(1) Most types of research questions can be answered with a survey if the behavior of interest can be observed through self-reports from the participants. However, most research using surveys addresses descriptive or predictive research questions. (2) open-ended, closed-ended. (3) Test-retest reliability is good when scores on a survey show a strong positive correlation for participants who take the survey multiple times over a period of time. Split-half reliability is good when sets of scores on a survey show a strong positive correlation. (4) The question requires a forced choice between two alternatives, neither of which may be liked by the respondent. There are several ways to improve the question, but one way might be to use the following question and response choices:

 

If you were to have a pet, which animal would you prefer?

____ dog

____ cat

____ bird

____ reptile

____ other (please specify _____)

____ I do not like pets

 

(5) Several aspects of the survey allowed the researchers to accurately predict the election outcome. For example, a random sample of likely voters instead of all Americans meant that they were sampling from people who were likely to cast a vote in the upcoming election. The large sample size kept the sampling error in the study low. The inclusion of cell phone users made the sample more representative, because some individuals in the population may not have landline phones. (6) d, (7) c, (8) a, (9) e, (10) b.