Examples of KEDS-Generated Times Series

This section shows some examples of the machine-coded data produced by KEDS from Reuters lead sentences, and compares these to a human-coded data set ("WEIS"; Tomlinson 1993) coded from the New York Times. The time series are the total activity in the dyad, aggregrated using the Goldstein (1992) scale.

Further information on these charts can be found in:
Schrodt, Philip A. and Deborah J. Gerner. 1994. "Validity Assessment of a Machine-Coded Event Data Set for the Middle East, 1982-1992." American Journal of Political Science 38:825-854.

Remarks:This chart shows Goldstein-scaled activity by Israel directed to Lebanon. The Goldstein scale converts the categorical WEIS codes to an interval-level conflict (negative) to cooperation (positive) scale.

Remarks: This chart shows Goldstein-scaled activity by Israel directed to the PLO and Palestinians. This includes any activity targeted at a group identified by the Reuters lead as "Palestinian", not just activities in the West Bank and Gaza.

Remarks: This chart shows a comparison of the KEDS-coded and human-coded data when used in a secondary analysis: the cross-correlation of activity in the ISR>PAL and PAL>ISR dyads. Note that the human-coded NYT data has a somewhat higher contemporaneous correlation, but the machine-coded Reuters data picks up an additional seasonal component in the +10 to +14 month range. This difference is probably due to the source rather than the difference in coding -- Reuters is big on anniversaries and may also be more likely to cover smaller Palestinian demonstrations that occur to commemorate annual events.

Remarks: This graph started as a classroom demonstration of the semilog transformation but shows some interesting features, particularly in the moving-average (MAV). The pattern in the pre-Madrid period generally tracks the vissitudes of the Israel-Syrian relationship in dealing with Lebanon and the Palestinians. Following the initiation of the Madrid talks, the behavior becomes substantially move chaotic.