Chapter 5 - Quality of Behavior Measurement

Question 1

Validity

Answer 1

Does the data actually measure what it’s supposed to measure?

Best done though direct measurement during times of most concern

Text book example: you would not measure soreness to determine the a amount of miles a bicyclist has ridden

Question 2

Accuracy

Answer 2

Does the measurement reflect the true value of the behavior?

Example:

Question 3

Reliability

Answer 3

Are measurements consistent?

Example: a scale that consistently says 3lbs lighter even though this is not accurate

Question 4

Conservative Sample

Answer 4

Used when assessing intervention

Times when the frequency of the behavior is most likely to be different (higher or lower) than the desired treatment outcome

Question 5

Calibration

Answer 5

Comparing data obtained to its true value

Process for establishing true values must be different than the procedures from the other measurement procedure

Question 6

Interobserver Agreement: Definition

Answer 6

Extent to which two independent observers agree on the occurrence and nonoccurrence of a target behavior

Data on this should be taken at least 33% of sessions

Question 7

Event Recording: Total Count

Answer 7

Smaller Total/Larger Total x 100 = % Agreement

**Simplest
Total count of responses per measurement period

Question 8

Event Recording: Mean Count per Interval

Answer 8

Divide total observation period into equal times and have observers record responses for each time period. Then divide by the total number of intervals and multiply by 100

Question 9

Event Recording: Exact Count per Interval

Answer 9

Percent of intervals in which the two observers recorded the same frequency of behavior

**most stringent

Question 10

Event Recording: Trial by Trial

Answer 10

agreements/ agreements + disagreements x 100

For response opportunity data (either occurrence (1) or nonoccurrence (0))

Question 11

Duration/Latency Data: Total Duration

Answer 11

Smaller Time/ Larger Time x 100

No assurance that the same durations are being recorded for the same occurrence of the behavior

Question 12

Duration/Latency: Mean Duration per Occurrence

Answer 12

Take: smaller/larger x 100 for each occurrence. Then, average this across all occurrences

Question 13

Interval: Interval by Interval or Total Interval

Answer 13

intervals Agreement / Total Intervals x 100

**how many intervals where they agreed on the occurrence and nonoccurrence of the behavior

Question 14

Interval: Scored Interval

Answer 14

Only comparing intervals where both observers note the occurrence of the behavior.

Good for low rate behaviors since it ignores chances where scoring a nonoccurrence is very likely

Question 15

Interval: Unscored Interval

Answer 15

Considering intervals where either one or both observers recorded a nonoccurrence

Good for behaviors that occur at high rates