Quality Score

Catch what standard industry checks consistently miss

Dynata’s QualityScore™ technology uses machine learning to identify patterns of behavior that suggest poor quality data related to fraud or low attentiveness.

QualityScore is a critical aspect of our holistic quality approach

With Dynata you can expect extremely high respondent accuracy, reduced field times and less time spent cleaning the final data set – with clients saying they are saving up to 85% of time previously spent on manual data cleaning!

Dynata’s QualityScore™ technology can catch what standard industry checks consistently miss, by identifying patterns of behavior that suggest poor quality data related to fraud or low attentiveness. What’s more? Fraudulent and low-quality respondents are removed and replaced live during field.

97%

Accept rate across all projects after our data cleaning

175+

Passively tracked behaviours

85%

Time saved from manual data cleaning

Benefits to using QualityScore

Confidence

QualityScore™ provides a consistent and reliable way to determine which respondents should be included in the final data set.

Time Savings

QualityScore™ reduces the time research teams spend manually cleaning data by removing respondents below a certain quality threshold before they are even counted as a complete. Automated backfilling of disqualified respondents in real time maintains quotas and eliminates a time-consuming return to the field to augment with new respondents.

Fraud Blocking

QualityScore™ monitors in-survey behavior — such as the standard checks for speeding, straight-lining, and open-ends — as well as hundreds of passive indicators — such as mouse movement, survey acceleration, hot-key usage, time on page, typing speed, page translation. QualityScore™ can catch sophisticated fraud that is difficult to spot with behavioral checks alone and uses this information to differentiate partial disengagement from complete disengagement (which can be a data quality concern).

Reduce Bias

QualityScore™ ensures your data are cleaned consistently from project to project and wave to wave within a tracker. By providing an objective score to help evaluate the quality of a respondent, you can be confident that your data are being cleaned consistently over time. This also helps to avoid human biases that could skew the findings, such as the ad hoc removal of certain respondent types.

Easy-to-use

QualityScore™ applies AI/machine learning to every end-to-end study Dynata hosts, placing respondents into three tiers: rejected; on the bubble, which could require a manual check; and good. Those that fall below key thresholds are automatically rejected and replaced in real time, and not counted against final quotas.

QualityScore is used against our panel, and on a survey-level as part of Dynata’s iron-clad quality approach

How it works

Uses AI to identify poorly engaged respondents
alongside the elimination of sophisticated fraud.

Evaluates >175 passively tracked behaviors
(copy/paste, illogical responses, network speed and latency to name a few data points) to determine

Identifies levels of engagement to determine qualification

Detects and removes ghost completes, bots & fraudster

Connect with the team about QualityScore

Learn more about Dynata’s data quality approach and more about how QualityScore™ works by connecting with us.