A study (N. Schwarz et al., Public Opinion Quarterly, vol. 49, 1985, p. 388) asked German adults how many hours a day they spend watching TV on a typical day. When the possible responses were the six categories (up to 0.5 hour, 0.5 to 1 hour, 1 to 1.5 hours, . . . , more than 2.5 hours), 16% of respondents said they watched more than 2.5 hours per day. When the six categories were (up to 2.5 hours, 2.5 to 3 hours, . . . , more than 4 hours), 38% said they watched more than 2.5 hours per day.
a. The samples could not have been random, or this would not have happened.
b. This shows the importance of question design, especially when people may be uncertain what the answer to the question really is.
c. This study was an experiment, not an observational study.