1. Starch, Daniel, Revised Study of Radio Broadcasting, New York, National Broadcasting Co., 1930. A similar trend in the return of mailed questionnaires occurs in Kirkpatrick, Clifford, Attitudes and Habits of Radio Listeners, St. Paul, Webb Book Co., 1933. In the latter case, although approximately 8% of the returned questionnaires come from Group I (as in the present study), an attempt is made by a process of averaging to show that this disproportionate return does not bias the sample. Occupational groups, however, cannot be added and divided by the number of returns any more than room numbers can be added and divided by the number of rooms to find the number of the average room.
2. Stanton, F. N., Checking the Checkers, Adv. & Sell., 1935, 26, 24. This study indicates that the error in estimating the average amount of time the radio is in use daily is about 10% of the time that it was actually turned on. Perfect agreement between estimates and the actual time the radio is in use occur only when the question breaks the listening day into periods.