Age-of-Acquisition
Lexical Decision Task
— Free to Take
A short browser-based reaction-time test. A letter string appears; you decide as fast as you can whether it is a real English word or a made-up non-word. Words you learned as a toddler are recognised faster than words you learned in school — this task measures that speed difference in milliseconds. Free with a free account.
What the task measures
The age-of-acquisition (AoA) effect is one of the most robust findings in psycholinguistics: words acquired earlier in life are recognised faster in a lexical decision task, even when word length and frequency are held constant (Brysbaert & Ellis, 2016). The effect is typically 20–80 ms and reflects the differential consolidation of early-acquired words in the mental lexicon.
This task uses 40 early-acquired words (mean AoA ≤ 4 years; e.g. ball, tree) and 40 late-acquired words (mean AoA ≥ 9 years; e.g. axiom, epoch), drawn from the Kuperman et al. (2012) norming study. Forty legal pseudowords require the same binary decision.
This is an experimental research paradigm, not a standardized test. Unlike our Semantic Priming LDT and ELP Lexical Decision Task, whose stimuli and design reproduce the original published studies verbatim, this item set was assembled by ImplicitifyAI from published open-access resources for this platform. There is no "original authors' version" of this exact task, and no published reference data exist for it.
Stimulus construction and lexical controls
Following the convention of documenting lexical controls in stimulus construction, the two word sets are controlled as follows. Length is matched: both sets span 4–7 letters with closely matched mean length (early 4.8 vs. late 4.9 letters). Word frequency is not matched: early-acquired words are, as in natural language, also more frequent (mean SUBTLEX-US Zipf 5.1 for the early set vs. 4.1 for the late set). The reported "AoA effect" therefore reflects the natural compound of acquisition age and frequency, not a frequency-controlled AoA estimate as in factorial studies such as Brysbaert & Ellis (2016). Non-words are orthographically legal, constructed following ARC Nonword Database conventions (Rastle et al., 2002), and matched to the words on approximate length.
From your trials we report:
- Mean reaction time for early-acquired words
- Mean reaction time for late-acquired words
- Mean reaction time for nonwords
- The AoA effect — how many milliseconds faster you are on early words
- Accuracy separately for early words, late words, and nonwords
These are descriptive reaction times in milliseconds. No percentile and no T-score are reported: no citable per-subject RT distribution for this exact AoA-band design is available in the published literature, so converting your raw times to a percentile would require fabricated norms. Your AoA effect is the honest within-subject measure.
Common questions
What is the age-of-acquisition effect?
Words learned earlier in life are processed faster on average — in controlled studies, even when word frequency and length are held constant. This AoA effect persists in adults and reflects how deeply early words are consolidated in the mental lexicon. It was first systematically documented by Carroll and White (1973) and has been replicated across languages and tasks. In this task the early and late sets are length-matched but not frequency-matched, so the contrast you see combines acquisition age with frequency.
Is this a standardized test?
No. It is an experimental research paradigm built by ImplicitifyAI from published open-access resources (Kuperman et al., 2012 AoA ratings; SUBTLEX-US frequencies; ARC-convention non-words). It has no original published version, no norms, and no validation study of this exact item set. Results are descriptive within-person reaction times only.
Where do the word norms come from?
AoA ratings are from Kuperman, Stadthagen-Gonzalez, and Brysbaert (2012), an open-access norming study of 30,000 English words collected from 1,960 raters. Word frequency is from SUBTLEX-US (Brysbaert & New, 2009). Both datasets are publicly available.
Why is there no percentile or IQ-style score?
Reaction times in browser-based tasks depend on your display, keyboard hardware, and browser timing. More importantly, no published study provides a per-subject RT distribution specifically for this AoA-band design and trial count. Reporting a percentile would require inventing a reference distribution — something we do not do. Your AoA effect in milliseconds is the scientifically honest summary.
Is it really free?
Yes. The task is free to take with a free account. Create one in under a minute and your results are shown immediately at the end.
How is this different from the semantic priming LDT?
The Semantic Priming LDT presents a prime word before each target and measures how much related primes speed recognition. This AoA task presents targets directly (no prime word) and measures how word-acquisition age affects baseline recognition speed. The two tasks are complementary but measure different aspects of lexical access.