Imputing Missing Data: A Good Starting Point?

Hello

I am working on my first attempt to impute missing data of a data 
set with systematically incomplete answers (school performance 
tests). I was googling around for some information and found 
Amelia (Honaker et al., 2010) and the mi package (Yu-Sung et al., 
n.d.). However, since I am new to this field, I was wondering 
whether some experts could give a good recommendation of a 
starting point for me, that is a point that combines theory as 
well as practical examples. Of course, My primary interest is to 
complete the task in time (1 week), however, I want to acquire 
skills for a program that provides some future, and of course I 
want some background on what I am doing (and what not). Could you 
help with some hints, experiences, and recommendations?

Thank you.

Regards
*S*

Sascha Vieweg, saschaview at gmail.com