|
|
|
|
|
CLE Dataset |
2007 Data Release (25 of 60 countries) |
2008/9 Data Release (25 of 60 countries) |
Frequently Asked Questions |
Data Links |
|
Who created the CLE dataset?
Dawn Brancati. She is an Assistant Professor in the political science department at Washington University in St. Louis.
Before coming to Washington University, Professor Brancati taught and held a post-doctoral fellowship at Harvard University and the Harvard-MIT Data Center, respectively.
Prior to Harvard, Professor Brancati was a fellow at the Center for the Study of Democratic Politics, Princeton University. She received her PhD from Columbia University.
How do I cite the CLE dataset??
If you use these data, please cite: Dawn Brancati. 2007. Constituency-Level Elections (CLE) Dataset. New York: New York.
http://www.cle.wustl.edu (The dataset is protected by copyright of the US Copyright Office.)
The data are for your personal use only. The data may not be transferred, copied or disseminated to third parties without prior written permission.
How did CLE originate?
The dataset began as part of my dissertation research at Columbia University and was initially housed in a single cardboard box crammed into the closet of my tiny NYC apartment.
It has since grown to 14 boxes and totals an estimated 280 pounds, which is more than an average newborn elephant.
I expanded the dataset exponentially while at Princeton University, as a fellow in the Center for the Study of Democratic Politics.
Now, the dataset is housed at Washington University, St. Louis through the Center for Empirical Research and Law, which is in the process of making the dataset for user-friendly.
Where do the election results in the CLE dataset come from?
With very few exceptions all the data in the CLE dataset are based on countries’ official election results, which I’ve amassed from government institutions
(e.g. electoral commissions, ministries of interior, statistical offices and legislative bodies).
In rare cases where official results are not available, the data are based on results published in national newspapers, or other reliable sources.
Why should I use constituency-level election data?
Constituency-level election data allows for more precise measures of many aspects of electoral systems, which vary at this level, and facilitates the analysis of different political issues that similarly vary on this level.
Why should I use the CLE dataset over other constituency-level datasets?
The CLE dataset, unlike other datasets, includes election results for all political parties than compete in an election.
It does not impose a threshold on parties as to the number of votes and seats that they have to win in order to be included in the dataset.
It also does not include an “other” category unless the official electoral data contained one, which is not very often.
The dataset also contains information for upper and lower houses, and in some cases, subnational elections as well, while most usually only provide data for lower houses.
The data, moreover, have been checked multiple times before being uploaded to ensure 100% accuracy. Finally, the depth and breadth of CLE’s post-WWII coverage is greater than other datasets.
Can I combine the CLE dataset with other datasets?
Maybe. If you want to combine the CLE dataset with other datasets, inspect the data you want to add very carefully to ensure it is consistent with the CLE dataset in terms of the level of information it contains.
Ex. Does this other dataset include information for all parties in an election or only parties that win seats?
Does the dataset contain large “other” categories? Are the data disaggregated according to tiers?
If it is not consistent, you may not be able to use the other data, or you may have to aggregate all of the data to the lowest common denominator.
How do I use the CLE dataset?
The data for each country is provided in a separate data file. Originally the data was only available in DTA format.
The dataset is now available in a wide range of formats, including .POR, .CSV, SAV, SPT and XLSX formats, making it more accessible to a wider demographic.
The type of legislature is represented in this file by the variable legid, where 1-2=national elections (1=lower house and 2=upper house) and 3-4=sub-national elections (3=lower house and 4 upper house). Parties are designated as p1, p2, p3, etc. so that data for different countries can be stacked on top of each other. The names of each party are defined in separate codebooks for each country, which also provide notes about the data. Parties are not grouped together in an “other” category unless the official results report an “other” category, which cannot be disaggregated. Wherever possible, the names of the parties are provided in English and the original language along with the party acronym. In many cases the original election data only indicates the parties’ acronym. Every effort was made to identify the full names of these political parties, but this was not always possible, especially in the case of very small parties that competed only briefly for political office. In other cases the original data did not identify parties consistently across years, changing languages, using different abbreviations, or using acronyms in some cases and not in others. Again, every effort was made to identify the parties completely and to define them consistently over time. Why isn’t there data for all elections held in some countries since 1945?
For some countries data were not available for the entire post-WWII period. Other countries did not hold democratic elections throughout this period. Only democratic elections are included in the CLE dataset.
Will the CLE dataset include more countries and elections in the future?
Yes! The CLE dataset is a work in progress and will include additional election results in the near future. Data for another 25 countries will be released in 2008 and even more data will be released subsequently.
To whom should I report problems with the data?
Please direct all problems/inquiries to Dawn M. Brancati brancati@wustl.edu. The data will be updated often so please check the website.
|
| Brancati Home Page | |
|
|
|