I would like to second Peter Bruce's comments about resampling. I've taught stats using resampling and have found the approach to be helpful for my students. They can get to the meat of the material and use the technique to do quite sophisticated problems and projects. All their work with simulation gives them a deep appreciation for the techniques and theories behind stats.
In two or three days many students are up and running doing great projects like: "Is there a difference between the number of hours of sleep juniors get vs that of seniors?" "How accurate is the college office in predicting a 'reach' for a senior?" "What is the approximate percentage of students in the school who lied on a recent poll?"