Psychometrics Part 1

psychometrics

the science of measuring mental capacities and processes

identifying what is normal

• measure group of people and obtain a range f scores

• identify the extreme ends of the range

• make the decisions on what we considered within normal limits (WNL)

intellectual disability definition

• significantly impaired cognitive functioning

• deficits in two or more adaptive behaviors

• onset before 18

language disorder definition

an impairment in comprehension and/or use of spoken, written or other symbol system (form, content or function of language in communication

• operational definitions:

  • 1.5 SD below the mean

  • at least 2 areas of language affected

  • ability to participate in educational activities negatively impacted

normal distribution

also called theoretical distribution, the range of values is spread out in predictable ways

• very RARE to have this appear in your data

• 68% within 1 SD, 95% within 2 SD, 99% within 3 SD

population

the collection of cases that comprise the entire set of cases with the specified characteristics (e.g. all living adult males in the US)

• any value generated from or applied to the population is a parameter

sample

a collection of cases selected from a larger population

• any value derived from the sample, such as mean, is a statistic

end purpose of assessment

to draw conclusions about an individual’s preformative and functional communicative and swallowing abilities

code of fair testing practices

1. select appropriate test that meets intended purpose and appropriate for test taker

2. administer and score correctly and fairly

3. repost and interpret results accurately and clearly

4. inform test takers about nature of test, rights and responsibilities

assessment methods

• norm-referenced tests

• criterion-referenced tests

• authentic/dynamic assessment approach

norm-referenced tests

• most common

• always standardized

• comparison on individual’s prefromae to “normative group” → “how does my client compare to the average?”

• normal distribution ( symmetrical and dependent on mean)

typical disordered ranges

• 1.25 Standard Deviations

  • 81 - Standard Score*

  • 10 - Percentile Rank

• -1.5 Standard Deviations

  • 77.5 (77 or 78) - Standard Score*

  • 7 - Percentile Rank •

• -1.75 Standard Deviations

  • 74 - Standard Score*

  • 4 - Percentile Rank

norm-referenced test advantages

• objective

• compare skills to large group of similar individuals

• efficient

• common (most can administer with instruction)

• generally covered by insurance

norm-referenced test disadvantages

• does not allow individualization

• static

• unnatural environment

• evaultes isolated skills

• exact administration to be valid and reliable

• may not be culturally or linguistically diverse

criterion-referenced tests

• may or may not be standardized

• allow comparison of individual’s performance to predefined criterion → “how does my client’s performance compare to the expected level of performance?”

• assume there is a level of performance that must be met for behaviors to be acceptable

when do we use criterion-referenced tes

• when it doesn’t make sense to rank people relative to each other (e.g. voice quality)

• when norms on a norm-referenced measure are inappropriate (e.g. cultural diversity)

• when specific information concerning behaviors or knowledge is required (e.g. establishing baselines)

criterion-referenced test advantages

• objective

• compre skills to predefined criterion

• efficient

• common

• generally covered by insurance

criterion-referenced test disadvantages

• don’t always have a standard score

• don’t always have appropriate criterion

authentic assessment (dynamic assessment)

• evaluates learning potential based on ability to modify response after teaching

• often used with clients with cognitive communication disorder or with culturally diverse backgrounds

• test-teach-retest method

authentic assessment advantages

• natural

• self-monitoring and evaluation

• individualization

• flexible

authentic assessment disadvantages

• lack objectivity

• not usually standardized (reduced reliability and validity)

• require high level of clinical ability

• inefficient

• may be impractical

• may not be covered by insurance

evaluating published test

1. basic evaluation → how useful is the test?

2. evaluating the norms → how good are the norms?

3. evaluating the test itself → how good is the test?

how useful is the test? WILL BE TEST QUESTION

• when was the test developed

• is the purpose clearly stated and does it match your intended use

• does my client fit into the age range of the norms

• is the administration time practical relative to your protocol

• are the administration and scoring guidelines easy to understand

• does the test allow you to obtain scores in a format that you might need

• does the test accurately differentiate between typical and disordered clients

classification

• sensitivity: correct identification of a disorder

• specificity: correct identification of normal development

• false positives: incorrectly identified as disordered

• false negatives: incorrectly identified as normal

how good are the norms

• how big is the sample

• how were the norms developed

• is my client sufficiently represented in the norms

how good is the test

• how valid is the test

• how reliable is the test

validity

how accurately does the measurement tool measure the characteristic it claims to assess

• construct

• content

• criterion-related

  • concurrent

  • predictive

reliability

how consistent are the test results

• interjudge (two judgers)

• intrajudge (one judger scores twice)

• test-retest

• splot-half