Upgrade to remove ads
Only $35.99/year

Psych 309 Braithwaite. BYU. Exam 2

Terms in this set (94)

Validity
Evidence for inferences that can be made about a test score
3 Types of Validity
Content, Construct, Criterion
Prerequisites for validity
Reliability is necessary but not sufficient. Validation studies justify validity
Face Validity
Whether or not the item appears to measure validity. It doesn't actually offer evidence for real validity
Content Validity
The degree to which a test or measure adequately represents the conceptual domain (i.e. getting all the nooks and crannies of the construct)
Construct Underrepresentation
Failing to capture important components of a construct. I.e. a math test with algebra but not geometry
Construct Irrelevant Variance
Scores are affected by something irrelevant to the construct such as survey fatigue
Content Validity Ratio
A panel of experts decides whether your items are measuring what you say they are. CVR = (# correct - total #/2)/(total #/2)
Criterion
Another measurement or result the test is supposed to predict. i.e. ACT --> success in college
Criterion-related Validity
How well the tests corresponds to a criterion (measured by pearson's r)
3 subtypes of Criterion Validity
Predictive, Concurrent, Postdictive
Validity Coefficient
Pearson's R -- comparing to a criterion. Square it => percent of variation explained
Incremental Validity
How much better your test measures the construct than existing tests.
Construct
A hypothetical trait you are trying to measure that is not necessarily directly measurable i.e. Courage
Construct-related Validity
The degree to which a set of measurement operations measures hypothesized constructs
2 types of evidence for construct validity
Convergent and
Convergent
High Correlation with existing measures
...
Demonstrates specific relationships that were expected
Reliability and Validity
A test cannot be valid without being reliable.
"Mother of all validities"
Construct Validity
Relationship with examiner
It can influence the score on the test for the examinee. A positive relationship tends to improve scores
Examiner Race Effects on IQ test
The test is fairly structured and examiners are usually professionals, so there are less expectancy effects
Bilingual test takers
They should take it in the language of the examiner. Translation can lead to problems
Expectancy Effects
Expecting a certain result can lead it to happen through psychological processes. Rosenthal.
Situationally change testing procedures
When dealing with a different population like children or those with special needs
Computer Administrated Tests
More efficient, less bias, standardization. Less depth/quality of information and hard to detect errors
Subject Variables in Administration
Medication, Illness, Age, Test anxiety
Reactivity
Knowing that someone will check a test rater's ratings makes them more accurate
Drift
Examiners start to rate behavior in the way they want rather than as instructed. Overcome through frequent training
Contrast Effect
Rate behaviours differently when observations are repeated. They get bored so see things that aren't there
Experimenter Expectancies
? Experimentors expect their hypothesis to be true, which can bias the results
Detecting Deception
Almost no better than what would be expected by chance.
Halo Effect
Knowing one good thing about someone leads you to infer other good things. i.e. attractive to kind
Interview vs Test
More detailed and rich but subject to bais
A good interviewer
Knows how to be warm and bright so the interviewee is relaxed and feels safe
Interpersonal Influence
The ability to get others to do what you ask
Interpersonal Attraction
A feeling of understanding and mutual respect
How attraction affects influence
People who are more likeable are more likely to have influence. If someone doesn't like you, you'll have a hard to influencing them.
Transitional Phrase
Used to keep the interaction flowing
Evaluative & Probing Statements
Avoid! Probe for more info, lead to hostile responses & false reassurances --> lose trust. People CANNOT feel judged
Direct Questions
Use only when there is no other way to obtain the information
Structured Clinical Interviews
Less likely to be biased and easier to administer and easier to compare. Less depth or quality of information
Mental Status Examination
Determine whether people are sane/functioning. Test orientation (what floor are we on). Name 3 objects and repeat
Standoutishness
If one trait stands out, we may anchor judgements on that trait
Interview Reliability in Structured Interviews
Twice as high as nonstructured. Different raters are more likely to get similar results.
Criticism of Structured Interviews
...
Social Facilitation
We tend to act like the models around
Error in Interviews
1st impressions, biases, blindspots, halo effect, physical appearance
Intelligence: 3 research traditions
Psychometric, information processing, cognitive approaches
Binet: 3 facilities
Judgement, attention, and reasoning
Age Differentiation
Differentiate older from young children by their performance on a task
General Mental Ability
Total product of various distinct elements of intelligence
Mental Age
Equivalent age capabilities regardless of actual age
Positive Manifold
When a set of diverse ability tests are administered to large unbiased samples, correlations tend to be positive
Binet task percentage
67 to 75 percent of age group
Spearman's g
G --> general intelligence. An underlying factor that explains all intelligence
Spearman's method
Factor analysis
GF-GC
Two basic types of intelligence -- Fluid (abilities that allow us to obtain new knowledge)
Crystalized
IQ Formula
Mental Age/Chronological Age
Deviation IQ
M=100 SD=16
Ceiling
# of incorrect items --> fail and move on
Basal
# of correct --> move forward
Weschler
A global capacity to act purposely??
Wechsler: binet criticism
Didn't like 1 score or nonintellective factors (i.e. speeded/timed portions).
Wechsler age range
2 to 90. Seperate tests for kids
Point Scale
Point scale allowed you do group scores and more easily perform factor analysis.
Items grouped together to get a better feel of ability
WAIS-IV subtests
Verbal Comprehension, Perceptual Reasoning, Working Memory, Processing Speed
Verbal Comprehension
Highest Correlation with total IQ. Not affected by age. (Vocab, similarities, information/trivia)
Perceptual Reasoning
Block design. Use 9 colored blocks to make specific design. Also matrix reasoning
Working Memory
Arithmetic and digit span. Affected by non-intellectual factors
Processing Speed
Digital symbol coding/symbol search
Means and SDs WAIS-IV
Scaled Scores: M=10 SD=3. Standard: M=100 SD=15
Calculate IQ scores
Add up 10 subscales, each with a score of 5-15
Index scores
The same as subtest
Pattern Anlalysis
Intelligence subscales and large differences hint at schizophrenia
Hold Subtest
A score that holds strong despite injury or mental illness
WAIS-IV: Crystallized Intelligence
Vocab, Information, Comprehension
WAIS-IV: Fluid Intelligence
Matrix Reasoning, Block Design
Traditional intelligence test
Measurement breaks down at extremes
Alternative Intelligence Tests Dis
Weaker psychometrics, standardization, limits the range of abilities
Alternative Intelligence Tests Adv
Used for specific populations or purposes. Not as reliant on crystallized intelligence. Nonverbal testing
Infant development tests
Do NOT predict future IQ
Screening
Administering a standardized tool to identify at risk children
Surveillance
Recognizing children with developmental delays. Document and maintain history, make accurate observations, identify risk and protective factors
Brazelton Neonatal Assessment
Disadvantages: No norms, poor reliability, doesn't predict IQ
Bayley Scales
Psychometrics break down at low ages. Doesn't predict IQ. BUT very low scores predict mental retardation
Sensitivity
Accuracy in identifying problems
Specificity
Avoiding false diagnoses
False positive
Diagnosing someone without a disorder
True negatives
People who are truly not delayed
Levels of specificity and sensitivity
70 to 80 percent
Learning disability identification
IQ and Woodcock Johnson. More than 2 SD difference leads to accomodation. Otherwise they are treated normally
Woodcock Johnson III
Assesses achievement to diagnose learning difficulties
Using test scores to diagnose LDs
This is a bad method because the discrepancy is often not demonstrated until 4th grade
Students also viewed