(
elections – parallel to the common finding that such domestic variables have little effect on use
of force rates.
The argument proceeds through five sections. First, a “media priming” model of rallies is
introduced, and reasons are given why this model predicts that peace events will be causes of
rallies. Next, the methods that were used to construct a dataset of peace events and code them on
relevant variables are described, as are the statistical techniques that are used to measure the
impact of these events on presidential approval. Results from regressions measuring the impact
of peace events on presidential standing are then presented; this is followed by a section that
analyzes how the frequency of peace events is affected by political variables. Finally, the
implications of these findings are discussed.
A New Model of “Rally Events”
Most authors agree that U.S. uses of force are usually followed by increases in
presidential approval (Mueller 1973;Kernell 1978;Marra; Ostrom, and Simon 1990;Lee
1977;Baum 2002;DeRouen 1995), although some point out that these increases are smaller and
less consistent than traditionally thought (Lian and Oneal 1993). The “conflict-induced
cohesion” effect is commonly invoked to explain this effect. Drawing on research in group
psychology, this theory predicts that when facing foreign threats, citizens rally around their
leader (Coser 1956;Simmel 1955); see also (Levy 1989;Stein 1976). This unifying ability is
seen as unique to external conflict, and so most studies of “rally effects” have either looked only
at the response to uses of force ((Baum 2002;Oneal; Lian, and Joyner 1996;Lian and Oneal
1993;DeRouen 1995;Edwards and Swenson 1997)) or to crises/militarized disputes ((Baker and
Oneal 2001;Oneal and Bryan 1995;James and Rioux 1998;James and Hristoulas 1994)). Others
Convention