Application of Para Data analysis to remote quality improvement in tablet-based surveys
64th ISI World Statistics Congress - Ottawa, Canada
Format: CPS Abstract
Session: CPS 38 - Survey statistics I
Tuesday 18 July 8:30 a.m. - 9:40 a.m. (Canada/Eastern)
The rapid development in information and communications technology (ICT) has enabled information to be generated and shared quickly nowadays. Electronic gadgets, such as cellular phones, Tablets, Global Positioning Systems (GPS), and scanning devices, and fora like social media and e-commerce create volumes of data on a daily basis, and in some instances by the second. The NSOs are at varying stages in digitally transforming their national surveys and census process, products and services through assimilating and using big data techniques with help of location-based data.
Further, many of crucial surveys like labour force survey, customer price index and those which are conducted by tablet devices produce a valuable source of para data as well. These para data are mostly geographical-based data (GPS info) with unique and primary keys and it can be used to analysis and simulation in many terms including but not limited to the enumerators’ daily performance, complementary trend of questionnaires, emerging errors, manipulations and edition during the implementation, time and date of completion, route mapping, overall performance and role of different executive roles. Analysis on this para data can be done in different phase from implementation to final tabulation. In order to improve the quality of survey during the reference period, a controlling room track each enumerators’ performance online and report on daily bases. Complimentary, para data produce spontaneously in tablet devices, is an extravagant advantage to assess and track the progress more efficient. Regarding to this analysis on tablet logs, important criteria like average completion time (urban and rural area), questionnaire errors, quality measurements were simulated. Based on the outcomes, practical suggestions were presented to be projected optimizing in future conduction of the survey. This study can be beneficial for other tablet-base surveys in other countries.