Login
Last updated on August 11, 2024

Docy

Recommended criteria for questionnaires to comprehensively analyse data quality with ReDem®

Estimated reading: 3 minutes

This article explains the criteria and requirements that your questionnaire should fulfill in order for ReDem® to make a meaningful assessment of the data quality. ReDem® allows a flexible application of the quality checks. Nevertheless, it is essential to observe a minimum number of quality checks for an objective quality control.

Before your survey goes into the field, it is important to ensure that your questionnaire design allows for a deep analysis of data quality. If a questionnaire does not meet the requirements below, we recommend making adjustments or enhancements. Open-ended questions in particular have proven to be extremely effective in ensuring data quality. The following criteria represent our recommended minimum requirements for ReDem® quality checks:

1. Open ended questions

  • Minimum 2 open ended questions - these should be answered by all respondents. Therefore, make it mandatory to answer the open questions. 
  • Questions that require answers in a specific, well-defined context are best, such as "What was the deciding reason for you to buy product XYZ?“ or "Please describe your ideal summer vacation.“
  • The more difficult it is to determine whether an answer to an open question makes sense, the less suitable it is for a quality check. An unsuitable question would be, for example, "What else do you want to tell us?". 
  • Due to their limited answer variety, semi-open questions (e.g. brand names) are unsuitable for quality checks.
    Such questions are used when the (brand) names cannot be covered by closed questions. This is the case, for example, with questions such as "In which online stores have you shopped within the last month?“or when unaided recall is required, such as in the question "Which brands do you know in the category XYZ?“.
  • The open-ended questions can be added for quality control purposes only and do not need to be evaluated.

2. Item batteries / Matrix questions

  • At least 1 item battery / matrix question with at least 7 items (the more, the better) and in each case at least 4 attribute values (e.g. Likert scale).
  • Item batteries are best suited when the repreated selection of the same scale values is contradirectory. For example, if someone strongly agrees that stricter penalties are needed for drink-driving; and at the same time strongly agrees that the blood alcohol limit for drivers should be raised.

3. Time durations

  • At least the total interview duration, which can be used to determine whether someone has an unusually fast or slow response behavior. 
  • We recommend additionally selecting additional time sections to reach an even more meaningful assessment (e.g. a matrix question with a lot of text or open-ended questions). 
  • Less suitable are durations for questions that can be answered very quickly, such as short yes/no questions or socio-demographic questions.
  • If interruptions are possible or permitted during the completion of the questionnaire, they must not affect the time duration.

nline surveys that have passed all quality checks according to the three recommended criteria and have subsequently been cleaned up by ReDem® are certified with the ReDem® seal of approval. If you have any further questions about the optimal design of questionnaires, please feel free to contact us at support@redem.io .

Share article
CONTENTS