3 - Measurement Properties Part 2

Question 1

___ measures relative reliability and is unitless

Answer 1

ICC

Question 2

___ measures absolute reliability and has the same units as the measure

Answer 2

Standard Error of Measurement (SEM)

Question 3

Internal Consistency

Answer 3

extent to which items measuring different aspects of the same construct are related

Question 4

Cronbach’s Alpha

Answer 4

measures internal consistency, includes correlation btwn each item and correlation of each item w/ the total score

Question 5

What is a good Cronbach’s Alpha value?

Answer 5

0.70 to 0.95

Question 6

Responsiveness

Answer 6

ability to detect change over time, measured by MDC

Question 7

Minimal Detectable Change (MDC)

Answer 7

Smallest change an instrument can detect that passes threshold for error.
If change > MDC, this change is not d/t error.

Question 8

Higher reliability = ____ (large/small) MDC

Answer 8

small

Question 9

True or False: MDC provides information about whether this is a meaningful clinical change

Answer 9

False

Question 10

What 2 statistics quantify magnitude of change in standardized units?

Answer 10

Effect Size (Cohens d)
Standard Response Mean (SRM)

Question 11

Difference between Cohens d and SRM

Answer 11

Cohens d: ratio of mean change to baseline SD.
SRM: ratio of mean change to change SD.
Both = difference in means / SD

Question 12

Cohen’s d criteria for large/moderate/small treatment effects

Answer 12

0.8 = large effect
0.5 = moderate effect
0.2 = small effect

Question 13

Interpretability

Answer 13

meaningful change over time, measured by MCID

Question 14

Minimal Clinically Important Difference (MCID)

Answer 14

smallest difference in the measurement that signifies an important change in the patient’s condition.
If change > MCID, then an important change has occurred.

Question 15

Distribution-based method of estimating MCID

Answer 15

assess the degree of change in a group of patients.
Within groups (pre-test and post-test scores compared) or between groups (placebo and intervention compared).

Question 16

Anchor-based method of estimating MCID

Answer 16

use an external criterion to determine whether changes in outcome scores are clinically meaningful

Question 17

Distribution-based MCID within groups is measured by

Answer 17

Effect Size
SRM

Question 18

Distribution-based MCID between groups is measured by

Answer 18

Effect Size

Question 19

Which types of measurements are NOT great for assessing change?

Answer 19

Nominal
Ordinal

Question 20

Sensitivity describes the ____ rate and is good for ruling ____

Answer 20

True positive rate
Rule OUT

Question 21

Specificity describes the ____ rate and is good for ruling ____

Answer 21

True negative rate
Rule IN

Question 22

Sensitivity =

Answer 22

= # correctly test positive / # true positive

Question 23

Specificity =

Answer 23

= # correctly test negative / # true negative

Question 24

Receiver Operating Characteristic (ROC) Curve

Answer 24

Helpful for determining the risk/benefit.
A plot of the true positive rate (sensitivity) on Y axis versus the false positive rate (1-specificity) on X axis.

Question 25

Youden Index

Answer 25

Point on ROC curve where it “turns”

Question 26

Area under ROC curve, highest and lowest

Answer 26

0.5 = useless test, no predictive value.
1.0 = perfect test, 100% true positives and 0% false positives

Question 27

Area under ROC curve criteria

Answer 27

.90-1.0 = excellent (A)
.80-.90 = good (B)
.70-.80 = fair (C)
.60-.70 = poor (D)
.50-.60 = fail (F)