Nothing can compare to interview and questionaire results IF the questions are posed appropriately and ask for the most pertinent information. All along though, they must have a good mix of open-ended questions to get the most out of your research. Remember, if you never ask, you may never get a direct response.
However, observation is an entity all of it's own...especially if you have a good group to observe that are not altering their behavor based on your presence. Try looking into the STROBE method of observing an environment. This is a very useful method of observation and is found in Systems Analysis and Design by Dr. Julie and Kenneth Kendall, plus others.