Flashcards in Reading and Eye Movements Deck (51)
Why do we need reading?
-Individual words – need to be able to identify
-Sentences->paragraphs->books – putting the words together
-Existing knowledge e.g. the alphabet
Why eye movements?
-Need to navigate the text
-A lot of research in the area
-Research has been going on for about 120 years
-Huey, 1898 – tried to look at someone’s eyes whilst they were reading, struggled to find participants
What is the eye-mind assumption?
-“the eye remains fixated on a word as long as the word is being processed. So the time it takes to process a newly fixated word is directly indicated by the gaze duration” Just & Carpenter 1980 p.330
-Non-invasive, relatively low-cost, sensitive online measure of cognitive processing during reading, use of a camera takes sample every ms
-Inform reading instruction, identify reading difficulty, inform design
What has eye tracking found when looking a children being read to?
-Using eye tracking
-Only interested in image when they don’t know how to read
-Look at the picture and some of the text when being read to if they know how to read
How do you perceive a word?
-Need highly detailed (central) vision for accurate perception of word form
- I.e. need to make a fixation on(almost) every word
-Make short saccades between the words
What are the Oculomotor patterns?
-Most words receive at least one direct fixation
-Saccades ~15-40ms, ~5-9 chars
-= average reading speed ~250-350wpm
What is a Saccade?
one it is set off can’t stop it, directive one shot movement to change gaze position
What is refixation?
What is fixation?
What is skip?
What is regression?
What is return sweep?
What are the factors influencing EM behaviour?
1. Characteristics of the visual system
3. Online cognitive processing of text
4. Personal characteristics
5. Task differences
6. Text differences
What does the Retina contain?
-Work at lower light intensities
What does the Fovea contain?
-Trichromatic (colour vision)
When is something considered high detail?
-Directly looking at something = high detail
-If you’re not looking straight ahead
What is text like with Online processing?
-Clear effects of certain text characteristics on metrics such as fixation duration
-Looking at effect of different manipulations on different measures can help us understand how text is processed during reading
-Manipulate a target word itself or sentence context and compare oculomotor behaviour on target word
-E.g. word length, word frequency, sentence predictability, sentence plausibility, syntactic ambiguity…
Can Word length have an effect?
-Short words are processed quicker
-Same pattern X-strings
-Fixation on longer words due to the amount of visual information
What is early processing?
‘Early’ processing –single fixation duration, first fixation duration, skipping probability –the initial familiarity/identification stages
What is late processing?
-‘Late’ processing –total gaze duration, regression probability, go-past duration –the stuff that comes after, integration into wider representation of the text
Where is your attention when reading?
-Mostly deployed to upcoming text (ahead of point of fixation) to begin processing
-Small allocation behind fixation
What is Word identification span and how do you investigate it?
-Word identification span –slightly asymmetric (~4L-7R)–close to average progressive saccade size (7 characters)
-Investigate with boundary change paradigm –gaze contingent boundary, different preview to fixated word –see if you notice the difference
What is the boundary change paradigm?
-Use a eye tracker which uses an invisible boundary, and if the boundary is crossed the sentence will change, participant will be asked if they noticed the change
-Underwood and McConkie 1985
What is perceptual span?
-Asymmetric –spans around 4 characters to left and 15 characters to right of fixation
-Constrained under higher foveal load – if you’re looking at something directly which is difficult to read, less attention around the fixation
-Information about word form, inter-word spacing
-Gaze-contingent moving window paradigm (McConkie& Rayner 1975) – using a window where the eye tracker can surround words with x’s
How do we know perception is to do with attention and not the visual span?
-Reversed for orthographies that are read the other way (e.g. Hebrew)
-Shrunk for more dense (/complicated orthographies; e.g. Chinese, 1L 3R)
-Not affected by parafoveal magnification – argued that if the surrounding text is bigger, acuity should be improved
What are the Two main competing models?
1. E-Z Reader –serial
2. SWIFT –parallel
What is the E-Z Reader Model?
-Attention deployed to one word at a time
-Two stages of lexical processing
-Saccade planned after first stage is complete
-If 2nd stage is completed as well before this is executed, attention shifts to the next word – saccades are planned for following words in this way until both stages of lexical processing of word (n+x) can’t be completed before the saccade programme is implemented (intervening words are ‘skipped’)
-If 2nd stage isn’t completed before the saccade is executed, a regression may be programmed to return and complete processing
What is the SWIFT model?
-Words around point of fixation activated to different extents, dependent on early low-level processing which occurs in parallel
-Saccadic shifts are programmed at random intervals to the word receiving highest activation
-Shift might be delayed up to a point by higher processing difficulty on foveated word n than planned saccade target –if delay isn’t sufficient to complete processing before a saccade, activation for this word will be increased and therefore increase likelihood of later regressive saccade
Can word frequency effect how you perceive words?
-Less frequently encountered words take longer to process
-Measure of successful word identification
-E.g. town = high frequency, cove = low frequency