James Raymond Vreeland presents...

The Return of Goldilocks in...
Civil War and The Three Regimes!

by Jimmy Ray
(click here for the background story)

Once upon a time, Goldilocks asked her Three Bear friends,
"Bear friends, what kind of political regime is most likely to suffer from civil war?"

"Not autocracies!" shouted Papa Bear.
"They're too hard!"

Goldilocks agreed,
"Autocracies must repress civil war."

"Well, not democracies either, dear," rejoined Mama Bear.
"They're too soft."

Again, Goldilocks agreed,
"Democracies can accommodate civil conflict peacefully."

"Civil wars are most likely in anocracies!"
exclaimed Baby Bear. "They're just right!"

"Yes," Goldilocks reasoned, "if civil war is unlikely in democracies and dictatorships, it must, by default, be more likely in the middle."

"Anocracy leads to civil war," they said all together.

Late that night, while the Bears were asleep, Goldilocks sneaked back into the Bears' home with a basket. True to her thieving nature, she stole the Bears' data on political regimes and took all of their codebooks!

Wherever she went, Goldilocks spread her message of anocracies and civil war to all she met.

One day, she came upon the Big Bad Wolf!

The Big Bad Wolf asked, "Whatcha got for me, li'l girl?"

Goldilocks whimpered and finally whispered meekly,
"I've got a correlation..."

Goldilocks desperately hoped to satisfy the Big Bad Wolf with some yummy political science knowledge.

"What's the correlation, Goldie?" said the Big Bad Wolf.

"Well," replied Goldilocks, "Civil war is most likely under anocratic political regimes."

"Anocratic?! $&@#!??
What the heck does anocratic mean?" growled the Big Bad Wolf.

The Big Bad Wolf grew impatient!
He grabbed the basket from Goldilocks and began to devour the data and the pages of the codebook!

"Aha!" blurted the Big Bad Wolf after a moment. "I found it!"
"Here's what anocracies are..."

The Big Bad Wolf read aloud slowly, as if stunned by the words in the codebook.

"Anocracies are countries that are suffering from political violence or civil war..."

Goldilocks gasped!

And the Big Bad Wolf snarled at her with his big bad teeth.

"That's more like a definition than a correlation," growled the Big Bad Wolf staring at Goldilocks with his big bad eyes. "Civil wars more likely where there is political violence?!? %$&@?#$!! That's just a truism!"

Goldilocks wasn't sure.

The Big Bad Wolf listened with his big bad ears.

"What if anyone is curious about political regime and civil war?" Goldilocks stammered.

"Tell them to use a measure of political regime that does NOT mix in political violence!" The pensive beast looked terribly frightening as he pronounced his instructions.

"Well, I will spread the message to all I meet," Goldilocks announced.

"Promise never to use this mixed up measure to study civil war," commanded the Big Bad Wolf.

"I promise! I promise!" cried Goldilocks, joyfully.

"And tell all you meet," added the Big Bad Wolf, "if they use the mixed up measure, I'll find them and eat them up!"

Goldilocks was persuaded. She never used the mixed up data again to study civil war. She later discovered that the Three Bears had constructed a new improved dataset on political regimes that did not mix in political violence!

Goldilocks, the Big Bad Wolf, and the Three Bears all lived happily ever after.

And Goldilocks spread the word.

No political scientist ever used the mixed up data ever again.


The above is a fairy tale.

Not the part about Goldilocks, the Three Bears, and the Big Bad Wolf.

All of that is more or less true.

The fairy tale part is that political scientists stopped using the mixed up data.

For the true parts of the story, the names were changed to protect the innocent.

Yes, the Three Bears and Goldilocks were all innocent. They used the same measure of political regime that everyone did, and they actually warned people that political violence was somehow mixed in. But the codebook was dense and convoluted. (Only a depraved and hungry and, let's face it, somewhat loony wolf would ever devour and dissect the codebook and the data.) Beyond this, they have made important contributions to the study of civil war in particular, and to political science in general. (As for the new improved dataset that the Bears constructed that doesn't mix in political violence, click here.)

The GUILTY are the political scientists who NOW go on using the same old mixed up measure of anocracy to test hypotheses about civil war.

Would you like to help me put a stop to this?

It's really quite easy. Below is a link to the Polity Data (yes, it is the Polity measure that mixes political regime with political violence. For the full story, click on this link).

The Polity index of political regime is made up of 5 composite variables. Two of the variables, PARCOMP and PARREG, are coded with direct reference to political violence and even civil war.

If you want to control for political regime in your studies of civil war, and you want to use the Polity data, there is an easy fix. Just remove PARCOMP and PARREG from the index.

There are only two obstacles.

First, the Polity components do not come ready-to-go. You must re-code them so that they take on the values that enter into the Polity index.

Then you must merge the components with your dataset.

The good news is that the data available here take care of both of these obstacles for you!

The data provided through this page include the original component data of Polity as well as the re-coded versions of each component. They are ready-to-go. The dataset also includes several country-codes so that you can merge it with your data easily.

So there are no excuses.

Here is the list of the variables included in the Stata datafile. At the end of the list is the link to the data.
country_name: Name of the country.
aclpcode: Country codes used by Przeworski et al. (2000) and Cheibub and Gandhi (2004).
wdicode: Country codes used by the World Bank. (Not recommended- the World Bank is bad at keeping track of countries that cease to exist or change size.)
imf_code: Country codes used by the International Monetary Fund. (Same issue as the World Bank codes.)
politycode: Country codes used by the Polity project. Highly recommended!!! They are the best at tracking countries.
year: Year
democ: Polity's Democracy variable.
autoc: Polity's Autocracy variable.
polity: The Polity index.
polity2: The Polity index with missing values filled in using certain rules of thumb.
xrcomp: Polity's competitiveness of executive recruitment variable - recoded as it enters into the overall Polity index.
xropen: Polity's openness of executive recruitment variable - recoded as it enters into the overall Polity index.
xconst: Polity's Constraints on chief executive variable - recoded as it enters into the overall Polity index.
parreg: Polity's Regulation of political participation variable - recoded as it enters into the overall Polity index.
parcomp: Polity's Competitiveness of political participation variable - recoded as it enters into the overall Polity index.
xrcomp_original: Polity's competitiveness of executive recruitment variable - as it is provided in the original Polity dataset.
xropen_original: Polity's openness of executive recruitment variable - as it is provided in the original Polity dataset..
xconst_original: Polity's Constraints on chief executive variable - as it is provided in the original Polity dataset..
parreg_original: Polity's Regulation of political participation variable - as it is provided in the original Polity dataset..
parcomp_orginal: Polity's Competitiveness of political participation variable - as it is provided in the original Polity dataset..
xpolity: The Polity index minus parcomp and parreg. Introduced in Vreeland (2008).

CLICK HERE FOR THE POLITY COMPONENT DATA

If you use the data I provide here, you should cite the original Polity source. Please also register. Click here for the link.

Please also cite my Journal of Conflict Resolution piece, which explains the variables above in detail:

Vreeland, James Raymond. 2008. The Effect of Political Regime on Civil War: Unpacking Anocracy. Journal of Conflict Resolution 52 (3):401-425.

The JCR piece explains the non-fairy tale version of the story above. It also explains the XPOLITY variable in the dataset. Interested readers may also wish to see the Web Appendix to the JCR article. Click here for the Web Appendix.