The Hartley (1946) and Politz–Simmons (1949) method. Suppose that all calls are made during Monday through Friday evenings. Each respondent is asked whether he or she was at home at the time of the interview, on each of the four preceding weeknights. Respondent i replies that she was home ki of the 4 nights. It is then assumed that the probability of response is proportional to the number of nights at home during interviewing hours, so the probability of response is estimated by ˆφ i =(ki + 1)/5. Let

a. Under what circumstances would you expect the method to reduce bias due to non response?
What assumptions must be made for the estimator to be approximately unbiased?
b. What are some potential drawbacks of the method for use in practice? How does it adjust for persons who were not at home during any of the five nights, or who refused to participate in the survey?