Careful data reduction is a necessary ingredient in getting the correct results. The methods I've recommended here result from good advice, repeated tinkering, and careful thinking about statistics. However, no warranty is express or implied, and if you have suggestions for improving these methods please let me know.
For multislit spectroscopy, try the LRIS(B) Multislit Reduction Cookbook and for longslit spectroscopy, there's the LRIS(B) Longslit Reduction Cookbook.