Shaping Chapter 8. Simple Reinforcement Before: Andrew has no gum Behavior Initial: Andrew moves his lips After: Andrew receives gum.

Shaping

Chapter 8

Simple Reinforcement

Before:

Andrew has no gum

Behavior

Initial:

Andrew moves his lips

After:

Andrew receives gum

Differential Reinforcement

Before:

Andrew has no gum

Behavior Intermed:

Andrew moves his lips

Behavior Intermed:

Andrew makes croaking sound

After:

Andrew receives no gum

After:

Andrew receives gum

Reinforcement

Extinction

Differential Reinforcement

Before:

Andrew has no gum

Behavior Intermed:

Andrew makes croaking sound

Behavior Terminal:

Andrew says “gum”

After:

Andrew receives no gum

After:

Andrew receives gum

Reinforcement

Extinction

Process of Shaping

• Used when the response is not currently in the repertoire

• Is used as an intervention

• We don’t shape people, we shape responses

Operant Level

• The frequency of responding before reinforcement

Terminal Behavior

• Behavior not in the repertoire or not occurring at the desired rate;

• This is the goal of the intervention

Shaping Behaviors

• Initial behaviors

• Intermediate behaviors

Initial behavior

• Behavior that resembles the terminal behavior along some meaningful dimension and occurs with at least a minimal frequency

Intermediate behaviors

• Behavior that more closely approximates the terminal behavior

The procedure of shaping with reinforcement

• The differential reinforcement of only that behavior that more and more closely resembles the terminal behavior

Shaping with Reinforcement

Before:

Andrew has no gum

Behavior Initial: NA

Intermed: Moves lips only

Terminal: Says words unclearly

Behavior Initial: Andrew moves lips

Intermed: Andrew makes croaking sound

Terminal: Says words clearly

After:

Andrew receives no gum

After:

Andrew receives gum

Reinforcement

Extinction

Differential Reinforcement vs. Shaping

Differential Reinforcement

Shaping

# of response classes

One A series

Successive Approximations to the terminal behavior

No Yes

Some terminal behavior at the start

Yes No

Differential Reinforcement vs. Shaping – fill in the table

Shaping Differential Reinforcement

# of response classes

Successive Approximations to the terminal behavior

Some terminal behavior at the start

Shaping with punishment

• The differential punishment of all behavior except that which more and more closely resembles the terminal behavior

Variable-outcome shaping

• Shaping that involves an increase in the magnitude of a reinforcer or a decrease in the magnitude of an aversive outcome as performance more and more closely resembles the terminal behavior

Variable Outcome Shaping

Before:

Rod has no rattle in his hand Behavior Initial:

weak force

Intermed: forceful

Terminal: very forceful

Behavior Initial: forceful

Intermed: very forceful

Terminal: superforceful

After Initial: no rattle

Intermed: Rattle in 10s

Terminal: Rattle in 6 sec

After Initital: Rattle in 10-sec

Intermediate: Rattle in 6s

Terminal: Rattle in 4s

Quicker Reinforcement

Slower Reinforcement

Fixed outcome shaping

• Shaping that involves the delivery of a fixed magnitude of a reinforcer, when performance meets the changing criterion,

• or the delivery of a fixed magnitude of an aversive outcome, when performance fails to meet the changing criterion.

Fixed Outcome vs. Variable Outcome Shaping

Fixed Outcome

Variable Outcome

# of outcome sizes

One Many

Regression to earlier levels

No reinforcers Weaker reinforcers

Usual Source of Shaping

Behavior Modifier (planned)

Nature (unplanned)

Operant Classes

R1

R2

R3

R4

SRA

R5

R6

R7

R8

SRB

Differential Reinforcement

Shaping

• “Creates” new behavior (creates classes)– adds responses to an individual’s repertoire

• Differential reinforcement of successive approximations

Shaping

• Artificial (i.e., deliberate)

Deliberate Shaping1. Select a final target behavior

2. Measure current distribution (baseline)

3. Differentially reinforce an approx.

4. Diff. reinf. new approx.• Place previous approx. on EXT

5. Continue until target behavior occurs

Eckerman et al. (1980)

• Interested in developing principles of shaping

• Two questions– Step Size? – How fast (i.e., step rate)?

Eckerman et al. (1980)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Nominal Operant

Class

Induction(Response Generalization)

Functional Operant Class

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

Eckerman et al. (1980)Independent Variables

• Step Size = Change in distance each step

• Step Rate = How long at each step

Eckerman et al. (1980)Dependent Variable

• Accuracy– Percentage of pecks at target locations

Eckerman et al. (1980)Results

• Accuracy was unaffected by step size or step rate

• Just as accurate with large steps and a criterion that changes rapidly

• Under these conditions, relatively large steps and relatively rapid change are best

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

vv

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Small Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Large Steps

Eckerman et al. (1980)Results

• Accuracy was unaffected by step size or step rate

• Just as accurate with large steps and a criterion that changes rapidly

• Under these conditions, relatively large steps and relatively rapid change are best