27
27
distant from our ratings. But simultaneously, they accord relatively well with our
ratings from the point of view of distribution. This holds for Polity too, even if it is
slightly behind Freedom House in this respect.
Trying to make sense of the substantial difference, in terms of average deviation,
between our score and Freedom House’s from 1990, the following could be told. The
difference between our score for elections and Freedom House’s for political rights
(which are close in conceptual focus) is as high as -4.07. Comparing the average score
for the each electoral subcategory in our index with the aggregate political rights
score (which is the only that is reported), we get the highest differences for the
openness and the effectiveness of elections: -4.42 and -4,72 respectively. Looking at
political freedoms vs. civil liberties (in the case of Freedom House) the average
deviation for the same period is lower: -2.00. Among the individual components, the
deviation is highest for organizational freedoms: -3.53. On the other hand we notice a
very small difference with respect to political violence and repression: -.26. Had we
only looked at this aspect, in other words, our ratings would almost have concurred
with the aggregate political liberty ratings by Freedom House. It should be observed
that for the sample under investigation, this is the component which gets the lowest
average score in our rating (after all, a third of the counties were marred by civil war
during the year of investigation).
One other thing about the Freedom House ratings is that the to two components,
political rights and civil liberties, often are awarded similar points. In fact, among the
cases under study, the score is exactly the same 17 times out of 37. By contrast, in our
data this happens only once for the corresponding components. This indicates that the
Freedom House ratings are driven by a more holistic view. Points are awarded, in
other words, in a more comprehensive manner; one general aspect seems to color the
whole picture. As it seems, this general aspect has been concerns about political
violence and repression.
As for Polity, for the same period (1990 onward), there is almost no difference
between its aggregate score and our aggregate score for elections: -0.09. Yet for
political freedoms (which are not explicitly accounted for by Polity) a substantial
difference emerges: 2.46. Among individual indicators, the difference is highest for
Convention