Intellicast S3E39 – Sales in the New Normal with Jason Silver of Gazelle Global
August 17, 2020
5 Great Market Research Podcasts from 2020 You May Have Missed
August 18, 2020
To conclude our data quality series, we have compiled some recommendations and best practices to help you feel confident in your data.
Best Practices in Evaluating Data Quality
The most important factors in determining data quality are to design the screener and questionnaire accordingly.
We recommend spending extra time on the screener to ensure it is not leading and that you are speaking with the right people. It is important to make sure that you are addressing the appropriate people in the right way. You want to be clear with what you are asking the respondent.
Furthermore, always spend time on the questionnaire and ensure that it is easy to take on all devices and that respondents can answer each question without feeling frustrated.
When measuring quality, we recommend:
- Always including at least one open-ended question because it is still the best method to determine a quality respondent.
- Being careful not to include too many data quality measures, particularly ones that don’t pertain to the topic (math questions, red herrings, attention checks.)
- Using a multiple check system rather than removing someone for one quality fail. For example, it might be best to implement a 3-strike system.
The most important people to remove are the biggest offenders—they are also the biggest outliers in the data. Depending on the risk of the study, it may not be necessary to nitpick and remove those that are in the grey area.
Be as specific as you can with your sample provider in terms of why you’re removing respondents so that the sample provider can act accordingly. They may take different actions depending on the severity, and it helps clean up the sample for future use.
Lastly, remember that respondents are people too!
The end goal is to always collect the most reliable data so that better business decisions can be made from it.
To read the previous installments in our Data Quality series, follow these links:
An Introduction to Data Quality Checks and Their Use Determining Data Quality
When It Comes to Data Quality—Not All Panels are Equal
Data Quality—Demographic Differences
Changes in Attitudes and Behaviors and How they Impact Data Quality
