Hyperstats also has a nice summary of the general formula for finding cofidence intervals.
Here are three java applets to remind us that the 95% confidence interval means we will only capture the true mean 95% of the time. (Of course we never know which time we fall in that other 5%.) Do each of these as they each address the idea in slightly different ways Applet one Applet two Applet three
You can sign in at the Yates/Moore/Starns web site and take Practice tests. (Leave instructor blank when asked and "continue"). For this chapter you want their chapter 10 quiz.
This link describes an experiment on the effects of calcium for treating blood pressure in African American males. The data is at the bottom. Use the data for "Begin" for both groups to construct a 95% confidence interval on the true mean of an untreated population. Use the "END" data from and on the true mean of a population if calcium treatment was used. Be sure to verify all necessary conditions. This data is also available at Cacium data on StatCrunch
Here is a video on a step by step approach to writing an acceptable Confidence interval statement.
and this one has a little history and information on uses of confidence intervals.
Unsure about finding critical values, here is a brief cover