How do interviewers influence data?
|Convenor||Miss Claudia Karwath (Leibniz Institute for Educational Trajectories )|
|Coordinator 1||Miss Manja Attig (Leibniz Institute for Educational Trajectories)|
In Japan, the study on the mode-difference in the CAIs is not so many. We conducted the experimental CAI survey to investigate the mode effects, especially the influence of the presence of interviewers in 2014. The result shows CAPI respondents tend to make more socially desirable responses than CASI respondents. In the comparison of subjective status identification using IPWE to control non response errors, the mean value of the subjective social status on CAPI is significantly higher than that on CASI.
These results suggest that in the situation using CAI, social desirability bias(the effect of interviewer) is problematic.
Collecting biomarkers and obtaining permission to link to administrative data within surveys holds significant promise, but introduces potential for nonresponse bias. We expand upon previous work by combining multiple data sources from the Health and Retirement Study, allowing for multifaceted exploration of consent dynamics in a face-to-face context. We first link requests for physical measurements, biomarker collection, and data linkage with respondent characteristics, moving beyond demographics to explore attributes such as personality that may correlate with consent. We then extend our focus on interviewers, incorporating results of behavior coding of the interviewer-respondent interaction and assessing interviewer characteristics
In this paper we ask whether interviewer characteristics influence the answers to a standard set of financial literacy questions. We fear that interviewer effects can arise because of interviewers perhaps helping respondents to answer the literacy questions or some other obscure interaction between them. We use data from Germany’s wealth survey “The Panel on Household Finance (PHF)” to investigate this issue. Controlling for the usual respondent characteristics, we find that interviewer fixed effects explain a substantial fraction of the variance of the financial literacy score and interviewers’ age, gender and education have statistically significant effects on it.
In the early childhood cohort of the National Educational Panel Study interviewers are involved in contacting the participants and collecting the data. Because of the age of the children (6 - 8 months) and the household situation, the interviewers are exposed to a special challenge, as they have to conduct the direct and indirect measurement assessments in a standardized way and in accordance with the administration rules as well as they need to handle up with the “household situation” (e.g., babies crying). We focus on interviewer effects, examine if interviewer characteristics influence the survey process and the collected data.