Human Cognitive Favorite Writers Response Question -Read Entirely FIRST!!
xzevo103
Chapter10.pptLANGUAGE II: READING AND COMPREHENDING TEXT
*
Top-Down and Bottom-Up Processes in Reading
Reading is an exercise in pattern recognition
Involves both bottom-up and top-down processing
Bottom-Up: data (text) drives processing
Top-Down: expectations and knowledge facilitates processing
Mechanical Aspects of Reading
*
Eye Movements
Methodology
Eye movements can be tracked with an infrared beam
A non-invasive technique
*
Saccades and Fixations
Saccades: discrete movements made by eyes from one point to another
Saccadic suppression
During saccades, we take in little or no visual information
Fixations
Brief pauses to allow information intake
Last 200 to 300 ms
Gazes = consecutive fixations
Gaze duration = summed duration of gazes
*
Where vs. When
Two major dimensions of interest in eye movements
When they occur (fixation duration)
Where they go (i.e., “landing spots”)
On average, 65% of words are fixated
80% of content words
40% of function words
You’ll note a number of interesting characteristics of eye fixations and eye movements. First, some words are fixated twice, and some not at all. Notice how content words, like devices and combustion receive more gaze time than do function words like and and that; sometimes, these function words are not fixated at all.
*
Fixation time determined by how challenging word is to process
Top-down factors—based on knowledge, experience
Word Frequency
Low frequency words receive longer fixations than high frequency words
Word Predictability
Unpredictable words (based on context) receive longer fixations than predictable words
Age at which word was acquired
The younger you acquired the word the less time it is fixated
*
Factors affecting “landing spot”
Bottom-up factors—based on physical characteristics
Length of saccade is determined by:
length of the current word being fixated
length of the word to the immediate right
Fixations “land” about ¼ of the way into the word
*
Eye Movements: Word Skipping
Word Skipping
30% of words are skipped (Rayner,1998)
Top-down and bottom-up factors affect the probability of word skipping
Words that tend to be skipped are:
highly constrained, high frequency (top-down)
short in length (bottom-up)
*
Eye Movements: Word Skipping
Regressive Saccades
10-15% of saccades are backwards
Occur when a word is difficult
Good and poor readers differ in the quality of regressive saccades
Good readers are good are regressing back to exactly where they encountered the problem
Poor readers must do more backtracking to zero in on where they had a problem
Regressive saccades are not the cause of poor reading
Correlation ≠ causation
*
Eye Movements: Regressive Saccades
Perceptual Span
The amount of text around a fixation point that is effectively covered by eyes
For English: 3 to the left, 15 to the right
Perceptual span differs based on:
Orthography
Hebrew is mirror image of English (read left to right)
Difficulty of material
Perceptual span is smaller for difficult material
*
Parafoveal information
Information that falls slightly out of fixation, but still within perceptual span
Aids in:
lexical access
detection of the length of upcoming words
detection of end of the word that constrains the possible identity of the word
*
Word Recognition
Theories of word recognition
Direct-access
Orthography (word’s written letter pattern) allows for direct access to the semantic memory representation
Indirect-access
Access to word’s semantic memory goes through its phonological representation first (access is indirect)
We access a word’s semantic memory representation by the way a word sounds
*
To contrast direct and indirect access, researchers employ homophones
Words with different spellings and meanings, but identical pronunciations (e.g., prince and prints)
*
van Orden (1987)
Task: category verification task
A category (flower) is followed by one of three types of targets:
Category member (e.g., rose)
Homophone of category member (e.g., rows)
Word with orthographic similarity to category member (e.g., robs)
Dependent variable: Accuracy at indicating if target is a member of the category
*
Prediction
Direct access view
Homophones (rows) and orthographically similar words (robs) should be mistakenly classified
Indirect access view
More misclassifications for homophones than orthographically similar words
Phonological representation is accessed—two possible interpretations exist (rows and rose)
Results
Orthographically similar words:
Homophones:
20%
3%
*
Investigating Direct vs. Indirect Access
(Van Orden, 1987)
Luo, Johnson, & Gallo (1998)
Tested direct vs. indirect access by using pseudo-homophones
“chare” (non-word, but phonology identical to chair)
Semantic related task: judge if two words are semantically related
“Chare-table”
Dependent variables
RT to make the judgment
Accuracy of the judgment
*
Investigating Direct vs. Indirect Access
(Van Orden, 1987)
Predictions
Direct access view
Pseudo-homophones should not affect judgments of semantic relatedness
No lexical entry
Indirect access view
Pseudo-homophones will cause problems
Sound pattern will activate lexical entry, leading to effects of relatedness
*
Other Research: Direct vs. Indirect Access
(Ashby, 2010)
Results
Data support the indirect access view
Pseudohomophones interfered with relatedness judgments
Activated phonological representation in spite of non-word status
*
Word Recognition:
direct-access vs. indirect-access
Many researchers favor a dual-route view
Recognition can occur via either direct or indirect routes
*
Ashby (2010)
Recorded ERPs during silent reading of target words
Targets were preceded by a syllable prime that was either congruent or incongruent with its pronunciation
Congruent: COM as a prime for COMMON
Incongruent: COM as a prime for COMMIT
Predictions
Direct access view
Congruent and incongruent should produce similar ERP responses
Indirect access view
Congruent and incongruent should produce different ERP responses
Results
Smaller N100 to targets preceded by a congruent prime
Phonological processing occurs as early as 100msec after stimulus presentation
Dyslexia
Involves problems in printed word recognition
Normal performance on measures of intelligence and comprehension
Dyslexic characteristics
Accurate, but slow word recognition
Poor spelling
Central issue is lack of phonological awareness
Difficulty mapping sounds of words to orthography
Must rely on direct route to word recognition
Non-remitting (no “cure”)
Special accommodations required for standardized testing situations
Dyslexia and the Brain
Dylsexia and the Brain
Root of dyslexia—deficits in phonological processing transcends cultures and languages
Brain locus of dyslexia
Posterior region of left hemisphere—occipito-temporal region
Spanning areas that process visual and auditory information
Some dyslexics develop compensatory strategies in other brain areas
Anterior regions of the right hemisphere (inferior frontal gyrus) are important in sound articulation
*
How Should Reading Be Taught?
Whole-word approach
Top-down approach
Rote learning of words
Phonics approach
Bottom-up approach
Emphasize relationship between letters and sounds
Whole-language approach
Broad top-down emphasis
No explicit instruction on sound to letter mapping
Guess at unfamiliar words using storyline and illustrations
*
Animation reveals the best approach
Phonics approach associated with enhanced proficiency in
Reading
Spelling
Comprehension
Phonics approach is consistent with the notion of indirect access
Reading instruction should be grounded in phonics, integrated with other (top-down) approaches
*
Anglo-Centric Language Studies
Most reading research is based on English orthography, so may not generalize well to other languages
Idiosyncratic spelling-sound relationships in English
Examples
kn know
pn pneumonia
cht yacht
Other differences include use of letters vs. characters, spacing, size/length, complexity
Highlights the need for reading research with other (non-English) orthographies
*
Example: Chinese
Translation: Cognitive psychology is fun
Radically different from English in
number of characters
spacing of characters
correspondence of a given character to an idea/concept
*
Bai, Yan, Zang, Liversedge, and Rayner (2008)
- Interested in whether words or Chinese characters are a more fundamental unit in reading Chinese
- Participants: Native speakers of Chinese
- Presented sentences in one of four spacing conditions
- normal spacing
- single character spacing
- word spacing
- non-word spacing
Sentence: The rapid development of science and technology brings great changes to society
*
- Task: read each sentence
- Dependent variable: reading speed (characters/minute)
- Predictions
- If fundamental unit is words:
- Grouping by words should facilitate reading
- If fundamental unit is characters
- Grouping by words should not facilitate (and may interfere with) reading
- Results
Characters/minute
- Words were read with equal proficiency in normal spacing and word spacing conditions
- Single character spacing decreased reading proficiency, relative to word spacing
- Implications
- Spacing not important in Chinese reading
- Fundamental unit in linguistic processing appears to be words
Reading Myths
Do We Rceoginze Wrods as a Wohle?
Apocryphal “study” at Cambridge University
Claim: any jumbled arrangement can be recognized if:
seen within a context
first/last letters present
*
Rayner, White, Johnson, and Liversedge (2006)
Participants read 80 sentences in one of four conditions
Control condition: sentences left intact
Internal letters of words transposed
Beginning letters transposed
Ending letters transposed
Dependent variables
Fixations on the words
Words read per minute
Results
Word jumbling comes with a cost
Normal text condition associated with best proficiency
Not all word jumblings are equal
Internal letter transpositions were the least costly
Initial letter transpositions were the most costly
Text Messages
Varieties of text messages
Orthographic abbreviations
pls = please
Phonetic respelling
c u later = see you later
Impact on comprehension? Perhaps minimal…
Important letters still present
Phonology somewhat intact
*
Perea, Acha, and Carreiras (2009)
Constructed sentences that contained either orthographic abbreviations or phonetic respellings
Measured sentence reading time
Results
Time for text-message style sentences was slower than normal text
Time saved in texting comes with a severe reading cost
*
If word recognition is mediated by phonology, why is text message comprehension so poor?
Word recognition ≠ comprehension
The role of pseudo-homophones in sentence and discourse comprehension has not been fully explored
Sentence Level Processing
Sentence Parsing
identification of a sentence’s component elements and their grammatical relation to one another
A seamless (i.e., unconscious) process
Influenced by both syntactic (sentence structure) and semantic (meaning-based) factors
*
The Importance of Syntax
Sequential arrangement of words is primary in sentence comprehension
Commonly employed word orders (e.g., subject-verb-object) provide expectations that facilitate parsing
*
Garden-Path (GP) Approach to Parsing
Word-by-word “path” is followed attempting to fit each successive word within the developing syntactic structure
We assume “professor” is subject, “argued” is the verb, and “position” is object
Readers assume the simplest structure, revising only if proved to be wrong
*
According to GP approach, two heuristics used for parsing
Minimal attachment (MA)
We assume the syntax to be as simple as possible
Readers are parsimonious in beginning new phrases
Late closure (LC)
Reader attempts to include a new word in the grammatical structure currently being processed
Garden path sentences
The professor argued the student’s position was indefensible
The implicitly assumed S-V-O structure is violated
MA and LC heuristics do not work, leading to problems in comprehension
Garden-path sentences are associated with
Longer reading times
Longer fixations
More regressive saccades
The Importance of Semantics
Case grammar approach
Sentence parsing is based on semantics, rather than syntax
Sentence comprehension involves the assignment of words to case roles
Agent
Goal
Patient
planted in the
Carrie
flowers
garden
*
Which Takes Precedence: Syntax or Semantics?
Autonomous view
Syntactic and semantic analysis proceed independently of one another
Serial approach
Syntactic analysis precedes semantics
Interactionist view
Syntactic and semantic analyses impact one another
Parallel approach
Syntactic and semantic analysis occur simultaneously
*
Does semantic processing affect syntactic analysis?
Garden-path sentence difficulty could be due to difficulties in processing semantics, rather than syntax
Is the garden-path effect due in part to meaning?
Pickering and Traxler (1998)
If semantics influence syntactic analysis:
Garden-path difficulty should be more pronounced when the initial interpretation is semantically plausible…
. . . than when it’s implausible
Conditions: read garden-path sentences that were:
Plausible
Implausible
Results
Stronger garden-path effects for plausible sentences
Semantic plausibility of the sentence induced more of a “commitment” from readers
Syntactic change is more disruptive in this case
Indicates that syntax does not necessarily dominate semantics in sentence processing
As in the processing ambiguous words, multiple interpretations are considered, with one biased by context
Consistent with the interactionist view
Levels of Representation Code: precise wording
Three levels for the mental representation of discourse
Surface code
The precise wording used
Text Base
Major facts and themes
Situation Model:
The “world” created by the discourse
Combination of information from the discourse and background knowledge of the reader
Involves mentally simulating what is happening in the text
Discourse Comprehension
*
Mental simulation of discourse action is basically the embodiment of language comprehension
Ditman, Brunye, Mahoney, and Taylor (2009)
Situation models should be memorable to the degree that they can be simulated from the reader’s perspective
A language parallel to the enactment effect
*
Procedure
Presented three-sentence discourse scenarios, all with the same structure
Descriptive statement
I am a 20-year-old college student
Occupational activity (with same pronoun)
I work in the college bookstore
Activity with temporal marker (with same pronoun and activity)
Right now, I’m preparing for book buy-back
*
Independent variables: personal pronoun used in sentences
“I” vs. “You” vs. “He”
24 sentences, 8 with each type of pronoun
Dependent variables
Recognition memory assessed after 10 minutes or 3 weeks
Measured by sensitivity (ability to discriminate new from old sentences)
Response time to make the recognition judgment
Prediction
Better memory when pronoun “you” is used
Reading “you” as pronoun will lead to embodiment (putting yourself in the situation)
*
Results
Sensitivity and response time were best for discourse that featured “you” as pronoun
A text-comprehension version of the enactment effect
Sensitivity
*
Structure and Coherence
Comprehension of discourse depends on more than the content of individual sentences
Discourse needs to have some type of structure, and links between ideas
Local structure
Connections (either explicit or implicit) between individual sentences
Global structure
General knowledge brought to what is being read
*
Difficult to comprehend due to lack of local structure
Difficult to comprehending due to lack of global structure
Both local and global structure relate to coherence
The degree to which discourse “hangs together”
Local coherence
Global coherence
Anaphoric Reference
When a current expression refers to something that was encountered earlier in a text
Parts of an anaphoric reference
Anaphor is the referring expression (“his favorite group”)
Antecedents are the events that correspond to the expression (Greg)
Allows for integration of discourse currently being read with previous information already read
*
Anaphoric reference places strain on immediate memory
When an anaphor is encountered, a person must think back to (remember) the antecedent
Increasing distance = more difficult comprehension
Immediate memory span is a good predictor of comprehension
*
When an anaphor is encountered, the antecedent is more easily retrieved if the antecedent
occurs frequently
occurred recently
received first mention
One of first concepts mentioned in a text
Do the latter two sound familiar?
Primacy effect
Recency effect
*
Given and New Information
Given-new “contract”
Implicit “agreement” that all communication includes:
Given information
Information assumed to be known
New information
Information that is assumed to be unknown
“That test next week is going to be really tough.”
Given: There’s a test next week
New: The test will be tough
*
Components in following the given-new contract
Reader must determine given and new information
Reader must figure out what the “given” information refers to in the text
“New” information must be linked to that part of the text
Inferences
Inferences are conclusions drawn by the reader
Not explicitly stated in discourse
“Greg left the game disgusted by his team’s inability to score”
Inference: The team lost
Inferences involve the combination of world knowledge with text information
Inferences become part of the text’s mental representation
*
Kintsch (1974)
Presented participants with sentences that:
Implied a fact (e.g., a cigarette started a fire)
A burning cigarette was carelessly discarded.
The fire destroyed many acres of virgin forest.
Explicitly stated the fact (control condition)
A carelessly discarded burning cigarette started a fire.
The fire destroyed many acres of virgin forest.
*
Task: Sentence verification task
Verify if a later target sentence was true based on the initially presented sentences
For the implicit condition, the target sentence (e.g., the fire destroyed many acres of virgin forest) was:
Explicitly stated
Implied
Dependent variable: RT to verify the sentence
Conditions: given immediately or after a 15 minute delay
*
Results
After a short delay, explicit statements more quickly verified than implied statements
After longer delay, no difference in verification time
Implication: Inferred “facts” became part of the text representation
*
Types of Inferences
Types of Inferences
The movie was incredibly tense and exciting.
Bridget was on the edge of her seat.
Inference: Bridget was watching the movie
Bridging inference
Constructed to connect two ideas that are not explicitly connected
Also termed backward inferences
Allows reader to “bridge” current information back to earlier information
*
Serena held the ticket and watched the cop drive away.
“I’ll never text and drive again,” she thought.
Inference: Serena received a ticket for texting while driving
Causal inference
Reader infers the cause of an event they have just read
Type of bridging inference
*
Jason felt satisfied as he finished wrapping the tiny box.
“She’s going to love it,” he assured himself.
Inference: Jason is giving his girlfriend jewelry
Elaborative inference
Adding extra information to representation of text
Forward inference
Read moves “forward” from (i.e., beyond) the text
Types of Inferences
Singer (1980)
Singer (1980)
Presented one of three types of passages
The pitcher threw the ball to first base. The runner was halfway to second.
Explicit presentation of pitcher throwing ball
The pitcher threw to first base. The ball sailed into right field.
Bridging inference of pitcher throwing ball
The pitcher threw to first base. The runner was halfway to second.
Elaborative inference of pitcher throwing ball
*
Types of Inferences
Singer (1980)
After sentence, participants had to verify if a related passage (The pitcher threw a ball) was true as quickly as possible
Dependent variable: RT
*
Results
RT did not differ for bridging inferences and explicitly stated information
Inference made at encoding because it’s necessary for coherence
Replication of Kintsch (1974)
Elaborative inference passage led to longer RT
Inference not made until retrieval
Sulin and Dooling (1974)
Classic demonstration of making an elaborative inference
Presented passages like:
For some participants “Gerald Martin” was replaced by “Adolph Hitler”
These participants made elaborative inferences (e.g., hatred of Jews)
When are Inferences Made?
Two views
Constructivist view
Forming inferences is an active process
We seek to form inferences
Top-down view
Minimalist view
Making inferences is automatic and memory-driven
Inferences engaged minimally, only when necessary
Bottom-up view
Both likely come into play, depending on circumstances
*
Discourse Memory and Representation
Memory for Discourse
Classic study–Bartlett (1932): “War of the Ghosts”
*
Gist vs. Verbatim Information
Gist information: basic ideas, main points
This is what we are most likely to recall
Bartlett’s participants tended to recall gist info
General happenings and ideas from War of the Ghosts
Verbatim information: exact wording
Quite unlikely to be recalled, especially after a delay
Bartlett’s participants recalled very little of the verbatim information from the story
Hard-to-understand segments were omitted or distorted to be consistent with knowledge
*
Sachs (1967)
Tested transition from verbatim to gist representation
Presented participants with this passage
There is an interesting story about the telescope. In Holland, a man named Lippershey was an eyeglass maker. One day his children were playing with some lenses. They discovered that things seemed very close if two lenses were about a foot apart. Lippershey began experiments and his “spyglass” attracted much attention. He sent a letter about it to Galileo, the great Italian scientist (T1). Galileo at once realized the importance of the discovery and set out to build an instrument of his own. He used an old organ pipe with one lens curved out and the other curved in. On the first clear night, he pointed the glass towards the sky. He was amazed to find the empty dark spaces filled with brightly gleaming stars! (T2). Night after night Galileo climbed to a high tower, sweeping the sky with his telescope. One night he saw Jupiter, and to his great surprise discovered with it three bright stars, two to the east and one to the west. On the next night, however, all were to the west. A few nights later there were four little stars. (T3)
*
Tested memory for the underlined sentence at 3 different delays (T1, T2, T3)
Recognition test had one of three types of sentences
Exact target sentence
He sent a letter about it to Galileo, the great Italian scientist
Transformed version, retained meaning
A letter was sent about it to Galileo, the great Italian scientist
Transformed version, changed meaning
Galileo, the great Italian scientist, sent him a letter about it
Dependent variable: determine whether the test sentence was identical to the target sentence in the passage
Results
At T1, recognition was highly accurate
At T2 and T3, sentence with different words but same meaning was distinguishable from original
Implication: relatively quick loss of verbatim info (exact wording), but maintenance of gist info (general idea)
T1 T2 T3
*
Some evidence suggests that verbatim information is not completely lost
Ferreira, Bock, Wilson, and Cohen (2008)
Investigated the phenomenon of syntactic persistence
Tendency to use syntactic structures that have recently been heard
Primed to use a particular syntactic structure after being exposed to it previously
An example of implicit memory
*
Participants: amnesics and non-amnesics
Procedure
Phase 1: presented a sentence
Phase 2: presented a picture, which they were to describe with a sentence
Implicit memory test—would they use syntactic structure from phase 1?
Phase 3: Explicit memory test for sentences from phase 1
*
Results
Syntactic persistence was evident for both amnesics and normals
Sentence structure (syntax) from phase1 was used to describe the picture in phase 2
Intact implicit memory ability
Amnesics’ recognition memory (phase 3) was poor, relative to normals
Impaired explicit memory ability
Models of Discourse Comprehension
Schemas and Scripts
Schema: general knowledge we possess about some person, place, or event
Guides coding and retrieval processes during comprehension
4 component processes in use of schemas
Important message components need to be selected and attended
Meaning abstracted and disposal of verbatim details
Appropriate schema must be activated
Integrate new information and previous schema information to form representation
*
Schemas and Scripts
Script: specific type of schema
Knowledge representing routine activities
Scripts are important in maintaining coherent representation of discourse
Irrelevant elements to the script are not remembered unless distinctive (distracting or salient)
Central elements to a script are comprehended faster than peripheral elements, even if not explicitly stated
Easier to find antecedent and make bridging inferences
Notion of schemas/scripts is broad and descriptive; vague
*
Story Grammars
A schema of the developments typically encountered as we read a story
Beginning, plot development, resolution
The representation of discourse consists of episodes that correspond to the important elements of the story
Each episode is processed as a separate “chunk”
Tend to be recalled in an all-or-none fashion
Length of an episode does not affect ability to remember another episode
Broad and vague descriptive account
*
The Construction-Integration Model
Proposed by Kintsch and van Dijk (1978)
Major ideas from the discourse are represented as propositions
True or false statements about relationships among events
*
The Construction-Integration Model
Kintsch & van Dijk (1978) - Kintsch (1988;1998)
Propositions expressed in terms of arguments and predicates
Arguments: focus of the proposition
Predicate: information given about the argument
Example: The instructor delivered the entertaining lecture to her enthusiastic students.
ARGUMENT: Instructor
PREDICATE: Lecture
PREDICATE: Delivered
ARGUMENT: Students
PREDICATE: Enthusiastic
*
The Construction-Integration Model
Kintsch & van Dijk (1978) - Kintsch (1988;1998)
As we read we construct a network of hierarchically-related propositions, which are integrated with world knowledge
Top: Important propositions representing major themes
Bottom: Tangential propositions representing relatively unimportant details
*
Immediate memory (IM) capacity limits ability to consider propositions as we are reading
We can only simultaneously consider 6-12 propositions
Those likely to be held in IM:
Important or central (high-level propositions)
Propositions that have occurred recently
Redundancy of text reduces demand on IM
Searching for antecedents and making bridging inferences is easier
*
Takes longer to read a text as the number of propositions increases
Levels effect
Information from higher levels in the hierarchy are more likely to be recalled than information from lower levels
Bottom-up approach, in contrast to the schema-based top-down approach
*
Structure Building
Proposed by Gernsbacher (1991; 1997)
Comprehension involves building structures (i.e., coherent mental representations) as we read text
Three processes involved in structure building
Laying a foundation
Mapping
Shifting
*
Structure Building
Gernsbacher (1991; 1997)
Laying a foundation
Occurs with the introduction of a new topic
Advantage of first mention
First-mentioned concept is highly accessible and tends to be well-remembered
Importance of primacy (Chapter 6)
Found if test is presented at a delay
Advantage of clause recency
Most recently presented clauses tend to be well-remembered
Importance of recency (Chapter 6)
Found if test is presented immediately
*
Mapping
Using world and linguistic knowledge to relate new information to information already encoded
Cues that provide coherence (e.g., anaphoric reference, inferences) are used
Shifting
A new structure begins when:
incoming material is unrelated to the current foundation
an overt cue such as “on the other hand,” or “later that same day,” is encountered
General aids to structure building
Suppression
A “fine-tuning” of comprehension in which the activation of irrelevant information is inhibited
Enhancement
Relevant information is activated, increasing its accessibility
Structure Building
Callendar & McDaniel (2007)
Individual differences in structure building
Callendar and McDaniel (2007)
Investigated the effectiveness of text adjuncts
Built-in aids to learning in a text (e.g., embedded review questions)
Does the effectiveness of these aids differ as a function of skill in structure building?
*
Participants
Administered a test to assess skill in structure building
Classified as high or low structure builders
Conditions:
Participants read a chapter from a social psychology text in one of two conditions
With embedded questions
With no questions; read the text twice
Received a multiple choice test with one of two types of questions
Target questions: focused on embedded question info
Related questions: focused on related concepts to that in the embedded questions
*
Predictions
Low (poor) structure builders
Embedded questions no questions
Poor structure builders find it hard to inhibit irrelevant information
Consequence: build too many structures hindering formation of a coherent representation
Embedded questions should provide an anchor, highlighting important information
High (proficient) structure builders
Embedded questions no questions
Embedded questions provide little benefit as able to successfully identify important information
>
=
*
Structure Building
Callendar & McDaniel (2007)
Results
High structure builders were superior overall
Embedded questions helped only low structure builders
Implication
Best type of reading/testing style for comprehension depends on the characteristics of the reader
*
Metacomprehension
A person’s understanding of the degree to which they understand the information presented in a text
People overestimate their comprehension
.27 correlation between what people think they know and what they actually do know (Maki, 1998)
Research has focused on
Methods for assessing metacomprehension
How metacomprehension can be improved
*
Problems in assessing metacomprehension
Much of the research used one overall global ratings of comprehension
Some information is comprehended well and some not at all—one rating is insensitive to this
Estimating one’s knowledge of a text is difficult if one did not understand the text
To improve metacomprehension and its assessment:
Engage in activities during reading that encourage the monitoring of one’s own learning
Re-read
Generate key terms
Summarize reading
Decrease grain size of estimation
Metacomprehension is better when assessing specific, rather than broad, areas of knowledge
Assessing understanding of garden path approach is better than understanding of Chapter 10
Metacomprehension accuracy improves if judgments of knowledge occur after one tests themselves
Another example of the testing effect (Chapter 6)
Speed Reading
Claim: Reading can be improved with some simple changes to reading mechanics
Increase information taken in during each eye fixation
Eliminate inner speech
Characteristics of speed reading
Minimal fixations
Length of fixation is the same as normal readers, but only fixate once per line
Read down left hand page and up right hand page
Claims primarily anecdotal
Empirical research fails to support the claims
*
Just, Carpenter, and Masson (1982)
One of the few valid empirical tests of speed reading
Compared three conditions
Speed readers
Normal readers
Normal readers instructed to skim
Dependent variables
Memory for gist
Memory for details of the passage
*
*
Results
Memory for gist
All conditions did well
Memory for details
Normal readers performed best
Results basically show a speed-accuracy tradeoff
Increased reading speed likely decreases accuracy (i.e., comprehension)
Reading is a comprehension task…not a perceptual processing task
