Measurement Flashcards
Define: continuous variable
Quantitative variable that an theoretically take on values along a continuum
Define: variable
Characteristic that can be manipulated or observed and can take on difference values (quantitatively or qualitatively)
Define: discrete variable
A variable that can only be measured in separate units and that can not be measured in intervals of less than 1
Define: dichotomous variable
A nominal variable having only two categories such as yes/no, male/female
Define: construct
Concepts that represent nonobservable behaviors or events; invented names for abstract variable that cannot be seen directly example: intelligence, strength, health
Define: nominal measurement
Classifies variables into mutually exclusive categories (red, yellow, blue)
Define: ordinal measurement
Rank-ordered measurement
Unequal intervals between categories
Greater than, less than
(Manual muscle testing)
Define: interval measurement
Rank-ordered
Equal intervals between categories
No true zero
(Temperature, range of motion)
Define: ratio measurement
Highest level of measurement
Rank ordered
Equal intervals between categories
True zero
Define: Reliability
The extent to which a measurement is consistent and free from error
Define: systematic error
Predictable and directional, consistent from trial to trial
Define: random error
Measurement error due to chances, bidirectional
Define: validity
The extents to which an instrument measures what it is intended to measure
Define: sensitivity
Test’s ability to obtain a true positive
Define: specificity
Test’s ability to obtain a true negative
Define: positive predictive value
Probability that a person who tests positively actually has the condition.
Define: negative predictive value
Probability that a person who tests negatively does not have the condition.
Define: Positive likelihood ratio
True positive rate divided by false positive rate.
Sensitivity/(1-specificity)
Define: negative likelihood ratio
False negative rate divided my true negative rate.
(1-sensitivity)/specificity
Define: measurement
Process of assigning numerals to variables to represent quantities of characteristics according to certain rules
Sources of measurement error (3)
Individual taking the measurement
Measuring instrument
Variability of the characteristic being measured
3 types of variables
Continuous
Discrete
Dichotomous
4 levels of measurement
Nominal
Ordinal
Interval
Ratio
Reliability coefficient
Ratio of true score variance to the total variance
T/(T+E)
.75 is good
>.90 is usually required for clinical decision-making
Types of reliability (4)
Test-retest
Rater
Internal consistency
Alternate forms
Test-retest reliability
Consistency of measurement from time 1 to time 2
Intrarater reliability
How consistently one individual measures a characteristic across toe or more trials
Interrater reliability
Consistency of measurement between two or more raters measuring the same characteristic
Internal consistency (type of reliability)
Extent to which multiple items agree when measure aspects of the same characteristic i.e. a pain survey may have many questions, higher consistency=higher reliability
Alternate forms (type of reliability)
Consistency between two instruments designed to measure the same characteristic
Validity vs reliability
To be valid, a measure MUST be reliable
Not all reliable measurements are valid.
Types of measurement validity (6)
Face validity Content validity Concurrent validity Predictive validity Prescriptive validity Construct validity
Face validity
An instrument “appears” to measure what it’s purported to measure; least rigorous form of validity
Content validity
Extent to which a measurement generalizes to a theoretical universal score. Are all parts of the whole defined? i.e. A pain questionnaire has more content validity than an 11pt scale
Concurrent validity
Variable measured with instrument of interest and “good standard” concurrently.
Predictive validity
Degree to whic a measurement is predictive of some future score
Prescriptive validity
Degree to which a measurement can be used as a basis for selecting an appropriate intervention i.e. Balance score dictates certain ambulatory device
Construct validity
Ability of an instrument to measure an abstract construct; not directly observable