Skip to content

Latest commit

 

History

History
105 lines (90 loc) · 8.09 KB

codebook.md

File metadata and controls

105 lines (90 loc) · 8.09 KB

Codebook

The codebook specifies the data types and other information for each column in the data files.

Table of contents

Word features

The word-level linguistic and lexical features, combined with reading measures (ET) or self-paced reading measures (SPR), plus psycholinguistic predictor variables.

Common columns

These columns are present in both ET and SPR datasets.

Column name Value type Description
subj_id String Unique participant identifier. Prefix indicates data collection site (zh = Zürich, pt = Potsdam).
text_id Integer Identifier of the text stimulus (16 texts in total).
screen_id Integer Identifier of the screen (each text was split into 6 screens, including title screen).
word_id Integer Identifier of the word within the screen.
session_id String Identifier of the recording session.
w_in_sent_id Integer Position of the word in the sentence.
word_length Integer Number of characters in the current word.
word_length_word_n_minus_1/-2 Integer Word length of the previous (n-1) or pre-previous (n-2) word.
lex_freq Float Normalized lemma frequency for the current word (lemma-based). Values come from an external lexicon/website and are normalized by dividing by a constant of ~122.323 (as provided by the dlexdb).
lex_freq_word_n_minus_1/-2 Float Normalized lemma frequency for the previous (n-1) or pre-previous (n-2) word
lex_freq_unk Boolean Indicates whether the word frequency was unknown in the lexicon.
surprisal Float Surprisal of the current word (negative log-probability from a language model).
surprisal_word_n_minus_1/-2 Float Surprisal of the previous (n-1) or pre-previous (n-2) word.
n_rights, n_lefts Float Number of right or left dependents of the current word in the dependency parse.
n_lefts_word_n_minus_1/-2 Float Number of left dependents of the previous words.
dep_distance Float Dependency distance of the current word (head-dependent linear distance).
dep_distance_word_n_minus_1/-2 Float Dependency distance of the previous (n-1, n-2) words.
deps String Dependency relation label of the word (according to German dependency grammar, e.g., subject (sb), object (oa), root).
contains_punctuation Boolean Whether the token contains punctuation.
session_label String Label for the recording session (e.g., Eye-tracking session, Self-paced reading session).
modality String Modality of the data collection (Eye-tracking = ET, Self-paced reading = SPR).

ET-specific columns

These columns are present only in the Eye-tracking (ET) dataset. All durations are in milli-seconds.

Column name Value type Description
FFD Float First-fixation duration: duration of the first fixation on a word if this word is fixated in first-pass reading, otherwise 0.
SFD Float Single-fixation duration: duration of the only first-pass fixation on a word, 0 if the word was skipped or more than one fixation occurred in the first-pass (equals FFD in case of a single first-pass fixation).
FD Float First duration: duration of the first fixation on a word (identical to FFD if not skipped in the first-pass).
FPRT Float First-pass reading time: sum of the durations of all first-pass fixations on a word (0 if the word was skipped in the first-pass).
FRT Float First-reading time: sum of the duration of all fixations from first fixating the word until leaving the word for the first time.
TFT Float Total-fixation time: sum of all fixations on a word (FPRT+RRT).
RRT Float Re-reading time: sum of the durations of all fixations on a word that do not belong to the first-pass (TFT-FPRT).
RPD_inc Float Inclusive regression-path duration: Sum of all fixation durations starting from the first first-pass fixation on a word until fixation on a word to the right of this word (including all regressive fixations on previous words), 0 if the word was not fixated in the first-pass.
RPD_exc Float Exclusive regression-path duration: Sum of all fixation durations after initiating a first-pass regression from a word until fixating a word to the right of this word, without counting fixations on the word itself.
RBRT Float Right-bounded reading time: Sum of all fixation durations on a word until a word to the right of this word is fixated.
Fix Categorical Fixation: 1 if the word was fixated, otherwise 0.
FPF Categorical First-pass fixation: 1 if the word was fixated in the first-pass, otherwise 0.
RR Categorical Re-reading: 1 if the word was fixated after the first-pass reading, otherwise 0.
FPReg Categorical First-pass regression: 1 if a regression was initiated in the first-pass reading of the word, otherwise 0.
TRC_out Float Total count of outgoing regressions: total number of regressive saccades initiated from this word.
TRC_in Float Total count of incoming regressions: total number of regressive saccades landing on this word.
SL_in Float Incoming saccade length: length of the saccade that leads to first fixation on a word in number of words; positive if progressive, negative if regression.
SL_out Float Outgoing saccade length: length of the first saccade that leaves the word in number of words; positive if progressive, negative if regression; 0 if the word is never fixated.
TFC Float The total fixation count on the word.

SPR-specific columns

These columns are present only in the Self-paced reading (SPR) dataset.

Column name Value type Description
word_rt Float Word-level self-paced reading time in milliseconds.

Participants

Participant-level information including psychometric test scores and task performance.
Psychometric tests were administered before the reading experiments. In Zürich (DILiLab), each participant completed two eye-tracking and two self-paced reading sessions.

  • Participant IDs follow the scheme:
    • zh1, zh2, …: Zürich participants
    • pt1, pt2, …: Potsdam participants

Psychometric tests included:

  • SLRT-II: Lese- und Rechtschreibtest (word and pseudoword reading speed)
  • MWT-B: Mehrfachwahl-Wortschatz-Intelligenztest (vocabulary knowledge)
  • RIAS: Reynolds Intellectual Assessment Scales and Screening (verbal, non-verbal, and general intelligence indices)
  • FAIR-2: Frankfurter Aufmerksamkeits-Inventar (attention indices)
  • WMC tasks: working memory measures (MU, OS, SS, SSTM)
  • Stroop and Simon tasks: inhibitory control measures
Column name Value type Description
subj Integer Unique participant ID.
SLRTWord Integer Word reading score (SLRT-II).
SLRTPseudo Integer Pseudoword reading score (SLRT-II).
MWTPR Float Vocabulary knowledge score (MWT-B).
RIASVixPR Float Verbal intelligence index (RIAS).
RIASNixPR Float Non-verbal intelligence index (RIAS).
RIASGixPR Float General intelligence index (RIAS).
FAIRLPR, FAIRQPR, FAIRKPR Integer FAIR-2 attention test scores (various indices: L, Q, K).
MUmean, OSmean, SSmean, SSTMRelScore Float Working memory task scores (Memory updating, Operation span, Symmetry span, Short-term memory).
total_memory Float Composite working memory score.
StrAccuracyEffect Float Stroop task accuracy effect.
StrRTEffect Float Stroop task reaction time effect.
SimAccuracyEffect Float Simon task accuracy effect.
SimRTEffect Float Simon task reaction time effect.