(Note: the following links go to pages with multiple graphs, and require some downloading patience.)
For example, a set of explorations of the Yugoslavia data was conducted to give us more insight into the structure of the data and the implications of our choices for analytical strategies, especially the block size. A second set of graphical and statistical alternatives looks at potentially revealing ways of displaying the data. A third collection of graphics, drawing on these explorations, works toward a standardized set of displays.

But first, a brief discussion of the purpose of the experiment, the nature of the data, and the primary analytical strategies:

The experimental hypothesis is that otherwise random data recorded during extraordinary global events will be slightly less random, resulting in an unusual deviation above or below the expected mean value. The fundamental elements in the measurement technology used by the GCP are the trials that make up the ongoing sequence of random events. When the random sequence is undisturbed, the trial deviations are symmetrically distributed around the mean, and an accumulation of the deviations over time will look like a random walk, with unpredictable excursions wandering above and below the mean, but no obvious trend. In contrast, if the sequence of data points is even slightly correlated with some other variable, the random walk will become less random, and take on a trend with a slope that corresponds to the strength of the correlation. The terminal value of such a cumulative deviation represents the "bottom line" for the question whether the correlation with the defined global event is significant according to the appropriate statistical criteria.