This video clip offers a visual representation of an extensive statistical analysis of the changes in levels of democracy in all the countries of the world, from 1972 to 1996. The contrasting colors, presented on a world map, graphically highlight the countries that experience some change in levels of democracy from one year to the next, and provide a visual test of the extent to which the statistical model actually fits historical patterns. It is intended as an aid, to provide a more intuitive grasp of the results of the statistical analysis and of historical patterns of democratic transition in the third wave of democratization. Our analysis generates two important kinds of information: a first stage probit produces an estimate of the probability that a given country will experience regime change in a given year, in light of the variables in the model; a second stage regression analysis predicts the magnitude and direction of change on the basis of a different mix of variables. In addition, we have information on the actual changes, derived from a database prepared by Freedom House, in which a country is given a democratic score for each of the years included in the study.

The video clip models first the probability that a particular country will experience a change in its level of democracy (for better or worse) as derived from our probit analysis, by flashing in red those countries that are predicted to change. The greater the intensity of the red, the greater the probability that the country will experience some regime change that year. Next, the red fades into blue to reflect the magnitude and direction of change predicted by the statistical model, or to a green background color for countries that do not change at all. The blue is also shown in varying levels of intensity: near white is used for countries in which the model predicts a fairly large negative change (a freezing of political and civil liberties, if you will), a more neutral light blue is used for the countries that will experience little change, and a deep blue for those countries that will move strongly in the direction of greater democracy. A well known political science article uses the color blue as a metaphor for greater quality of democracy, and so it appropriately represents change in that direction. Finally, the actual magnitude and direction of change is shown in various shades of purple. This process is repeated for every year.

By tracking the change in intensity of the colors of particular countries, the viewer can quickly garner a great deal of information. The flashing red on the map draws the eye to the countries in which conditions are ripe for regime change (at least according to our variables). The fading of red into blue or white reflects the extent to which a predicted probability becomes actual change; blue or white coming out of nowhere illustrates the extent to which our model failed to identify instability; and red into green shows a prediction of change that did not come to pass. This provides an impressionistic test of the statistical significance of the first stage probit model. Similarly, the fading of blue into purple reflects the accuracy or significance of the second stage model: if our predictions are accurate, the blues will fade into similar shades of purple. Moreover, watching particular regions more closely, watching the same country over time, or looking at the map as a whole provides a visual representation of regional patterns of change, the extent of instability within a country, or the years in which regime change sweeps over large areas of the world at the same time. This information can, of course, be conveyed using tables, graphs, coefficients and tests of statistical significance, as we do in our paper. But the video clip presents the same information, at a higher level of generality, in just over 5 minutes.