Companies in the competitive hotel industry need to keep up with the tastes of their visitors. If the rooms seem dated and the service is slow, travelers will choose a different destination for their next visit. One means of monitoring customer satisfaction is to put cards in guest rooms, asking for feedback. These are seldom returned, and when they are returned, it is usually because of some unusual event.
To get a more representative sample, a large hotel frequently used by business travelers decided to conduct a survey. It contacted every guest who stayed in the hotel on a randomly chosen weekday during the previous two months (June and July). On the date chosen (Tuesday, July 19), the hotel had 437 guests. With several follow-up calls (unless the customer asked to be excluded), the hotel achieved an 85% response rate. The hotel could have mailed questionnaires to guests much more cheaply, at about one-tenth of the cost of the telephone calling.
The key question on the survey was, “Do you plan to stay with us on your next visit to our area?” In the survey, 78% of guests responded Yes to this question.
(a) What can the company hope to learn from this survey that it could not get from the cards left in guest rooms?
(b) The response rate from mailed questionnaires is typically less than 25%. Would it make sense to mail out survey forms to more customers rather than go to the expense of calling?
(c) Carefully describe what each observation in this survey represents. What is the population?
(d) Does the procedure yield a random sample of guests at the hotel, or does this design make the survey more likely to include some guests than others?
(e) In calling the customers, one of the interviewers was a man and one was a woman. Might this difference produce a difference in answers?
(f) Some customers were called repeatedly in order to improve the response rate. If the repeated calling annoyed any of these customers, how do you think this might bias the results?
(g) In describing the results of the survey, what points about the randomization should be made in order to justify the reliability of the results?
(h) How should the hotel deal with the 15% of customers from that day who either did not reply or asked to be excused from the survey and not called again?