PureSpectrum Updates PureScore System

US research tech and marketplace provider PureSpectrum has rolled out a new version of its PureScore respondent-level quality scoring system. New to version 6.0 are expanded behavioral and device intelligence capabilities, boosting its ability to spot the full range of unnatural and fraudulent behavior.

The company was founded in 2016 by former Cint US MD Michael McCrary, as a marketplace for buyers and sellers of online sample. The PureScore respondent-level scoring system was first launched at the beginning of 2020, as an advanced machine learning-driven system designed to block 'low-quality' respondents (scoring five or less out of ten) from taking surveys. Further features were added in 2023, and the latest update is the second in just fifteen months, reflecting the need to keep pace with the rise of AI and the proliferation of ways in which AI tools can intersect with online surveys.

Version 5.0 introduced continuity tracking, to detect mid-session handoffs from human respondents to AI agents. Version 6.0 builds directly on this, analyzing interaction patterns including typing cadence, mouse movement and navigation flow to assess whether behavior reflects genuine human engagement, regardless of whether a hand-off occurred. The firm says this allows PureScore to identify a broader range of quality issues, from AI-assisted fraud to fraudulent human behavior, distinguishing both from 'well-intentioned respondents participating in good faith.' PureScore is now embedded directly into users' own surveys via its own In-Survey Script.

Version 6.0 also features improved device intelligence, for better assessment of the device and network behind each session. 'The question we're always asking is whether the data reflects what real, genuinely engaged people think,' says CPO Mark Menig (pictured). '6.0 gives us a sharper answer to that question by expanding what we can see in both behavior and device signals.'

The firm's home page is at www.purespectrum.com .