canada       Finland Great Britain Portugal 2002 Portugal 2005 
#>           670          1086           756           759          1855 

# fit logit model
f <- contacted ~ district_competitiveness * electoral_rules
fit <- glm(f, family = binomial, data = rainey)
summary(fit)
#> 
#> Call:
#> glm(formula = f, family = binomial, data = rainey)
#> 
#> Coefficients:
#>                                             Estimate Std. Error z value
#> (Intercept)                                  -0.9000     0.2413  -3.730
#> district_competitiveness                     -0.5691     0.3363  -1.692
#> electoral_rulesSMD                           -1.6448     0.4266  -3.856
#> district_competitiveness:electoral_rulesSMD   3.9960     0.6283   6.360
#>                                             Pr(>|z|)    
#> (Intercept)                                 0.000191 ***
#> district_competitiveness                    0.090626 .  
#> electoral_rulesSMD                          0.000115 ***
#> district_competitiveness:electoral_rulesSMD 2.01e-10 ***
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> (Dispersion parameter for binomial family taken to be 1)
#> 
#>     Null deviance: 5995.4  on 5125  degrees of freedom
#> Residual deviance: 5732.4  on 5122  degrees of freedom
#> AIC: 5740.4
#> 
#> Number of Fisher Scoring iterations: 4
#>