How SAMPLE|QUALITYGUARD detects anomalies before they distort results
You can rely on your data
Monday, 9 a.m.
Over the weekend, 300 interviews for an online survey were received. Some were completed in three minutes instead of the expected twelve. One open-ended response consisted of just five letters: "sdfgh". Other responses were surprisingly detailed and seemed almost too perfect.
Which interviews are reliable? Which answers provide valuable insights? And which, due to poor quality, could skew your results?
Anyone who asks these questions only after the fieldwork is complete risks unnecessary follow-up surveys and a data set that is difficult to defend. That's why SAMPLE|QUALITYGUARD reviews incoming interviews during the fieldwork phase.
Over the weekend, 300 interviews for an online survey were received. Some were completed in three minutes instead of the expected twelve. One open-ended response consisted of just five letters: "sdfgh". Other responses were surprisingly detailed and seemed almost too perfect.
Which interviews are reliable? Which answers provide valuable insights? And which, due to poor quality, could skew your results?
Anyone who asks these questions only after the fieldwork is complete risks unnecessary follow-up surveys and a data set that is difficult to defend. That's why SAMPLE|QUALITYGUARD reviews incoming interviews during the fieldwork phase.
Conspicuous does not automatically mean unusable
Automated checks detect typical anomalies in large datasets, for example:
- unusually short processing times
- uniform answer patterns such as straightlining
- striking click patterns in matrix questions
- unusual typing behavior
- incorrect answers to control questions
- logical inconsistencies between multiple responses
- inappropriate or strikingly similar open-ended answers
However, an anomaly is initially just an indicator. It could point to a lack of attention or automated processing. But it could also reflect a plausible characteristic of the respondent or the target group. Those who rely solely on fixed rules for filtering may also lose valid interviews that could yield valuable insights and accept the bias that can arise from removing interviews.
Therefore, SAMPLE|QUALITYGUARD combines automated checks with expert evaluation: The technology makes anomalies visible, and the research team categorizes and analyzes them.
This is how quality assurance works
- Define criteria
At the start of the project, standardized tests and project-specific rules are defined, derived from the target group, questionnaire, and research objective. - Continuously monitor
During the fieldwork phase, SAMPLE|QUALITYGUARD flags conspicuous interviews as soon as they are received. - Expertly evaluate
Experienced researchers decide which interviews remain in the dataset, which are excluded, and which are plausible despite conspicuousness. - Replace if necessary
Problematic interviews can be replaced even during the fieldwork phase. - Document
Test criteria and decisions are recorded transparently.
What you get out of it
Data quality isn't just a methodological issue. The more important the decision, the more often the internal question arises: "Can we really trust the results?" The SAMPLE|QUALITYGUARD helps you provide a reliable answer:
You can rely on your data.
How reliable is the data basis of your next online study?
We'll discuss with you:
Contact us.
- Early detection: Quality problems become apparent during the ongoing data collection process, rather than only at the end of the project. This creates room for improvement.
- Traceable: Not only results are documented, but also criteria and decisions. This makes it possible to explain why an interview was retained or excluded.
- Valuable interviews are preserved: No interview is discarded solely because of an automated notification. This protects insights that would otherwise be lost.
You can rely on your data.
How reliable is the data basis of your next online study?
We'll discuss with you:
- which quality risks are typical for your target group and your questionnaire,
- which standardized and project-specific testing criteria are appropriate,
- how you can transparently document quality decisions.
Contact us.
Resquest