In the process of testing their codes, computational scientists frequently encounter challenges that are not typically encountered by software engineers who develop testing techniques. In response to this oversight, we have developed a research technique, called Mutation Sensitivity Testing. Using this technique we found that a few well-designed tests were able to detect a high percentage of the code faults that were introduced into small MATLAB functions. Our experiments also showed that it is often more effective to reduce error tolerances than to conduct more tests. These results suggest that, in the field of testing, software engineers and computational scientists have much to offer each other.