Changes in cohabitation question on its answers: evidence from a panel study of Taiwan
Jul 25, 09:55
In Chinese Society, whether someone is cohabiting with another person is a sensitive question for a majority of people, especially the elder ones. According to an estimate based on the 2010 Census data of Taiwan, the rate of cohabitation of women aged 15 or above is 2.2%. Whether the estimate underestimated or overestimated the true rate of cohabitation are still controversial in Taiwan. In the field of survey methodology, there exist only few studies on how the design of cohabitation question affects the estimated rate of cohabitation. Whether changes in cohabitation question affect its answers is worthy of exploration in both the fields of survey research and demography. In the Taiwan Panel Study of Family Dynamics (PSFD) established in 1999, the design of cohabitation question has changed since the 2016 survey. Prior to the changes, the respondents were asked to choose one answer from the following categories: married, widowed, divorced, separated, never married, or living with a partner. Since the 2016 survey, the question has been divided into two sub-items. The first sub-item asked whether the respondent is now married, widowed, divorced, separated, or never married. For those who answered “widowed,” “divorced,” or “never married,” the interviewers would ask a follow-up question regarding whether the respondent is now living with a partner. A preliminary look at the data suggests that the proportion of respondents answering cohabitation has become higher after the changes in question. For example, within the 1,527 respondents who were born in 1964-76, only 15 of them chose the answer of cohabitation in the 2014 survey. Among the 1,494 respondents who successfully finished the follow-up survey in 2016, 543 persons reported that they were widowed, divorced, or never married at the time of survey, in which 36 persons replied that they were living with a partner. Using the panel data collected from the PSFD, this study shall analyze the effects of changes in cohabitation question on the responding pattern of the respondents. In addition, whether the genders of the interviewer and respondent matter for the answers to cohabitation question shall be explored. The possible intertwining effects of the interviewer’s age and respondent’s age shall be examined as well.