Why do telephone interviews last longer on cell phones than landline phones? Common explanations for this phenomenon include differential selection into subsets of questions, activities outside the question-answer sequence (such as collecting contact information for cell-minute reimbursement), respondent characteristics, behaviors indicating disruption to respondents’ perception and comprehension, and behaviors indicating interviewer reactions to disruption. We find that the time difference persists even when we focus only on the question-answer portion of the interview and only on shared questions (i.e., eliminating the first two explanations above). To learn why the difference persists, we use behavior codes from the U.S./Japan Newspaper Opinion Poll, a dual-frame telephone survey of US adults, to examine indicators of satisficing, line-quality issues, and distraction. Overall, we find that respondents on cell phones are more disrupted, and that the difference in interview duration occurs because cell phone respondents take longer to provide acceptable answers. Interviewers also slow their speed of speech when asking questions. A slower speaking rate from both actors results in a longer and more expensive interview when respondents use cell phones.