Assessment Concepts Test 1

Question 1

What defines a good assessment?

Answer 1

Defines the problem

Drives Intervention

• Both dynamic & static
• Gather info.
• Both formal & informal

Question 2

Assessment Purposes?

Answer 2

1. Compare to peers
2. Qualify for services
3. Measure progress

Question 3

What is “Compare to peers”?

Answer 3

Determine their status of difficulty

Compare:

• MLU
• Brown’s grammatical morphemes (syntax)
• Type-token ratio (semantics)

Question 4

Dynamic Assessment

Answer 4

looking @ performance over time

Question 5

Static assessment

Answer 5

one time

Question 6

Calculating Age

Answer 6

Todays date- yr, month, day

12 09 10 Today date

01 04 11 Birthday

11 04 29 Age

yrs. months years

_LOOk AT NOTES _

Question 7

Informal assessment

Answer 7

• natural; no strict rules
• limited to seeing what client produces

Ex. language sample

Question 8

Formal assessment

Answer 8

• Standardized
• More rigid
• Do exactly as says

Question 9

Criterion-referenced

Answer 9

• a test-taker’s performance is compared to a pre-defined set of criteria
• have they shown a certain or set of skills

Question 10

Norm-referenced

Answer 10

comparing a person’s score to another person’s score to classify students across a continuum

• a normal distribution results from a norm-referenced assessment
• Ex. GRE

Question 11

Reliability?

Types?

Answer 11

Measurements can be replicated

1. Inter-rater: b/w ppl Ex. both clinicians should get same results
2. Intra-rater: within examiner @ different times or different days

Question 12

Validity?

Answer 12

test truly measures what is claims to measure

Question 13

Validity Types

Answer 13

1. Face
2. Concurrent
3. Construct
4. content
5. criterion-related
6. predictive

Question 14

Basal

Answer 14

Set starting point depending on the age group

• criteria for a standard mostly based on age to obtain a basal
• stastical way of seeing what one knows

Question 15

Ceiling

Answer 15

when to stop administering subtest/ whole test

• follow specific rule-can assume client will not get any of those above correct

Purpose

• reduce client fatigue
• test administration efficiency

Question 16

Raw scores

Answer 16

actual # when grading tests/ # correct

Question 17

Types of Scores

Standardized

Standard scores based on 100 or 10 as mean, also called scaled scores

Answer 17

For explaining in percentile rank:

• 70 standard score– “Out of 100 children, 97 children answered better on this test”
• 55 standard score– “out of 1000 children, 988 children performed better”

Question 18

Types of Scores

Standardized

Z-scores

Answer 18

how many standard deviations the raw score is away from the mean

Question 19

Types of Scores

Standardized

Stanine

Answer 19

score based on a 9-unit scale

• educational & testing reports
• 2-7 –> normal limits
• 4-5 –> most average limits

Question 20

Types of Scores

Standardized

Standard deviations

Answer 20

represents the distribution away from the group average

Question 21

Types of Scores

Standardized

Composite Scores

Answer 21

an average

• statistically meaningful only if subtest scores are closely
• if large gaps in subtest scores, the doesn’t mean anything

Question 22

Standard Error of Measurement

Answer 22

error within instrument; how likely are we to have found the person’s true ability

LOOK AT NOTES FOR FIGURE

Question 23

Confidence Intervals

ON TEST

Answer 23

ranges from 68% to 90-95% confidence on the bell-shaped curve

• 95% will be a doubled SEM @ 68%
• 90% will range or is 1 less than 95%

NEED to DO On TEST:

• Which confidence interval to use and why
• 95% & 90% are too big to make clinical decisions with @ times

ASK STEPH

Question 24

Percent vs. Percentile Rank

Answer 24

percentage of people scoring at or below a particular score

Below 77.5standard score = 5.5 percentile rank qualify for services

Ex. 4 percentile rank–> “out of 100 children taking this test, you child scored better than 3 children”

OR

“out of 100 children you child scored less than 96 children”

Question 25

Normal Distribution

Answer 25

If have normative data, is there a good representative sample: * Gender * Ethnicity * Geographic (all over) * Enough people On TEST **_Why would a test not have a normal distribution?_** Ask Steph

Question 26

**_Test Sensitivity_**

Answer 26

React to many conditions * Catches condition (all disordered) and catches no condition too but makes sure it catches all that it wants * Screening test

Question 27

True negative?

Answer 27

A person is NOT identified with a disorder and in fact they DON'T have a disorder

Question 28

False positive? Which test does this risk?

Answer 28

Identify people as disordered when they're not Sensitivity test

Question 29

False negative? Which test does this risk?

Answer 29

Real disordered people are identified as not having a disorder Specificity test

Question 30

True positive?

Answer 30

A person is identified with a disorder, and they do in fact HAVE a disorder

Question 31

**_Test Specificity_**

Answer 31

Not sensitive, pays attention to one certain detail

Question 32

Incidence?

Answer 32

the frequency of a new illness in a population in a certain time

Question 33

Prevalence?

Answer 33

current number of people suffering from an illness in a given year