On stability questions for pendulum-type equations