Threats to Validity. As mentioned earlier, our survey was conducted with only one group of students taking a particular course of a particular university. However, we also mentioned that our sur- vey participants were final-year undergraduate students majoring in computer science who can be considered entry-level developers. Our survey fulfilled its purpose of gauging initial response to our change contract language; our students easily learned and used our change contract language. However, given the number of participants, a larger-scale study is necessary to confirm our results. In particular, more sophisticated study is required to see the validity of several interesting initial observations such as higher correctness rates in structural changes than in behavioral changes and little difference between the correctness rates for artificial programs and real-life programs.
Appears in 2 contracts
Sources: Software Change Contracts, Software Change Contracts