Web Survey Bibliography

Title Panel Online: how important is the recruitment mode ?
Year 2016
Access date 29.04.2016
Presentation PDF (1.02MB)
Abstract
Relevance & Research Question: Non probabilistic online panels are gaining more and more market shares. The 2 main factors are determining this success are the low cost and speed in the data collection. We focalized our research on the the way panellists were recruited: offline through CATI interview or online through banners and advertisements on facebook. Our main question is : Is there any difference between panelists recruited via CATI and panelists recruited online in terms of data quality ?

Methods & Data: For our research we used data collected using our own panel: Opinione.net. This panel was created on January 2011, during the first 4 years the panelists were recruited exclusively offline. In the last years we also started to recruit panelists online. At the time when we’ve carried out the surveys 6674 panelists were subscribed to our panel, 3632 recruited offline, 3042 recruited online. To assess our hypothesis we analysed the data collected by 5 different web surveys carried out on our panel.

To measure the data quality we’ve used several indicators (response time, number of DK, non differentiation index)

Results: Panelists recruited online are generally quicker than the ones recruited offline, in all 5 surveys the length of interview for the web recruited is shorter (significative difference) and 4 times out of 5 the panelists recruited via cati take 20% more time to complete the questionnaire.

Also the non differentiation index shows differences between two groups: panelists recruited via CATI give more differentiated responses (T test significative).

Panelists recruited online are generally younger, high percentage of women and housewives, while panelists recruited via CATI show a greater percentage of pensioners. Panelists recruited via web are more active (significantly higher number of clicked links/day ), have an higher percentage of 'Completed' and a lower percentage of 'Early Screen Out'.
Year of publication2016
Bibliographic typeConferences, workshops, tutorials, presentations
Print

Web survey bibliography (8390)

Page:
Page: