Flanker Task Online — Selective Attention Test
Press the direction of the center arrow while ignoring the four flanking arrows — 5 practice trials, then 40 scored trials measure your flanker effect in milliseconds.
What this test measures
Selective attention and interference control — your ability to respond to a target while suppressing conflicting information right next to it. Each trial shows five arrows after a 500ms fixation cross; press the direction of the CENTER arrow within 2 seconds, using the arrow keys or the on-screen buttons, and ignore the four flankers. In congruent trials all arrows agree (← ← ← ← ←); in incongruent trials the flankers point the other way (← ← → ← ←). After 5 unscored practice trials you complete 40 scored trials — 20 congruent and 20 incongruent in shuffled order. Your score is the flanker effect: mean reaction time on correct incongruent trials minus correct congruent trials, in milliseconds. Lower is better, no response within 2 seconds counts as an error, and if accuracy drops below 70% your percentile is capped — you can't trade guessing for speed.
The science behind it
The arrow flanker task descends directly from Eriksen & Eriksen 1974 (Perception & Psychophysics), one of the most cited paradigms in attention research. The core finding has replicated for five decades: responses to a target surrounded by incompatible flankers are reliably slower, because the conflicting flankers activate the competing response, which must be suppressed. Neuroimaging ties this conflict resolution to the anterior cingulate and prefrontal control network (Botvinick et al. 2001, Psychological Review), and a flanker subtest anchors the Attention Network Test (Fan et al. 2002). Modern replications put the typical arrow-flanker effect around 65ms with test-retest reliability near 0.80 — one of the more stable interference measures. Green & Bavelier 2003 (Nature) used a flanker compatibility task to show action gamers distribute visual attention differently from non-gamers.
How to improve interference control
Honest answer: interference control is one of the harder abilities to train directly. Practicing the flanker task itself mainly speeds up your overall reaction time; the interference effect shrinks much less, and reviews of commercial brain training find little far transfer (Simons et al. 2016, Psychological Science in the Public Interest). What does help, per the evidence: aerobic exercise reliably benefits executive control (Hillman, Erickson & Kramer 2008, Nature Reviews Neuroscience); short mindfulness programs improved attention-network scores in randomized work (Tang et al. 2007, PNAS); and sleep deprivation measurably worsens conflict resolution, so test rested. Action video game players show attention advantages in this family of tasks (Dye, Green & Bavelier 2009), though self-selection explains part of that gap. Expect gradual gains, not miracles.
Frequently Asked Questions
What is a good flanker effect score?
Against our literature-based norms (mean 65ms, SD 25ms), an effect under 40ms puts you roughly in the top 16%, and around 25ms is roughly the top 5%. A smaller effect means the conflicting arrows barely slow you down. The percentile only counts fully when your accuracy stays above 70%, so a tiny effect built on guessing won't rank well.
What do congruent and incongruent trials mean?
In congruent trials the flanking arrows point the same way as the center target (← ← ← ← ←); in incongruent trials they point the opposite way (← ← → ← ←). You do 20 scored trials of each. Your flanker effect is the difference in mean reaction time between the two conditions, computed from correct responses only.
Why is my flanker effect 0 or negative?
A negative effect — responding faster when the flankers conflict — is theoretically not expected and is almost always measurement noise from a limited trial count. We treat negative values as 0 when computing your percentile. Forty scored trials give a reasonable estimate, but lab studies typically use hundreds of trials to pin the effect down precisely.
Why was my score capped for low accuracy?
If your accuracy falls below 70%, your percentile is capped at 30. The flanker effect is only meaningful when computed from genuine correct responses — rushed or random answering produces a meaningless reaction-time difference. The cap stops speed-for-accuracy trading; note that not responding within 2 seconds also counts as an error.
Can I use this flanker task online for a psychology class?
Yes — this is a browser-based arrow version of the classic Eriksen & Eriksen 1974 paradigm: 5 practice plus 40 scored trials (20 congruent, 20 incongruent), a 500ms fixation cross, and a 2-second response window, free with nothing to install. It works well for demonstrations and coursework, but browser timing adds a few milliseconds of jitter compared with lab software, and formal research typically uses larger trial counts.
More Cognitive Tests
Color words flash in sometimes-mismatched ink — respond to the INK color, not the word. 40 scored trials measure your Stroop effect: the interference cost in milliseconds.
Press for green circles, hold back for red — 45 rapid trials (5 practice + 40 scored, about 75% green) build an urge to press that your score depends on resisting.
Click when the screen turns green — and don't fall for the distractor colors. 20 rounds measure your average visual reaction time in milliseconds.