First fixation duration (FFD) is an eye-tracking measure of how long a reader’s first fixation on a target word lasts. If the reader fixates the word twice during the first pass, FFD counts only the first stop, while gaze duration includes both. The measure helps researchers investigate early visual and lexical word processing, but cannot by itself establish whether the person understood the word or its sentence.
Did you know that a pupil can recognise a printed word quickly and still misunderstand the sense chosen by the paragraph? The eyes provide measurable information about where and when words are inspected, while comprehension depends on grammar, context and concept knowledge. FFD is one carefully defined observation within that larger system. Vocabulary becomes practical technology when visual recognition leads to an accurate explanation, not merely a short number on an eye-tracking graph.
This guide explains first fixation, single fixation, gaze duration, refixation probability, skipping and regression path duration with original arithmetic examples. It draws on Rayner’s reading research, the E-Z Reader framework and 2025 methodological guides, examining why word frequency, length, preview and motor targeting affect interpretation. It also provides Primary and Secondary English, Mathematics, Science, History, GP and translation applications.
Editorial note: All fixation times in classroom calculations are invented to explain the formula. No child was measured or diagnosed, and scientific claims are linked to identifiable research.
Contents: what does the first eye stop really measure?
Open 58 chapters, calculation checks and classroom cases
- 1. The first fixation is one measurable pause
- 2. Why vocabulary scientists measure the first stop
- 3. The eye fixates between rapid saccades
- 4. First fixation has a position as well as a time
- 5. The earliest landing is the key event
- 6. An invented two-fixation example
- 7. A single fixation case behaves differently
- 8. Single fixation duration uses a subset of trials
- 9. Gaze duration includes more of the first pass
- 10. Refixation probability is not a duration
- 11. Regression path duration includes backward detours
- 12. Total viewing time counts later returns
- 13. A skipped word has no initial target fixation
- 14. The denominator for an average is not trivial
- 15. Eye trackers need consistent fixation detection
- 16. Calibration can move a fixation between words
- 17. Word boundaries must be defined in the display
- 18. First fixation can be surprisingly short
- 19. An initial fixation can be prolonged by visual conditions
- 20. Word length can affect early sampling
- 21. Word frequency is a well-studied predictor
- 22. Contextual predictability is a different factor
- 23. A fixation is not a lie detector
- 24. Parafoveal preview precedes direct fixation
- 25. The boundary paradigm changes what appears next
- 26. Parafoveal-on-foveal claims concern an earlier region
- 27. Optimal viewing position changes information access
- 28. The inverted optimal viewing effect deserves caution
- 29. Launch site and targeting error influence duration
- 30. Return sweeps create special first-fixation paths
- 31. Undersweep fixations can be linguistically informative
- 32. Reading task changes the interpretation
- 33. Reading aloud adds articulation demands
- 34. Eye tracking and self-paced reading are different
- 35. The N400 is an electrophysiological measure
- 36. The P600 is also a distinct signal
- 37. The 1998 Rayner review provides broad context
- 38. The E-Z Reader model predicts multiple outcomes
- 39. The 2025 methods guide defines FFD directly
- 40. The 2025 fundamentals paper adds a counting framework
- 41. Multiple comparisons can create false discoveries
- 42. A mean hides the response distribution
- 43. Participant and item variability both matter
- 44. A careful apparatus check can prevent misleading data
- 45. An original three-outcome table
- 46. A correction may be useful rather than slow
- 47. Primary English: recognise and explain the word
- 48. Primary English: do not rush negation
- 49. Secondary English: one sense is not the whole word
- 50. Secondary English: ambiguity may be resolved later
- 51. Mathematics: the whole condition governs the answer
- 52. Science: a long technical word needs a real definition
- 53. History: verify names and source wording
- 54. General Paper: fluency is not truth
- 55. Translation: written similarity can deceive
- 56. Professional instructions: a second look can prevent error
- 57. A reproducible five-column data sheet
- 58. The final lesson: first is not everything
Quick Test: the same word can produce several measures
| Observed eye path | First fixation duration | Gaze duration |
|---|---|---|
| One 210-ms fixation, then forward exit | 210 ms | 210 ms |
| 180-ms fixation, 140-ms refixation, then exit | 180 ms | 320 ms |
| 160-ms fixation, backward exit to earlier text | 160 ms | 160 ms before first exit |
| Word skipped on the initial pass | No eligible first fixation | No eligible first-pass target gaze |
These are fictional, not real recordings. The table illustrates how one source of fixation data can produce several measures and why skipped targets should not simply be assigned zero milliseconds.
1. The first fixation is one measurable pause
First fixation duration, or FFD, is the elapsed time of the first valid eye fixation on a designated word or region in a reading experiment. It is measured in time units such as milliseconds and usually belongs to the reader’s first encounter with that target during the relevant pass.
The measure is precise about what was observed: where the eyes first stopped and how long that stop lasted. It is not a direct measure of the time needed to understand every meaning associated with the word. A fixation can include visual sampling, partial lexical processing and preparation for the next eye movement.
2. Why vocabulary scientists measure the first stop
Readers recognise printed word forms at different speeds depending on experience, spelling, context and the task. A controlled eye-tracking experiment can ask whether changing a particular lexical property alters the duration of the first fixation.
That makes FFD useful as one piece of evidence about word recognition. Its meaning depends on the comparison and the participant’s eye path. A single short fixation does not establish that the child knows a rare term deeply; a long fixation does not prove a vocabulary weakness.
3. The eye fixates between rapid saccades
During reading, the eyes typically alternate relatively stable fixations and rapid saccades. A fixation provides a period of high-quality visual sampling of the current area, while the next movement relocates gaze to another part of the text.
The temporal boundaries of a fixation must be defined by an eye tracker and analysis algorithm. A human observer cannot accurately infer millisecond-level FFD simply by looking at another person’s face. The method’s instrumentation is part of its scientific value.
4. First fixation has a position as well as a time
A valid fixation has a spatial location on the display and an interval in time. Researchers must first determine whether the fixation landed inside the chosen word region, then calculate how long it lasted.
A fixation assigned to the wrong word creates a misleading duration record. Calibration error, tight word spacing and line transitions complicate classification. A reliable report states how gaze coordinates were mapped to text regions before interpreting lexical effects.
5. The earliest landing is the key event
For a target with multiple first-pass fixations, FFD counts only the first fixation. It stops when that initial fixation ends and the eye makes the next movement, even if the next movement stays within the word.
This is how FFD differs from gaze duration, which adds additional first-pass fixations on the target. The two values can be related but need not be equal. A student should inspect the actual sequence rather than assume ‘reading time’ has one meaning.
6. An invented two-fixation example
Imagine a fictional reader who first fixates the word garden for 180 milliseconds, then moves within that word and fixates for another 140 milliseconds before moving to the next word.
The first fixation duration is 180 ms. Gaze duration is 180 + 140 = 320 ms. These figures are invented solely to teach the definitions. They are not data from a real reader or a benchmark for school performance.
7. A single fixation case behaves differently
In another fictional trial, the reader fixates the target word exactly once for 210 milliseconds and then leaves the word. The first fixation is also the entire first-pass gaze in that simple sequence.
Here FFD and gaze duration are both 210 ms. The equality is a consequence of one observed fixation, not proof that the word was fully understood. A research measure must distinguish the number of events counted from their psychological interpretation.
8. Single fixation duration uses a subset of trials
Single fixation duration is typically computed only for eligible trials where a word received exactly one fixation during the first pass. FFD, by contrast, can be measured on both single-fixation and multiple-fixation trials.
The distinction matters statistically. A group of targets requiring refixation can contribute FFD values but not ordinary single-fixation-duration values. These two datasets may differ in word length, familiarity or difficulty. An analysis should not treat their means as directly comparable without considering selection.
9. Gaze duration includes more of the first pass
Gaze duration sums every fixation on a word before the reader first exits it, under the specified region definition. It can include a first fixation and one or more refixations.
FFD is a narrower early time measure. Gaze duration incorporates additional inspection when needed. They can respond differently to word length or other lexical properties because some effects change fixation duration and others change the probability of making another fixation.
10. Refixation probability is not a duration
Refixation probability counts the fraction of eligible word encounters in which another fixation occurs on the same word, often before the first exit. Its unit is a proportion rather than milliseconds.
The probability can influence the average gaze duration while leaving the mean first fixation duration comparatively unchanged. A researcher should not confuse a twenty-percent refixation rate with a 200-millisecond first fixation. The vertical axis determines what a graph means.
11. Regression path duration includes backward detours
Regression path duration or go-past time often counts fixation durations from the first entry into a region until gaze first moves forward beyond it. Under an inclusive convention, backward fixations on earlier text can contribute.
FFD ends much sooner, at the end of the very first target fixation. A later regression can reflect sentence reanalysis without changing that earlier duration. The measures offer different views of the reading process and must not be collapsed into one number.
12. Total viewing time counts later returns
Total viewing time can include fixations on the target during later passes after the reader has moved beyond it. A pronoun ambiguity or difficult clause can cause the eyes to return and inspect the original word again.
That later reading may be important for final comprehension, but it is excluded from the initial FFD. A paper reporting only FFD cannot necessarily conclude that no subsequent difficulty occurred. The chronology of fixations matters.
13. A skipped word has no initial target fixation
Some printed words receive no direct fixation when first passed. If a target is skipped, its first fixation duration is not an observed zero; no eligible first-pass target fixation occurred.
Researchers may analyse skipping probability separately and restrict duration calculations to fixated targets. Treating skips as zero milliseconds can distort the average and change the meaning of the measure. The inclusion rule should be explicit.
14. The denominator for an average is not trivial
If a study begins with 100 target appearances and some are skipped or excluded for tracking error, the mean FFD should be calculated from the observations meeting the declared criteria.
A classroom arithmetic exercise can show the difference between all trials and fixated valid trials. The researcher must report exclusions before interpreting the result. A neat decimal conceals little unless the sample supporting it is clear.
15. Eye trackers need consistent fixation detection
Eye-tracking analysis software groups gaze samples into fixations based on spatial and temporal criteria. A very short event may be removed, merged or classified as a saccade under some settings.
Such preprocessing decisions can affect the measured first fixation duration. The 2025 methodology guides emphasise documenting these choices rather than using a single rigid recipe for every reading task. Instrument precision must be matched by procedural transparency.
16. Calibration can move a fixation between words
Small systematic tracking errors can place a recorded gaze point just inside an adjacent region. A first fixation that actually landed on word N+1 might be assigned to word N.
This is especially important with short words, small fonts or unusual landing positions. Researchers should inspect calibration quality and text geometry. A wrong region assignment can create an apparent lexical effect that has little to do with processing the intended target.
17. Word boundaries must be defined in the display
The word region can include just the visible letters, additional surrounding space or a polygon defined by the software. The choice affects whether border fixations qualify as initial target fixations.
A research report should state its region-of-interest definition. Two laboratories using different word boundaries can produce subtly different datasets. Scientific vocabulary is useful when it makes the measurement reconstructible rather than assumed.
18. First fixation can be surprisingly short
A brief fixation does not necessarily mean that a word was fully recognised and understood. The reader may have landed at an unintended location or a corrective movement may already be planned.
Some short fixations are associated with motor-control adjustments. A researcher who interprets every short interval as easy lexical access risks confusing eye-movement mechanics with language cognition. FFD is informative but not self-explanatory.
19. An initial fixation can be prolonged by visual conditions
Low contrast, tiny print, unfamiliar fonts or crowded letter spacing can slow visual sampling independently of word meaning. Reading material presentation is therefore a relevant variable in fixation experiments.
If two conditions differ in typography as well as vocabulary, the FFD difference cannot safely be attributed to lexical complexity. A teacher should first check whether a worksheet is readable before concluding that the pupil finds its words conceptually difficult.
20. Word length can affect early sampling
Longer words place letters at different distances from a landing position and may change the amount of information acquired from one fixation. They can also increase the need for additional fixations.
A study of FFD should examine length separately from gaze duration and refixation probability. A long technical word may still be familiar to a specialist. The number of letters does not establish the difficulty of the underlying concept.
21. Word frequency is a well-studied predictor
Higher-frequency words often receive more efficient recognition under comparable conditions. Eye-movement studies use FFD and other measures to investigate whether frequency contributes to early reading behaviour.
The effect is probabilistic, and frequency estimates depend on the corpus. A rare scientific term may be familiar to a particular pupil through repeated subject study. A word’s population count is not equivalent to an individual’s knowledge.
22. Contextual predictability is a different factor
Sentence context can make an upcoming word more or less likely. A highly predictable short word may be recognised rapidly, whereas an unexpected but valid completion may require more interpretive work.
Researchers may use cloze norms to estimate expectations. A target’s word frequency and its cloze probability must be kept distinct. One asks how common the word is across sampled language; the other asks how often people supply it after this specific context.
23. A fixation is not a lie detector
A person may read a common false statement quickly because its words and phrasing are familiar. A surprising true statement may take longer because it introduces a new concept.
FFD cannot verify the truth of a proposition. Schoolchildren must learn to distinguish ease of recognition from factual evidence. Vocabulary supports critical interpretation only when meanings and sources are checked separately.
24. Parafoveal preview precedes direct fixation
Before the eye lands on a word, the reader may obtain limited information from parafoveal vision. That can influence processing when the first target fixation occurs.
A FFD difference between preview conditions may therefore reflect information acquired during an earlier fixation, not only processing after the target was directly viewed. The boundary paradigm can manipulate preview information to test such hypotheses. Reading is an unfolding system across movements.
25. The boundary paradigm changes what appears next
In a gaze-contingent boundary experiment, an upcoming target location shows one preview before the eye crosses an invisible coordinate. The actual target replaces it during the saccade.
Researchers can measure FFD on the target after the change and compare conditions. This helps examine which prior visual information facilitated target processing. A reported difference depends on accurate timing and valid eye paths, not merely the diagram of the experiment.
26. Parafoveal-on-foveal claims concern an earlier region
A parafoveal-on-foveal effect is a proposed influence of upcoming word N+1 on the fixation duration of current word N. FFD on N+1 is a different measure from FFD on N.
This distinction prevents a common research error. A preview benefit at the target does not automatically prove a change on the preceding word. The experimental report must identify which word received the measured first fixation.
27. Optimal viewing position changes information access
Where the eye first lands within a word affects which letters lie nearest high-resolution central vision. Under some tasks, centre-near landing positions support more efficient identification than extreme edge landings.
FFD may interact with landing position, but a long first fixation does not uniquely reveal a poorly chosen location. A researcher should record the position as well as the duration. Visual geometry and lexical processing both contribute to the result.
28. The inverted optimal viewing effect deserves caution
Some studies report longer initial fixations for centre-near landing positions than for certain edge landings, even though central positions often reduce refixation probability. This is sometimes described as an inverted optimal viewing position effect.
One proposed account involves short corrective fixations after mislocated eye movements at word edges. The finding illustrates why longer initial fixation does not always imply less efficient word recognition. Different dependent measures can show different patterns.
29. Launch site and targeting error influence duration
The preceding fixation determines the launch site for the saccade that reaches the target. Long or inaccurate movements can produce landing errors and rapid corrections.
A first fixation observed near a word boundary might reflect motor mislocation as well as lexical processing. A valid analysis considers launch distance and landing position. Without this geometry, a frequency-related timing claim may be vulnerable to an alternative explanation.
30. Return sweeps create special first-fixation paths
When readers move from the end of one printed line to the next, the long return sweep may undershoot the true line beginning. A brief undersweep fixation can occur before a corrective saccade.
Such events complicate conventional first-pass word timing because a later word on the new line may be encountered before the first intended word. Researchers may treat return-sweep events separately. An ordinary FFD rule cannot simply be applied without checking the actual fixation sequence.
31. Undersweep fixations can be linguistically informative
Research by Slattery and Parker and later by Parker, Kirkby and Slattery found that short undersweep fixations can influence subsequent reading of text on the line under the studied conditions.
That does not mean the first brief landing always conveys a complete word meaning. It shows that motor corrections and visual sampling can overlap. For FFD analyses, the researcher needs to know whether the event was an ordinary target fixation or a return-sweep correction.
32. Reading task changes the interpretation
Reading a novel for comprehension, visually searching a document for one name and proofreading technical instructions place different demands on attention and lexical recognition.
An FFD measured during visual search should not automatically be compared with one measured during natural sentence reading. The experimenter’s instructions and the participant’s goal must be named. A vocabulary assessment should test the actual skill of interest rather than borrow a metric indiscriminately.
33. Reading aloud adds articulation demands
A participant who reads words aloud must prepare pronunciations and coordinate speech with the eyes. Silent reading normally does not impose the same external production requirement.
Eye-movement timing may therefore differ between the two tasks. An adult laboratory mean from silent reading is not a normative target for a child reading a story aloud. Accurate interpretation requires a task-matched population and outcome.
34. Eye tracking and self-paced reading are different
A self-paced reading task commonly measures the time between revealing a word or phrase and the participant advancing with a keypress. Eye tracking measures actual fixation locations and durations while the person reads.
The self-paced interval includes the choice to press a key and may not correspond to one fixation. Calling it ‘first fixation duration’ would be incorrect unless eye movements were actually measured. Researchers should label instruments and units carefully.
35. The N400 is an electrophysiological measure
The N400 is an EEG-derived component associated with meaning processing in controlled experiments. It is not a fixation duration and does not share the same underlying unit or recording device.
A study might compare eye movements and ERPs to explore complementary stages of word comprehension. Agreement can be informative, but a 400-millisecond ERP name is not an instruction that FFD must last 400 milliseconds. Different measurements answer different questions.
36. The P600 is also a distinct signal
A P600 is a later positive-going ERP effect studied in relation to sentence difficulty or reanalysis under many conditions. It is not produced by adding several eye fixations together.
The distinction matters for educational explanations. A pupil rereading a confusing sentence does not allow the tutor to infer the presence of a P600. Neurological and behavioural claims need their own experimental evidence.
37. The 1998 Rayner review provides broad context
Keith Rayner’s 1998 review surveyed eye movements in reading and language processing across decades of studies. It describes how fixation durations, eye-movement control and lexical factors interact.
The review helps readers avoid treating FFD as an isolated magic measurement. Individual effects emerge within a larger system involving word recognition, visual input and sentence interpretation. A study’s specific design still governs the claim.
38. The E-Z Reader model predicts multiple outcomes
Reichle, Pollatsek, Fisher and Rayner’s 1998 E-Z Reader paper developed a computational account linking aspects of lexical access with eye-movement control. The model addressed gaze duration, fixation probabilities and later individual fixation measures.
Such models are not direct neural recordings. They provide testable predictions that can be compared with real data. A model’s success with FFD does not mean it has proven how every reader’s thoughts work. Scientific models must be examined through their assumptions and limitations.
39. The 2025 methods guide defines FFD directly
A 2025 Behavior Research Methods guide by Schotter and Dillon reviews eye-tracking design choices, preprocessing and common reading measures. It states the distinction between first fixation and other first-pass or rereading measures.
The authors emphasise that procedures should be chosen according to the research question, not copied mechanically across fields. That is the proper lesson for vocabulary research: define the observable event, choose a defensible analysis and report uncertainty.
40. The 2025 fundamentals paper adds a counting framework
Another Behavior Research Methods review describes first-pass outcomes such as skipping, single fixations and refixations alongside first fixation duration, gaze duration and total fixation time.
The measures are correlated because they share underlying eye events. A paper testing many of them must consider statistical multiplicity and not report only those with attractive p-values. A measurement framework prevents a single stopwatch reading from becoming an overclaimed cognitive conclusion.
41. Multiple comparisons can create false discoveries
When a study reports tests for FFD, single fixation duration, gaze duration, go-past and total time, the outcomes are not independent. Repeated significance testing can increase false-positive risk without an appropriate plan.
Von der Malsburg and Angele’s work examined this problem in common eye-movement analyses. Transparent studies state their planned hypotheses and how they handled the number of tests. Research literacy requires attention to statistics as well as the word definitions.
42. A mean hides the response distribution
Some words receive very short initial fixations and others longer ones. Averaging them produces one number that may conceal outliers, skew or differences in the types of trials included.
A researcher may inspect distributions and use appropriate statistical models rather than relying solely on a grand mean. A classroom should not translate a group average into a personal benchmark. Actual learner support needs broader evidence.
43. Participant and item variability both matter
Participants differ in reading skill, language experience and eye-movement patterns. Words differ in length, frequency, predictability and spelling properties. Both kinds of variability can affect observed FFD.
Modern analyses often account for participant and item differences where appropriate. A claimed group effect should be evaluated against its uncertainty and sampling context. A single word read by one person offers far less evidence than a controlled design.
44. A careful apparatus check can prevent misleading data
Eye trackers require calibration, suitable lighting, stable stimulus displays and procedures for handling blinks or tracking loss. If the target’s position is measured inaccurately, the duration may be attributed to the wrong item.
An attractive chart cannot repair faulty data. Methods should document tracking quality and exclusions. For children, recordings also require appropriate consent and privacy safeguards. The right intervention should never be selected from an unvalidated measurement.
45. An original three-outcome table
Imagine three fictional target encounters. In the first, the word receives one 190-ms fixation. In the second, the word receives 160 ms followed by 140 ms within-word. In the third, the word is skipped entirely on the first pass.
The FFD values are 190 ms, 160 ms and unavailable for the skipped target. The corresponding first-pass gaze durations are 190 ms, 300 ms and unavailable. This constructed example shows why zero is not a valid substitute for an unobserved initial fixation.
46. A correction may be useful rather than slow
A pupil might notice a word resembling another familiar word and inspect it again. The first fixation can be short even when the full word takes additional time to identify.
That is not inherently a reading failure. The second look may prevent a wrong interpretation. A tutor should evaluate the final spelling recognition and meaning rather than demand that every word be identified within one fixation of a fixed duration.
47. Primary English: recognise and explain the word
A Primary pupil may look at garden briefly, pronounce it correctly and still need to learn a figurative or technical use of the word. A short first fixation would not settle that semantic gap.
Ask for the intended meaning in the sentence and a fresh example. If pronunciation or spelling is incorrect, teach the relevant pattern. An eye-movement statistic does not replace an age-appropriate literacy lesson.
48. Primary English: do not rush negation
Words such as not, only and unless are short but can determine whether an instruction permits or prohibits an action. An easy-looking token may carry the sentence’s decisive logical force.
Ask the child to paraphrase the complete condition and apply it to a new scenario. The instruction is understood when the answer is accurate, not when the first fixation is as short as possible. Vocabulary technology protects meaning before speed.
49. Secondary English: one sense is not the whole word
A Secondary student may recognise the written form charge but need to decide whether the paragraph concerns money, electricity, an accusation or movement. That decision depends on grammar and context.
The teacher should ask for the relevant sense and supporting words. FFD can be a research indicator of visual processing, but it does not prove sense selection. Reading comprehension requires a complete interpretation of the sentence.
50. Secondary English: ambiguity may be resolved later
A long sentence can encourage one initial structural interpretation and then require revision after a later clause. The eye may already have left the first word before the reanalysis becomes necessary.
The tutor can underline the main clause and identify the critical disambiguating word. The first fixation on an earlier noun cannot describe all later syntactic work. The educational goal is a faithful paraphrase that survives the revision.
51. Mathematics: the whole condition governs the answer
A word problem may use at least five or more than five. The individual words are familiar, but their relation to the numerical endpoint determines the inequality.
Students should write the correct symbols and test whether the boundary value is included. A fast fixation does not demonstrate correct mathematical reasoning. Vocabulary is useful when it supports a valid operation.
52. Science: a long technical word needs a real definition
A learner might require another look at photosynthesis or another specialist term. That visual demand is separate from knowing what process the term describes.
Teach the scientific mechanism, conditions and an original example. An FFD result from adult reading research cannot be used as a threshold for the pupil’s conceptual mastery. The answer must be checked through explanation.
53. History: verify names and source wording
An archival document may contain similar names or dates whose exact characters matter. A reader who recognises an outline quickly can still attach the text to the wrong person.
Teach careful source comparison and faithful transcription. Extra inspection may be appropriate. The first fixation duration cannot certify historical attribution or evidence quality. Accurate labels need independent verification.
54. General Paper: fluency is not truth
A familiar slogan may receive an easy first reading while expressing an unsupported claim. An unfamiliar technical phrase may require more visual effort yet introduce evidence that changes the argument.
Ask students to identify the claim, evidence and hidden assumptions. The credibility of an essay does not follow from reading speed. Vocabulary supports reasoned analysis when its meanings and sources remain explicit.
55. Translation: written similarity can deceive
A source-language word may resemble a target-language form while meaning something different. A quick initial recognition is not evidence that a translation is faithful.
The translator must verify the intended concept, grammatical roles and register. Eye tracking can inform research about translation processes but cannot directly certify equivalence. Exact source content remains the authority.
56. Professional instructions: a second look can prevent error
Workplace documents contain identifiers, deadlines and exceptions. A single word or digit may control permission, money or safety. Reading it again can be the responsible action.
Use the current authoritative procedure and check the complete string. No eye-movement result justifies replacing verification with a guessed familiar term. The practical purpose of vocabulary is reliable action.
57. A reproducible five-column data sheet
For a valid eye-tracking study, record target word ID, fixation coordinates, first fixation duration, later within-word refixations and whether the target was skipped. Add region boundaries, participant and condition identifiers.
This allows another analyst to reconstruct FFD and distinguish it from gaze duration. A single value without an event sequence or inclusion rule can be misinterpreted. Transparent data definitions are essential to good research.
58. The final lesson: first is not everything
First fixation duration is the length of the first measured stop on a target word, not the total time required to understand the text. It can contribute to hypotheses about early lexical processing when the study has suitable controls.
For real education, the learner must continue beyond the first look: recognise the full form, identify the appropriate meaning and use it accurately. Vocabulary becomes a technology when efficient visual processing is paired with sound judgement and independent communication.
Full case: a fictional two-fixation route
A hypothetical reader encounters garden, landing on it for 180 ms before making another within-word fixation of 140 ms. The first fixation duration is 180 ms and gaze duration is 320 ms. The pupil can calculate the values from the event sequence without implying that either duration measures the word’s full meaning.
The teacher then presents ‘The children gathered in the garden’ and asks which sense applies and why. That independent language task supplies comprehension evidence. The eye-movement example is for measurement literacy; the sentence explanation is for vocabulary mastery.
Seven-stage eye-tracking interpretation checklist
- Identify the exact target word and the boundary of its screen region.
- Locate the first eligible fixation and record its measured duration.
- Keep within-word refixations separate when calculating FFD.
- Calculate gaze duration from all valid first-pass target fixations.
- Treat skipped words and tracking failures under a documented inclusion rule.
- Control or report word frequency, length, preview, landing location and task.
- Test actual word meaning through a separate explanation or independent use task.
Frequently asked questions
1. What is first fixation duration?
The elapsed time of the first valid fixation on a target word during the designated reading pass.
2. How is it different from gaze duration?
Gaze duration adds additional first-pass within-word fixations.
3. Is first fixation the same as single fixation?
Not always. FFD includes targets later refixated; single fixation duration uses one-fixation trials.
4. Does FFD include later rereading?
No. Later visits are measured through other reading-time outcomes.
5. What if the word was skipped?
There is no observed eligible first-pass target fixation duration.
6. Do short fixations always show easy comprehension?
No. Visual, motor and task effects can also produce short intervals.
7. Can a first fixation be longer for a familiar word?
Yes. Individual trials and conditions vary, even when group tendencies exist.
8. Does FFD record an N400 or P600?
No. Those are EEG-derived ERP responses, not fixation time.
9. Can FFD diagnose a child’s reading disorder?
Not from an informal activity; professional assessment requires broader validated evidence.
10. What is the useful educational goal?
Recognise the word accurately, explain its context and use it independently.
Primary research and connected vocabulary guides
- Schotter and Dillon (2025): A Beginner’s Guide to Eye Tracking for Psycholinguistic Studies of Reading — careful event definitions and analysis choices.
- The Fundamentals of Eye Tracking, Part 2: From Research Question to Operationalization — first-fixation, gaze and skipping measures.
- Rayner (1998): Eye Movements in Reading and Information Processing — Twenty Years of Research — major overview of reading and eye movement research.
- Reichle, Pollatsek, Fisher and Rayner (1998): Toward a Model of Eye Movement Control in Reading — E-Z Reader.
- Slattery and Parker (2019): Return Sweeps in Reading — Processing Implications of Undersweep Fixations — why line transitions complicate fixation definitions.
- Von der Malsburg and Angele (2017): False Positives and Other Statistical Errors in Standard Eye Movement Analyses — multiple-outcome reporting risk.
- Vocabulary | What Is Gaze Duration? — sum of multiple initial target fixations.
- Vocabulary | What Is Refixation Probability? — when another stop occurs.
- Vocabulary | What Is Regression Path Duration? — backward detours and later reading.
- Vocabulary | What Is Word Skipping? — targets without direct initial fixation.
- Vocabulary Learning Hub — integrate word form, grammar and meaning.
First fixation duration is a clear and valuable observation, but it is only the beginning of a reader’s journey through a word. Accurate vocabulary learning asks more than how quickly the eyes first stopped: it asks whether the person recognised the correct form, resolved its intended sense and communicated that meaning faithfully.