Saturday, November 12
Questionnaire Design
Sat, Nov 12, 1:45 PM - 3:10 PM
Hibiscus A
Questionnaire Design for Establishment Surveys

Intelligent Validation in Online Questionnaires, Including Establishment-Specific Prefill of Known Information for Cross Validation (303613)

Lise Stahl Jacobsen, Statistics Denmark 
*Pia Thomsen, Statistics Denmark 

Keywords: intelligent validations, business, establishment, surveys, questionnaires, data quality, response burden, data collection, data providers, online, digital

Data collection by dynamic online questionnaires offers a range of options with regard to dynamic verification of data while the respondent is filling out the questionnaire. At Statistics Denmark we use these possibilities in online questionnaires for business surveys in order to minimize burden and improve data quality and cost effective production of business statistics.

We use different kinds of dynamic data verification: The “soft” or indicative verification in which the respondent meets a dynamic warning, and may review or ignore; the “semi-hard” verification in which the respondent must correct, explain or confirm responses; and the “hard” verification in which the respondent must correct data before it can be submitted.

Verification rules are run DURING the response process and verification results are presented to respondent immediately - when the respondent is focused on the task and expects it. This reduces response burden as the extent of re-contact AFTER submission can be reduced or even avoided.

The various kinds of verification has been utilized in validation of individual responses, in cross-validation of relations between individual responses, and in cross validation between responses and pre-known information, which may or may not be displayed to the respondent.

We will present examples of intelligent validation - including use of prefill of known information for cross validation implemented in business questionnaires at Statistics Denmark. Furthermore, we will present some results in terms of e.g., the amount of conspicuous values found in the submitted data or the extent of re-contacts - before and after implementation of validation in the online questionnaires.