What is individual representation?

Political representation is usually understood to be about groups: left-wing parties represent people with left-wing views (“substantive” representation), female MP represent women in the population (“descriptive” representation). When Matt Golder and Jacek Stramski examined ideological congruence, that’s what they did, too. They included a footnote to outline the possibility that individuals are represented by a parliament as a whole, but seemed to dismiss this perspective as irrelevant in practice — this was the same comment I got from an examiner once when still working on my doctorate. This did not discourage me to write up how to conceptualize and measure individual representation in 2012.

Let’s begin with the different representational relationships:


Here, Z’ is an individual citizen, and R’ and individual representative (parliamentarian). ZZ refers to citizens as a group, and RR to representatives as a group.This gives us:

    dyadic representation: how well does a single MP represent citizens, e.g. how well does the MP of a district represent the views of the citizens in this district?
    collective representation: how well do parliamentarians collectively represent citizens, e.g. how well are the views of the citizens in a country represented in their legislature?
    direct representation: how well does a single MP represent a single citizen, e.g. does the president of a country represent the views of this particular person?
    individual representation: how well do parliamentarians collectively represent a single citizen, e.g. how well are the views of this particular person represented by their legislature?

Individual representation is expressed as a score for each individual. Rather than combining views and looking at averages or distributions, we consider the position of an individual citizen. To keep the information of all citizens, we first compare the position of each citizens vis-à-vis the positions of the other citizens. This gives us a measure of “marginality”. The intuition is that individuals are aware of their position among citizens to some degree, and a person with far-left views will not expect to have his or her views represented in a legislature if everyone else is centre-right or far-right. Similarly, a person with centrist views will expect his or her views represented in a legislature when many other citizens have centrist views. This is the comparison at the top of the figure.


Once we have figured out how common or marginal each citizen’s position is in the population, we do a thought experiment, and look at how common or marginal each citizen’s position would be among the representatives. This is the comparison at the bottom of the figure.

We can then simply compare (=subtract) the marginality among citizens and the (imagined) marginality among representatives to derive an individual representation score. Normatively, we assume that smaller Euclidian distances are “better”. This is an important note, because we could also assume that an individual only cares about having “perfect” direct representation through one MP. When moving beyond single issues, I think most individuals will cease to have a “perfect” direct representative who shares their views on all dimensions exactly, and considerations of minimizing positional distance will play a role.

Why should we care about individual representation? Because individual representation allows us to examine different aspects of representation than the other perspectives. For instance, individual representation is very flexible for exploring how membership in different socio-demographic groups (“intersectionality”) affects substantive representation. We can readily compare the level of representation of old women with that of left-wing men (if this makes sense, of course), or test whether individuals who are better represented also feel better represented – rather than make the assumption.

Golder, M., and J. Stramski. 2010. “Ideological Congruence and Electoral Institutions.” American Journal of Political Science 54 (1): 90–106.

Ruedin, Didier. 2012. “Individual representation: A different approach to political representation.” Representation 48(1): 115–29. doi:10.1080/00344893.2012.653248

Polrep package for R to calculated individual representation scores: http://polrep.r-forge.r-project.org/irs.html

Legislative Power and the Representation of Ethnic Groups

In their 2014 article, Leslie Schwindt-Bayer and Peverill Squire show that the political power of legislatures can affect gender representativeness of legislatures. In the article they discuss likely mechanisms and suggests that the same result applies to ethnic groups. The argument is that in a legislature with more professional power, need to provide representatives with incentives to compensate for their investments like long sessions. These incentives, in turn, encourage incumbents to preserve their seats and discriminate against under-represented groups. Sounds reasonable enough, but ever since collecting information on the ethnic composition of legislatures worldwide, I have been keen to empirically check such claims.

I did so using the spreadsheet from the DICE Database and my own data on ethnic representation. This gives me 35 countries to have a quick look at the claim: there is no such correlation among the countries examined.


Ruedin, Didier. 2009. ‘Ethnic Group Representation in a Cross-National Comparison’. The Journal of Legislative Studies 15 (4): 335–54. doi:10.1080/13572330903302448.

———. 2010. ‘The Relationship between Levels of Gender and Ethnic Group Representation’. Studies in Ethnicity and Nationalism 10 (2): 92–106. doi:10.1111/j.1754-9469.2010.01066.x.

———. 2013. Why Aren’t They There? The Political Representation of Women, Ethnic Groups and Issue Positions in Legislatures. Colchester: ECPR Press.

Schwindt-Bayer, Leslie, and Peverill Squire. 2014. ‘Legislative Power and Women’s Representation’. Politics & Gender 10 (4): 622–658. doi:10.1017/S1743923X14000440.

Mapping Ethnic Representation Scores in R

Here’s a demonstration of how easy it is to map with the R package rworldmap by Andy South. I map the ethnic representation scores in my 2009 JLS article, available from my Dataverse. I used the tab-delimited file, which contains the country name, Q-scores, R-scores, and a binary indicator of minority presence. I searched&replaced the tabs with commas, and added a new column for the ISO3 country codes.

After that, it’s just a few lines in R:

dta <- read.csv("Representation.csv") # these are the data described above.

> head(dta) gives:
Country ISO3 QScore RScore Present
1 Afghanistan AFG 0.803 0.697 1
2 Albania ALB 0.980 0.510 1
3 Algeria DZA NA NA 1
4 Andorra AND 0.980 0.000 0
5 Angola AGO NA NA 1
6 Antigua & Barbuda ATG 0.910 0.000 0

Next we have to identify the countries. joinCode specifies that I used ISO3, nameJoinColumn specifies the variable with the country abbreviations:

jcd <- joinCountryData2Map(dta, joinCode="ISO3", nameJoinColumn="ISO3")

Next a line from the package vignette that makes the plot use the space available.

Now, while mapCountryData(jcd, nameColumnToPlot="QScore") would suffice to draw a map, I used some of the options available (e.g. a blue ocean, light grey for missing data), and drew the legend separately for a little extra control:

mapParams <- mapCountryData(jcd, nameColumnToPlot="QScore", addLegend=FALSE, mapTitle="Ethnic Representation Scores", oceanCol="light blue", missingCountryCol="light grey")
do.call(addMapLegend, c(mapParams, legendWidth=0.5, legendMar = 4))

The title is a bit off, but other than that, I’m pretty happy for a first cut with so little coding.

Ruedin, Didier. 2009. ‘Ethnic Group Representation in a Cross-National Comparison.’ The Journal of Legislative Studies 15 (4): 335–54. doi:10.1080/13572330903302448.

How (Not) to Study Ideological Representation

David Broockman has an important paper on political representation apparently forthcoming in LSQ.

He notes two ways to study the political representation of issues, policies, and preferences. On the one hand we can examine citizen-elite congruence issue by issue. On the other hand, we can calculate “policy scores” to capture ideal points of overall ideologies and compare these between citizens and the elite. The paper convincingly demonstrates that the latter approach is flawed in the sense that it doesn’t really capture political representation in the way we generally understand it.

Broockman, David E. 2015. “Approaches to Studying Policy Representation.” Legislative Studies Quarterly.

New Publication: The Gap between Public Preferences and Policies on Immigration

Just days after announcing the “SOM book” (Politicization of Immigration), I have the pleasure to announce another product from the SOM project: The Gap between Public Preferences and Policies on Immigration: A Comparative Examination of the Effect of Politicisation on Policy Congruence in JEMS. In this paper Laura Morales, Jean-Benoit Pilet and I examine the purported gap between (restrictive) public opinion on immigration and (expansive) policies by the elite.
Using data from the SOM project and a range of public opinion polls, we consider the situation across seven countries and 15 years (1995 to 2010). This provides a better insight in what is one of the most salient policy domains in contemporary Europe than was done previously. There is no evidence that strong anti-immigrant parties have anything to do with differences between public opinion and elite policies. Just like what I found in my monograph on political representation, it turns out that salience plays a key role. When negative attitudes in the population are combined with extensive media coverage, we observe high levels of policy congruence.