What is measurement? Module 1: Measurement

Module 1: Measurement

James Carr Linda LeBlanc

Jessa Love

Western Michigan University

What is measurement?

? Assigning numbers and units to features of objects or events

? Direct and frequent measurement is the foundation of applied behavior analysis

? We use measurement to detect and compare the effects of environmental events on behavior

Why is measurement important in research?

? Describes phenomena in precise, consistent, and publicly verifiable ways

? Decreases subjectivity and prejudices of the researcher

? Provides reliability - measures are consistent and repeatable

? Provides validity - the extent to which a measure of X actually measures X and not Y

Why is measurement important in practice?

? Provides valuable information:

? During baseline, before intervening ? During intervention, allowing for modification ? After intervention, on effectiveness

? Prevents important mistakes:

? Continuing an ineffective intervention ? Discontinuing an effective intervention

What is an operational definition?

? Defines a behavior in terms that are precise so the behavior is measurable and an occurrence can be agreed upon by two or more observers

? In general, the more detailed the operational definition, the better

A good operational definition is:

? Objective

? Defines behavior in terms that are observable

? Does not make inferences about mental states or processes (e.g., the person's intention)

? Clear

? Unambiguous

? Complete

? Includes examples and non-examples of the target behavior

1

How can we evaluate an operational definition?

? Test it with multiple observers - when they are able to agree on occurrences and non-occurrences of the target behavior, the operational definition is sufficient.

Let's try an example:

? Define hand raising:

? lifting one arm... ? so the arm is straight (elbow and wrist are

not bent)... ? the bicep is level with the ear... ? and the hand is open (fingers are not

bent)... ? for a maximum duration of 5 s or until the

child's named is called ? only code during an instructional session

How do we select the dimension of behavior to measure?

? This will depend on the research question

? What aspect of the behavior are you trying to change?

? Example: If you are trying to decrease how often a problematic behavior occurs, you would want to measure and record how often the behavior occurs (frequency)

Methods of Data Collection: Frequency or Rate

? Event recording - count each response as it occurs

? Requirements for use:

? Target behavior is discrete, or countable ? Target behavior does not happen too frequently

(extremely high rate behavior can be difficult to count accurately)

Methods of Data Collection: Frequency or Rate

? Examples:

? Rate of aggressive behavior during oneon-one teaching (occurrences of aggression per minute)

? Rate of social initiations made during group play (initiations per hour)

Methods of Data Collection: Frequency or Rate

? Procedure: ? Record date ? Record start and stop time for observation/session ? Record each occurrence of the target behavior (example: tally)

2

Methods of Data Collection: Frequency or Rate

? Sample data:

Participant: __X__

Date:

Observer: __Y__ Start time:

Target Behavior: __head hitting__ Stop time: Number of occurrences:

Primary/ Secondary Total:

1/21/08

9:05am 9:15am

12

1/22/08

10:30am 10:40am

9

1/23/08

9:45am 9:55am

15

Methods of Data Collection: Frequency or Rate

? How are data reported?

? If each observation period is the same length, you can simply report the number of occurrences

? FREQUENCY

? If the length of the observation period varies, convert the frequency count to rate and report the number of occurrences per unit time (example: 6 responses per minute)

? RATE

FREQUENCY: Sidener et al. (2006)

RATE: Esch & Carr (in press)

Methods of Data Collection: Momentary Time Sampling (MTS)

? Record occurrences of the target behavior only at the end of an brief intervals during a longer-duration observation period

? When to use:

? Target behavior has a longer duration, not readily countable

? Target behavior occurs at a high rate ? You do not have the means to observe the target

behavior continuously

Methods of Data Collection: Momentary Time Sampling (MTS)

? Examples:

? Out-of-seat behavior during one-on-one teaching

? Engagement in interactive play with peers during group play

3

Methods of Data Collection: Momentary Time Sampling (MTS)

? Procedure: ? Divide the observation period into equal intervals (at least 10) - set a timer to indicate the end of the interval ? Watch for target behavior at the END of the interval ? Ignore occurrences DURING the interval ? At the end of the interval

? Record "Y" if the behavior occurred at the moment the interval ended

? Record "N" if the behavior did not occur at the moment the interval ended

? The timing device should automatically restart

Methods of Data Collection:

Momentary Time Sampling (MTS)

? Sample data:

Participant:__X__ Observer: __Y__ Target behavior: Primary/Secondary Toy engagement

Obs. #:

Target behavior?

Obs. #:

Target behavior?

1

Y

N

6

Y

N

2

Y

N

7

Y

N

3

Y

N

8

Y

N

4

Y

N

9

Y

N

5

Y

N

10

Y

N

Total Y's: __6__ Percentage: 60%

Methods of Data Collection: Momentary Time Sampling (MTS)

? How are data reported? ? Count the number of observations during which the target behavior occurred and divide by the total number of observations/intervals

? Report as percentage of observations/intervals with target behavior

Alvero et al. (2007)

Methods of Data Collection: Trials

? Measures whether a target behavior occurs following the relevant opportunity or instruction

? When to use:

? Target behavior is a restricted operant - the occurrence of the target behavior is limited to when trials are presented

Methods of Data Collection: Trials

? Examples:

? Imitating the motor behavior of a teacher when instructed to do so

? Responding correctly to questions about personal information

4

Methods of Data Collection: Trials

? Procedure: ? Present the relevant instruction or opportunity to engage in the target behavior ? Record "Y" if the target behavior occurs ? Record "N" if the target behavior does not occur

Methods of Data Collection: Trials

? Sample data:

Participant:__X__ Observer: __Y__

Trial #: 1

Independent Response?

Y

N

Target behavior: Gross motor imitation - clap

Trial #:

6

Primary/Secondary

Independent Response?

Y

N

2

Y

N

7

Y

N

3

Y

N

8

Y

N

4

Y

N

9

Y

N

5

Y

N

10

Y

N

Total Y's: __4__ Percentage: 40%

Methods of Data Collection: Trials

? How are data reported? ? Percentage accuracy: report the percentage of correct responses out of the total number of opportunities, or out of a block of trials

? Trials-to-criterion: report the number of trials needed to reach a predetermined level of performance

Percentage Accuracy: Chong & Carr (2005)

Trials to Criterion: Lechago et al. (2008)

Practical Tips: Collecting data live vs. from video

? Live

? Can the individual implementing the intervention record data while doing so?

? Is a second observer available periodically to collect secondary data?

? If "yes" to both, live data collection should work well.

5

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download