Elim-First attribute identification

Description

Identifies the attributes users apply first to eliminate options before any trade-off, so designs surface and satisfy these screening criteria.

Tversky's elimination by aspects model describes choice as sequential: people pick an important attribute, discard all options that fail it, and repeat until one option remains. It is non-compensatory, so excellence on other attributes cannot rescue an option eliminated early. Elim-First attribute identification finds which attributes users eliminate on first and the cut-offs they use. In UX this explains why a technically superior option is never considered: it failed a screening aspect such as 'works offline', 'in my language' or 'no upfront fee' before comparison began. Rigour comes from process-tracing methods (information display boards, think-aloud, eye tracking or click-stream order) rather than stated importance ratings, which systematically confuse screening and trade-off attributes.

Attributes

Loss type
Unmigrated
Data
Mixed
Effort
Weeks

Steps

  1. List 8 to 20 candidate attributes for the decision from interviews and competitor analysis, defining each with clear levels.
  2. Build an information display board or mock comparison interface where attribute values are hidden until clicked, so search order and stopping can be logged.
  3. Run 15 to 40 participants through 3 to 6 realistic choice tasks while thinking aloud, logging which attribute they open first, which options they discard and why.
  4. Classify each attribute by whether it was used to eliminate options, the order of use, and the stated cut-off (for example 'more than 2% fee, gone').
  5. Rank attributes by elimination frequency and median position in the search sequence; attributes used for elimination by more than about 40% of participants in the first two steps are screening attributes.
  6. Validate with a larger survey using a screening question format (for example 'Would you consider an option that...?') and check the design surfaces screening attributes early and visibly.

Sources

  • Heuristics & Biases (Kahneman ed.)
  • Judgment Under Uncertainty
  • Choices Values & Frames (Kahneman/Tversky)
  • The Inquiry Series, Ch 21 — Heuristics, Biases & the Architecture of Judgment

Also known as

  • Elimination-by-aspects analysis
  • EBA attribute screening
  • Non-compensatory screening attribute identification

Categories

  • Analytics & Behaviour
  • Behavioral Science & Economics