Field work, survey completion times and data quality in Citizen Panel 4 - 2012. LORE working paper 2013:1
Field work, survey completion times and data quality in Citizen Panel 4 - 2012
Författare och institution:
Love Christensen (Statsvetenskapliga institutionen); Johan Martinsson (Statsvetenskapliga institutionen)
This report examines the inflow over the period of field work of a web survey, the amount of time it takes a respondent to complete the survey (duration) and a few data quality indicators.
The first part shows us a very large part of the total completed questionnaires are received during the very first day of field work, close to 50 percent. However, after a field work period extended to three weeks, twice as many completed surveys are received in total. It is also made clear that the time of day when most people tend to answer their questionnaires is between 8 am and 11 am. These hours account for almost 40 percent of all completed questionnaires. We also discover a high amount of variation when it comes to the time respondents spend filling out the web questionnaire. On average, this particular survey took 25 minutes to complete. When it comes to predictors of survey duration it is found that people who answer late at night or who are highly interested in politics spend more time on the survey than others.
When it comes to data quality the respondents of the Citizen Panel 4 seem be straight-lining to a fairly high extent. As many as one out of four respondents had straight-lined at least one out of seven question batteries. We also confirm a curve-linear relationship between time spent on survey and data quality. Those filling out the questionnaire just slightly slower than the average respondent are those who are least likely to be straight-lining.
Ämne (baseras på Högskoleverkets indelning av forskningsämnen):
Sannolikhetsteori och statistik ->
survey methodology, web panel, data quality
LORE working papers
Visa i Endnote-format