Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
27 Cards in this Set
- Front
- Back
In regard to Precision Vs. Accuracy, Reliability is ____ while Validity is _______
|
Precision = Reliability
Accuracy = Validity |
|
Definition of Reliability
|
Measure repeated outcomes with consistent scores; the extent to which a measurement is consistent and free from error
|
|
What is Test-Retest Reliability
|
a basic premise of reliability is the stability of the measuring instrument; will obtain the same results with repeated tests
|
|
What is intrarater reliability
|
Refers to the stability of data recorded by one individual
|
|
What is interrater reliability
|
concerns variation between two or more raters who measure the same group of subjects
|
|
What is the primary criteria for choosing an appropriate time interval in test-retest intervals
|
the stability of the response variable and the test's intended purpose.
|
|
Time intervals in test-retest intervals should be far apart enough to avoid
|
fatigue, learning, or memory effects, while being close to enough to avoid genuine changes in the measured variable
|
|
reliability can be influenced by the ____ of the first test on the _____ of the second test
|
effect, outcome,
and example is that a test of dexterity may improve because of motor learning. |
|
Rater Bias is
|
for when the rater has a vested interest in the outcome, or when one rater takes two measurements, because raters can be influenced by their memory of the first score.
|
|
what is systematic error
|
predictable errors of measurement, consistently overestimating or underestimating the true score, constant and biased
|
|
If a systematic error is detected, you can correct it by
|
recalibrating the system or to adjust for it by adding or subtracting the appropriate constant
|
|
What is Random Error
|
measurement are due to chance and can affect the score in an unpredictable way from trial to trial.
|
|
An example of Random error is
|
measuring a patient's height and the patient moves while his height is being measured each time, so it will be inconsistent
|
|
What is regression toward the mean
|
single test's potentially extreme score; multiple tests reveal score closer to group average
|
|
what is the classical reliability theory
|
Single score made up of true score & random error gives best estimate of actual value
|
|
what is the generalizability theory
|
single score made up of true score and various types of error. Must consider the specific measurement scenario when considering reliability of measure, including test-retest error, rater error, random and other measurable errors
|
|
What is Reliability Coefficient
|
an estimate of the extent to which a test score is free from error. Is expressed as a ratio True score Variance/ True score variance + Error Variance. Worst value is 0, best reliability value is 1
|
|
What is correlation
|
Correlation reflects the degree of association between two sets of data, or the consistency of position within two distributions. (relationship)
|
|
what is agreement
|
Recording same actual values. You can have perfect correlation but if there is no agreement it can illustrate poor reliability
|
|
For statistical tests of reliability, the ICC (intraclass correlation coefficient) has become the preferred index because it reflects both ____ and ____
|
correlation and agreement
|
|
Kappa Statistic and Percent Agreement are used with
|
nominal data / categorical data
|
|
Pearson's and Spearman's correlation coefficients are used with
|
interval-ratio data and ordinal data in the continuous data category
|
|
Standard error of measurement is
|
how confident you are that the measure falls within a certain range. SEM = std dev x sqrt(1- reliability coefficient)
|
|
The greater the reliability, the _____ the SEM (standard error of measurement)
|
smaller
|
|
Variation around a measure example. While using grip strength and a dynomometer, if there was a 95% variation around a measure = +/- 3 ft.lbs, and the subject measured to have 25ft.lbs of grip strength you would be ____ confident the true grip strength lies between ___ and ___ ft.lbs
|
95%, 22 and 28 ft.lbs
|
|
minimal detectable change/difference (MDC) is
|
used to define the amount of change in a variable that must be achieved to reflect a true difference
|
|
minimal detectable change is not a clinical measure for measuring meaningful improvement, it is only a _____ measure
|
statistical
|