A campus survey uses randomly selected dorm-room numbers, but researchers skip residence halls where security rules make entry difficult. Which feature of the sampling procedure is most likely to introduce sampling bias?