Lexical Decision Task

Version: v1 (current)

A fundamental word recognition paradigm measuring the speed and accuracy of accessing stored lexical representations.

Overview

The Lexical Decision Task is one of the most widely used paradigms in psycholinguistics and cognitive psychology. Participants view letter strings and must quickly decide whether each is a real word (e.g., "TABLE") or a pronounceable non-word (e.g., "BLINT"). This simple binary decision taps into core processes of lexical access, word recognition, and semantic memory retrieval.

Response times reveal the automatic and rapid nature of word recognition: real words are typically identified in 500-700ms, and various word properties (frequency, length, neighborhood density) systematically affect decision speed. The task is sensitive to individual differences in reading ability, vocabulary knowledge, and language proficiency.

Scientific Background

Classic Findings

Word Frequency Effect: High-frequency words (e.g., "time") are recognized faster than low-frequency words (e.g., "whim"). This effect demonstrates that word access is influenced by how often we encounter words in language.

Lexicality Effect: Real words are recognized faster than non-words, even when non-words are pronounceable and follow spelling rules. This difference (typically 100-200ms) reflects the presence or absence of stored lexical representations.

Word Length Effect: Shorter words are typically processed faster than longer words, though the magnitude varies by language and orthographic transparency.

Neighborhood Density: Words with many similar-looking words (orthographic neighbors) can show facilitation or inhibition depending on task demands and language.

Key Mechanisms

Lexical Access: Rapid lookup in mental lexicon based on visual word form
Decision Criterion: Threshold for accepting a letter string as a familiar word
Familiarity vs. Identification: Task can be performed via global familiarity without full word identification

References

Meyer, D.E., & Schvaneveldt, R.W. (1971). Facilitation in recognizing pairs of words. Journal of Experimental Psychology, 90(2), 227-234.
Balota, D.A., & Chumbley, J.I. (1984). Are lexical decisions a good measure of lexical access? Journal of Experimental Psychology: Human Perception and Performance, 10(3), 340-357.

Why Researchers Use This Task

Reading Research: Investigate word recognition processes and reading development
Semantic Memory: Study organization of concepts and word associations
Language Proficiency: Assess vocabulary knowledge and reading fluency
Clinical Assessment: Evaluate reading disorders, dyslexia, and language impairments
Bilingualism: Compare lexical access across languages and proficiency levels

Configuration Options

Response Mode

Parameter	Default	Description
Time-based mode	On	If enabled, trials auto-advance after timeout; if disabled, participant must click button to continue

Visual Parameters

Parameter	Default	Range	Description
Font size	48px	8-400px	Size of letter strings

Practice Trials

Parameter	Default	Description
Enable practice	Off	Turn on to include practice trials with feedback

When practice is enabled:

Participants see visual feedback after each practice response (green checkmark for correct, red X for incorrect)
Helps participants learn to distinguish words from non-words
Main trials begin after practice is complete

Keyboard Shortcuts

Researchers can customize the keyboard bindings used during the task:

Parameter	Type	Default	Description
Show keyboard hint	boolean	True	Display an on-screen hint showing the configured keys
Word key	key	W	Key for "word" response
Word action label	text	"Word"	Label shown in the keyboard hint
Nonword key	key	N	Key for "nonword" response
Nonword action label	text	"Nonword"	Label shown in the keyboard hint

Trial Configuration

Configure trials in the spreadsheet with the following columns:

Column	Description	Example
stimulus	The letter string to display	"TABLE" or "BLINT"
lexicality	Whether it's a word or non-word	"word" or "nonword"
fixation_ms	Fixation cross duration	500
stimulus_ms	How long to display (0 = until response)	2000
iti_ms	Inter-trial interval (optional - empty cells use general setting)	500
block	Optional grouping label	"high_frequency"

Example Trials

| stimulus | lexicality | fixation_ms | stimulus_ms | block          |
|----------|------------|-------------|-------------|----------------|
| TABLE    | word       | 500         | 2000        | high_frequency |
| CHAIR    | word       | 500         | 2000        | high_frequency |
| WHIM     | word       | 500         | 2000        | low_frequency  |
| BLINT    | nonword    | 500         | 2000        | nonwords       |
| TRALE    | nonword    | 500         | 2000        | nonwords       |

Creating Non-Words

Pseudowords are the standard type of non-word:

Created by changing 1-2 letters of real words (e.g., "table" → "tible")
Follow spelling rules of the language (legal letter combinations)
Pronounceable like real words
Matched to real words on length and structure

Important: Avoid pseudohomophones (non-words that sound like words, e.g., "brane" for "brain") unless specifically studying phonology, as they are harder to reject and introduce additional processing demands.

Participant Experience

Trial Sequence

Main Instructions: Explains to respond "word" for real words, "nonword" for made-up words
(Optional) Practice Instructions: If practice enabled
(Optional) Practice Trials: With feedback (checkmark/X)
(Optional) Trials Instructions: Before main trials
Main Trials: Each trial follows:
- Fixation cross appears
- Letter string appears (e.g., "TABLE")
- Participant responds via keyboard or button
- Trial advances to next

Response Methods

Keyboard (recommended):

Press W key for "word" (default -- configurable by researcher)
Press N key for "nonword" (default -- configurable by researcher)

Buttons (if time-based mode disabled):

Click "Word" or "Nonword" button

Both methods record response time from stimulus onset to response.

All keyboard bindings are configurable by the researcher in the study configuration. The keys listed above are the defaults.

Data Output

Markers and Responses

The task records high-resolution timestamps in two separate collections:

Markers (stimulus_shown):

{
  "type": "stimulus_shown",
  "ts": "2024-01-01T00:00:01.000Z",
  "hr": 1234.56,
  "data": {
    "trial_index": 1,
    "stimulus_id": "lexical_decision_0_1",
    "stimulus": "TABLE",
    "lexicality": "word",
    "block": "high_frequency",
    "is_practice": false
  }
}

Response Data:

{
  "trial_index": 1,
  "stimulus_id": "lexical_decision_0_1",
  "source": "keyboard",
  "raw_key": "w",
  "stimulus": "TABLE",
  "lexicality": "word",
  "response_value": "word",
  "response_correct": true,
  "latency_ms": 623,
  "block": "high_frequency",
  "is_practice": false
}

Summary Artifact

A JSON file (ld_summary_<taskIndex>.json) with aggregated statistics:

{
  "task_kind": "lexical_decision",
  "task_index": 0,
  "total_trials": 40,
  "overall": {
    "total": 40,
    "valid_responses": 38,
    "correct": 35,
    "accuracy": 0.921,
    "mean_rt_ms": 687,
    "mean_correct_rt_ms": 654,
    "timeouts": 2
  },
  "by_lexicality": {
    "word": {
      "total": 20,
      "accuracy": 0.947,
      "mean_correct_rt_ms": 598
    },
    "nonword": {
      "total": 20,
      "accuracy": 0.895,
      "mean_correct_rt_ms": 710
    }
  },
  "lexicality_effect_ms": 112,
  "trials": [ /* per-trial data */ ]
}

Key metrics:

lexicality_effect_ms: Difference in mean correct RT between nonwords and words
accuracy: Proportion of correct responses overall and by lexicality
mean_correct_rt_ms: Average reaction time for correct trials only

Instructions

The task uses four types of instructions you can customize:

Main Instructions: Shown on a dedicated page before the task begins. Explain the word/non-word distinction with examples.
Hint Instructions: Available via the "?" button during the task. Provide quick reminders.
Practice Instructions (if practice enabled): Shown before practice trials. Explain that this is practice with feedback.
Trials Instructions (if practice enabled): Shown after practice, before main trials. Note that feedback will no longer be shown.

All instructions support rich text formatting and can be customized for your study.

Design Recommendations

General Guidelines

Balance: Use equal numbers of words and non-words (50/50) to prevent response bias
Non-word Quality: Non-words should be pronounceable and follow spelling rules
Word Selection: Control for frequency, length, and concreteness depending on research question
Trials: Minimum 40-60 trials (20-30 per type) for stable effects; 80-120 trials typical for research

Creating Matched Non-Words

Match non-words to your word list on:

Length: Same number of letters
Syllable count: Similar complexity
Orthographic structure: Similar letter patterns
Initial letters: Balanced across alphabet

Avoid:

Letter combinations that don't exist in the language (e.g., "qxvz")
Pseudohomophones unless studying phonology
Non-words that are too easy to reject (e.g., obviously foreign words)

Timing Recommendations

Parameter	Standard	Fast-Paced	Self-Paced
Fixation	500ms	300ms	500ms
Stimulus Duration	2000ms	1500ms	0 (unlimited)

Most participants respond within 1000ms. Shorter durations add time pressure; unlimited duration removes it.

Population-Specific Adaptations

Children (8-12 years):

Use high-frequency, age-appropriate words
Obvious non-words (e.g., "gribble")
Larger font (60-72px)
Self-paced responses (no timeout)
Extensive practice (20 trials)

Older Adults (65+):

Larger font (60px)
Generous response window (3000ms+)
Age-appropriate vocabulary
Clear instructions about response keys

Non-Native Speakers:

Adjust word difficulty to proficiency level
Use concrete, imageable words
Allow longer response times
May show reduced frequency effects if vocabulary is limited

Common Issues and Solutions

Issue	Possible Cause	Solution
Low accuracy (<80%)	Non-words too subtle, task confusing	Use clearer non-words, improve instructions, add practice
No frequency effect	Insufficient frequency range, too few trials	Use wider frequency contrast, increase trials per condition
Response bias (favoring one response)	Unbalanced word/non-word ratio	Ensure 50/50 balance
Very slow RTs (>1200ms)	No time pressure	Reduce stimulus duration, add speed emphasis to instructions
Very fast RTs (<400ms)	Guessing	Increase task difficulty, use more subtle non-words

Example Study Configurations

Standard Word Recognition Study

50 words (25 high-frequency, 25 low-frequency)
50 non-words matched on length and structure
500ms fixation, 2000ms stimulus
Keyboard responses (W/N keys)
Analyze: Frequency effect, accuracy, lexicality effect

Reading Fluency Assessment

40 graded words (easy → difficult)
40 matched non-words
Faster pace (1500ms timeout) to measure fluency
Button responses for simplicity
Analyze: Accuracy by difficulty, overall RT

Clinical Screening

30 high-frequency words
30 obvious non-words
Self-paced (no timeout)
Larger font (60px)
Extensive practice
Analyze: Basic word recognition ability, error patterns

Overview​

Scientific Background​

Classic Findings​

Key Mechanisms​

References​

Why Researchers Use This Task​

Configuration Options​

Response Mode​

Visual Parameters​

Practice Trials​

Keyboard Shortcuts​

Trial Configuration​

Example Trials​

Creating Non-Words​

Participant Experience​

Trial Sequence​

Response Methods​

Data Output​

Markers and Responses​

Summary Artifact​

Instructions​

Design Recommendations​

General Guidelines​

Creating Matched Non-Words​

Timing Recommendations​

Population-Specific Adaptations​

Common Issues and Solutions​

Example Study Configurations​

Standard Word Recognition Study​

Reading Fluency Assessment​

Clinical Screening​

See Also​