This seems a rather stupid question to ask, and especially in Zimbabwe where we talk about this endlessly. However, this is not a trivial question, and we remember 2008 and 2002 more clearly than we do 2005. Simply put, is the killing, beating, and raping of citizens worse from the point of elections than the threatening, terrifying, and starving of the them? It all depends on the purpose and the consequence.
If the consequence is to change the result of the vote and hence who governs, then surely both are equivalent as regards the final result: that those who use either strategy subvert the real purpose of elections? Which is what? Surely that the citizens can ensure, freely, that those that govern have the mandate to govern?
So, we need to be very clear, when we talk about elections, and we talk about election violence, that we are clear about what this is. So, when killing, beating, and raping do not happen, but threatening, terrifying, and starving does, we are certain that election violence still happened. We need no repeats of 2008 and 2002, or even 2005!
So what do we mean when we talk about election violence? Consider this definition:
…Acts or threats of coercion, intimidation, or physical harm perpetrated to affect an
electoral process or that arises in the context of electoral competition. When
perpetrated to affect an electoral process, violence may be employed to influence the process of elections – such as efforts to delay, disrupt, or derail a poll – and to influence the outcomes: the determining of winners in competitive races for political office or to secure approval or disapproval of referendum questions.
As Timothy Sisk points out above this is considerably broader than the presence of physical violence: it is the range of activities aimed at subverting the will of ordinary citizens to freely exercise their choice.
Electoral violence is a sub-type of political violence in which actors employ coercion in an
instrumental way to advance their interests or achieve specific political ends. Similarly,
societies prone to experiencing election-related violence are normally vulnerable to
broader kinds of political violence; Kosovo, India, Indonesia, Sri Lanka, Kenya, or
Colombia are examples of instances in which electoral violence is embedded in a
broader, often ongoing context of deep-rooted social conflict.
Electoral violence includes acts, such as assassination of opponents or spontaneous
fisticuffs between rival groups of supporters and threats, coercion, and intimidation of
opponents, voters, or election officials. Threat and intimidation is a form of coercion
that is just as powerful as acts of violence can be. Indeed, one purpose of acts of
terrorism such as tossing a grenade into a crowd of rival supporters is an act
diabolically designed to induce fear and to intimidate (e.g., to suppress mobilization or
voting by that group).
Violent acts can be targeted against people or things, such as the targeting of
communities or candidates or the deliberate destruction of campaign materials, vehicles,
offices, or ballot boxes.
Electoral violence is more than just physical violence: it is the purpose behind violence, and the oscillation between physical violence and psychological violence that enable us to understand this purpose in Zimbabwe. The results of the elections in 2005 can only be understood in the context of the violence of 2002 and 2008. That 2005 was less violent than the two previous elections is not really the point, and it would be useful here if the South African Government would stop contesting the release of the Khampepe/Moseneke report: we could then see the nexus between 2000/2002 and 2005.
And, just maybe, SADC would own up to the Principles that it promulgated so piously in 2005, and start to insist that the GPA required constitutional change and reform, then elections, rather than accepting the weak compromise offered by the GNU of constitutional change, then elections and reform. Then maybe the SADC Treaty would be a real, substantive document as opposed to a loose-leaf folder from which pages are removed whenever they are inconvenient! And they are especially inconvenient when elections (and sometimes courts and court decisions) leave the members in potential conflict with each other over who has the right to rule.
 Sisk, T. D, Elections in Fragile States: Between Voice and Violence. Paper Prepared for The International Studies Association Annual Meeting. San Francisco, California. March 24-28, 2008.