Description
Participants see a name and must click the matching icon among distractors as fast and accurately as possible, giving speed and accuracy per icon and style.
A speeded visual-search experiment for icons. Each trial shows a prompt word and an array of icons; the participant selects the match as quickly and accurately as possible. Accuracy and response time are the outcome variables; semantic studies stress accuracy, visual-style studies stress speed. Arledge built 20 concrete, unambiguously named icons in filled and outline styles and two polarities (80 stimuli), presented all 20 around the prompt in a randomly ordered circle, timed trials client-side in milliseconds and randomised positions to prevent position learning. Success was analysed with chi-square and time with 2x2 ANOVA and Tukey tests, overall and per icon, then icons were cross-ranked by speed and accuracy. With 1,260 participants, outline white-on-black icons were about 170 ms slower. It differs from comprehension tests, which ask what an icon means.
Attributes
- Discipline
- Usability & Evaluation
- Loss type
- Siloed
- Stage
- Design
- Data
- Quantitative
- Effort
- Weeks
Steps
- Choose concrete icons with unambiguous names to limit semantic effects.
- Create every icon in each level of the visual factors under test at consistent size and line width, checking a minimum visual angle of about 0.7 degrees.
- Pre-load stimuli and time each trial from array reveal to selection; show an untimed orientation screen of all icons with names.
- On each trial show the prompt word surrounded by all icons in random, evenly spaced positions; record time and correctness, and record which wrong icon was chosen.
- Run hidden warm-up trials, then balanced test trials in random order so each condition appears equally often.
- Remove failed tests, warm-ups and time outliers; analyse success with chi-square and time with ANOVA and post-hoc tests, overall and per icon, and check trial-order effects.
- Cross-rank icons by median time and success rate to flag slow or confusable icons.
Sources
- Filled-in vs. Outline Icons (Arledge, 2014)
Also known as
- Search and select experiment
- Icon search task
- Icon recognition test
Categories
- User Research
- Usability & Evaluation