Ppsy 674 Cognitive Assessment

Front 1

What is Standard Error of Measurement?

Back 1

Error variance for rel. coef. Compare two tests for which is more reliable. Calc'd on full sample.

Front 2

What is the Standard Error of the Estimate?

Back 2

Error variance for the val. coef. To compare an indiv. perf on two tests. It is calc'd on ind performances. Measures the predictive ability of the regression eq. The higher the corell, the lower the SEE, the more accurate the pred.

Front 3

Cronbach's Alpha

Back 3

Reliability coefficient that measures internal consistency:
r = (n/n-1)*{Var(t) - Sum[Var(i)]}/Var(t)
Use for Likert Scales

Front 4

Kuder Richardson

Back 4

Reliability coefficient:
r = (n/n-1)*{Var(t) - Sum[pq]}/Var(t)
p=prob an item will be passed; q=probability it won't be passed.

Front 5

5 purposes of assessment

Back 5

1. Screening; pop. prevalence
2. Problem solving
3. Diagnosis
4. Counseling & rehab.
5. Progress or program eval.

Front 6

4 pillars of assessment

Back 6

1. Norm referenced tests
2. Interview(structured,semi,un)
3. Observations (structured, semi and un)
4. Informal Assessments
Use no less than 3 sources for DX

Front 7

Assessment Process

Back 7

1. Battery approach-everyone gets the same tests
2. Hypothesis testing - like a single subject exper design; tests are selected based on est cog functioning

Front 8

Social Validity

Back 8

How likely is it that a recommendation will be followed

Front 9

Spearman's g

Back 9

FSIQ is made up of VIQ and PIQ (perfmormance IQ)

g stands for general mental ability and is a weighted avg of more specific abilities.

Front 10

Concurrent validity

Back 10

Degree to which test scores relate to some currently available criterion.

Front 11

Criterion-related validity

Back 11

Degree to which test scores relate to a type of criterion.

Front 12

Content validity

Back 12

Degree to which the test content represents the domain to be measured.

Front 13

Construct validity

Back 13

Degree tow which a test measures a psychological construct or trait.

Front 14

Predictive validity

Back 14

Degree to which the test score relates to criterion performance (requires a time delay to determine).

Front 15

Factor analysis

Back 15

Extracts first factor that accounts for the largest proportion of variance, then the next and so on. In intelligence tests the first factor is g.

Front 16

Age-equivalent and grade-equivalent scores

Back 16

Average test scores by age (aka Mental Age) and grade are used to report test scores. MA/CA X 100 = IQ

Front 17

Clinical significance

Back 17

Once troubled and disordered clients are now, after treatment, not distinguishable from a meaningful and representative nondisturbed reference group. The effect of sample size is neutralized.

Front 18

Practical significance

Back 18

Practical significance looks at whether the difference is large enough to be of value in a practical sense. Effect size puts the change on the normal curve.

Front 19

Statistical significance

Back 19

Statistical significance means that the observed mean differences are not likely due to sampling error. Tests hypothesis.

Front 20

Theories of Intelligence

Back 20

1. Factor Analytic
2. Information Processing

Front 21

Factor Analytic Theories of Intelligence

Back 21

g is one FA theory and may include s or specific factors, making it a two factor theory.
Another is multifactor theory that equates the factors. An example is Horn's Crystallized vs fluid intelligence.
Crystallized intelligence depends on culture while Fluid intelligence is culture free mental efficiency.

Front 22

Information Processing Theories of Intelligence

Back 22

Structural component relates to the physical brain structure while the functional (executive) component relates to the learned component.

Front 23

Norm refenced vs. Criterion referenced tests

Back 23

Normed tests are comparisons to othter subjects on the same measure.
Criterion referenced tests compare to an established standard and measure mastery.

Front 24

WAIS III

Back 24

Ages 16-89

Front 25

Executive Functions

Back 25

Class of skills that enable independent, productive and goal directed behaviors.
Diff from cog functions which assess what we are capable of and how much of that capability we have. EF assess how we do something.

Front 26

Ceiling Level

Back 26

Level at which the test ends for an age group (usually 2 SD's). Most subjects get the answers wrong.

Front 27

Floor Level

Back 27

Level at which the test starts for an age group and rep the level where most subjects (95-98%)get the answer right.

Front 28

Scales of Measurement

Back 28

Nominal - values are names. does not rank order.
Ordinal - Rank order w/o regard to interval diff. Interval - Rank order with regard to interval diff. Eg, WAIS. Arbitrary zero and equal units.
Ratio - True zero with equal units of equal ratios. No psych test does this. eg. weight

Front 29

Measures of dispersion

Back 29

Range, Variance and Standard Deviation.
Corellation ind. direction and magnitude but not causality. Rsqrd is the variance (?).

Front 30

Regression to the mean

Back 30

Scores at the extremes regress to the mean on retests.

Front 31

Age equivalent score
Grade equivalent score

Back 31

Mean raw score for an age level.
Mean raw score for a grade
Both are useful in academic & achievement tests. Age is prfrd unless the child is held back.

Front 32

Standard score

Back 32

arbitray mean (10 or 100)with standarized dev (3 or 10)

Front 33

Error Variance

Back 33

Rel coeff ind stability of test and the amt of error variance. Represents any condition that is irrelavent to the purpose of the test.

Front 34

Reducing Error Variance

Back 34

Standardized adm instructions
Time limits
Standardized scoring directions

Front 35

Test-retest Reliability

Back 35

As T rises, Corr coeff decreases. Problems include practice and memory effects. Time interval shb reported. 2 test admins.

Front 36

Alternate Forms Reliability

Back 36

Fewer practice and memory effects than T-R, but difficulty level may not be ident. Time interval shb reported. Alt forms may be "immediate" (1 admin) or "delayed" (2 admins).

Front 37

Split-half Reliability

Back 37

Single admin. Reducing # items reduces reliability of S-H rel below T-R rel. Content consistency. 1 test admin.

Front 38

Kuder Richardson (KR20)

Back 38

Evaluates intercorellations of all items in a dichotomous form. 1 Test admin.

Front 39

Coefficient Alpha

Back 39

Evals intercorellations for Likert scales. 1 test admin.

Front 40

Interrater Reliability

Back 40

Error variance is due to examiners.

Front 41

Sources of Error

Back 41

T-R time sample
A-F Immediate Content sampling
A-F Delayed Time & Content
S-H Content sampling
KR20 - Content sampling
Alpha - Content and homogeneity sampling
Interrater - Inter-scorer diffs

Front 42

Confidence Intervals

Back 42

Uses SEM. Range of scores in which an indiv's true score will likely fall. The higher the SEM, the greater this range.

Front 43

Validity

Back 43

Measures corell b/w test and criterion performance. On a continuum and should not be typed.

Front 44

Factors affecting validity (and rel coef)

Back 44

Nature of the sample
Sample heterogeneity - restricting range reduces val coef.
Form of the relationship - assumes homoscedasicity. Heteroscadiscity red. corell.

Front 45

Content Validity

Back 45

Sys exam. of test content to see if rep sample of behavior. Procedures should be discussed.

Front 46

Face Validity

Back 46

Not a valid measure. Subjects more likely to take a face valid (relavent) test.

Front 47

Criterion-related validity

Back 47

Effectiveness of a test in predicting an indiv. perf in a specified activity (IQ, pilot tests.
A. Concurrent - both tests given at the same time.
B. Predictive - a time lapse exists.

Front 48

Construct-related validity

Back 48

Extent to which a test measures a theoretical construct or trait.