Humans vs. Automated Systems

1
Humans vs. Automated Systems

• Social Cognitive Considerations

Teddy McCalley Technical University
Eindhoven Department of Technology Management
2
Pop Quiz 1. What is domotics?

1. A Dutch word for smart homes.
2. A computer that runs the house.
3. Household appliances that talk to each other.
4. A house that makes things easier for old people.
5. An automated system to save energy.
6. A status symbol.

3
Definition
A dwelling incorporating a communications
network that connects the key electrical
appliances and services and allows them to be
remotely controlled, monitored or accessed. -
both within the dwelling and from outside the
dwelling.

1
4
What makes a home smart?

1. Internal network between appliances wire,
   cable, wireless
2. Intelligent control gateway to manage the
   systems
3. Home automation products within the homes and
   links to services and systems outside the home.

5
(No Transcript)
6
How

• Wireless laptops
• PDA devices (personal Digital Assistants) a PDA
  device can be used as a remote control device in
  the house. Status information and energy savings
  can be stored and checked.
• WAP (Wireless Application Protocol) Phones
  capable of internet access can be used to check
  conditions in the house.
• Office PC  

7
A home can be smart in 6 areas

• Environmental (heating/water, lighting, energy
  management, metering)
• Security (alarms, motion detectors, environmental
  detectors)
• Home entertainment (audio, visual, internet)
• Domestic appliances (cooking, cleaning,
  maintenance alerts)
• Information and communication (phone, Internet)
• Health (telecare, home assistance)

8
Using the system to save energy

1. Human-system interaction
2. Motivation, feedback and goal setting
3. Conflict

9
Goal-Directed Energy Feedback Conservation
Actions and the Goal Hierarchy
10
Goal-Feedback Relationship

• A goal without feedback cannot be reached.
• Feedback that does not relate to a goal is
  useless.

11
The Feedback Project

• Is there a goal to save energy present in todays
  user?
• If not, will creating a goal generate energy
  conservation behavior?
• What type or level of goal is best to use and why?

12
Product-Integrated FeedbackExample Washing
Machine

• Programmed into the appliance.
• Immediate.
• Frequent or continuous.
• Per act or cumulative.
• Antecedent or consequence information.
• Inexpensive.

13
Psychological Functions of Feedback

• Goal attainment.
• Teaching skilled responses.
• Providing rewards for conservation actions.
• Motivating conservation behavior.

14
Simulated Control Panel
15
Experiments tested...

• Simple consequence feedback.
• Antecedent information.
• Goal plus feedback interventions.
• Social strategy with goal plus feedback.

16
Experiments

• cost per wash
• cumulative kWh
• cumulative cost
• kWh per wash
• kWh only.
• Antecedent information only.
• Antecedent information plus kWh.

17
Self-set vs. Experimenter-set Goal
22
20
18
Main effect of Goal p .03
16
14
12
10
8
Percent kWh Saved
6
4
2
0
FB-ESG
FB-SSG
FB-NG
NFB-NG
Feedback and Goal Condition
18
Why did the goal plus feedback strategy work?

• Feedback Intervention Theory (Kluger De Nisi,
  1996)
• Minimal Justification (e.g. Katzev Johnson,
  1983)

19
FIT Theory

• Behavior is regulated by comparisons of feedback
  to goals or standards.
• Goals or standards are organized hierarchically.
• Attention is limited and therefore only
  goal-feedback gaps that receive attention
  actively participate in behavior regulation.

20
Theory cont.

• Attention is normally directed to a moderate
  level of the hierarchy.
• Feedback interventions change the locus of
  attention and therefore affect behavior.

21
Task Performance HierarchyExample Washing
clothes

• Meta-task processes Im a clean person.
• Task-motivation processes I must do the laundry
  today.
• Task-learning processes Whats the best
  temperature for washing a wool sweater?

22
Minimal Justification Principle

• An individual is more likely to attribute
  behavior to the self if a weak, rather than a
  strong, external justification is given.
• Weaker justifications allow the individual to
  develop a more powerful internal control (e.g.
  attitude).
• Foot-in-the-door technique.

23
Minimal Justification
30
Goal effect p lt.0001
20
MJ x Goal p .04
10
Percent kWh Saved
MJ Treatment
No
0
Yes
Self-set
None
Goal Condition
24
Minimal Justification and Goal Setting
20
10
Energy Saving Goal
5
10
15
20
N
0
Treatment
No Treatment
MJ Treatment
25
Conclusions

• Feedback works at the task level.
• Manipulations that draw attention to a higher
  level goal attenuate the effects of feedback.
• Minimal justification techniques might positively
  influence conservation attitude but a higher
  level goal to conserve is susceptible to
  distraction.
• Limited attentional resources determine response
  to feedback.

26
Information Specificity as a Moderator of Goal
Setting

• Impacts on Energy Efficient Consumer Behavior in
  the Programming of the Thermostat
• .

27
Research Questions

• Does the level of information affect conservation
  strategy planning?
• Does information moderate goal response?
• Does planning time affect energy conservation
  performance?
• What is the role of experience?
• Is there a conflict between conservation and
  comfort?

28
Information and Goals
Information -gt Planning time -gt Strategy
development -gt Performance Goal -gt Motivation -gt
(Strategy Development) -gt Performance
29
Experiment

• 2 (goal) x 2 (information) x 2 (experience)
• Dep. Var. Planning time (minutes)
• Dep. Var. Percent of change score

30
Results Strategy Planning Information Level x
Goal Level
4.5
4.0
3.5
Mean Planning Time (Minutes)
3.0
2.5
2.0
1.5
1.0
.5
0.0
Low Info-High Goal
Low Info-Low Goal
High Info-High Goal
High Info-Low Goal
Specificity Levels (Low and High)
31
Results Energy Conservation Information Level x
Goal Level
High Low Specificity
10.0
5.0
8.0
2.5
2.8
2.1
6.0
Mean Percent Temp. Change
5.8
4.9
4.8
4.0
4.2
Experience
2.0
yes
0.0
no
NSI-SG
NSI-NSG
SI- SG
SI-NSG
32
Information and Goal Level Set
11
10
9
Mean Goal Level Set ()
8
Experience
7
no
6
yes
yes
no
Specific Information
33
Goal Response Inexperienced Users
34
Goal Response Experienced Users
35
Goal Level, Experience Perception of
Conservation-Comfort Conflict
Count
Energy-Comfort
Conflict
Thermostat Experience
no
yes
Total
.00
Goal
no
2
4
6
Level
5.00
5
5
10.00
4
4
8
15.00
3
3
20.00
3
1
4
17
9
26
Total
.00
Goal
yes
2
2
Level
5.00
3
4
7
10.00
5
2
7
15.00
1
1
8
9
17
Total
36
Experience and Goal Level Set
Count
Thermostat Experience
no
yes
Total
.00
Goal
6
3
9
Level
5.00
5
10
15
10.00
8
8
16
15.00
3
1
4
20.00
4
1
5
Total
26
23
49
37
Summary

• Without a specific goal, information has little
  effect.
• Specific goal plus specific information increases
  planning time.
• Specific goal without specific information
  decreases planning time.
• Experienced users do not conserve as much as
  inexperienced users.

38
Conclusions

• Information campaigns alone will have little
  effect.
• Energy conservation loses to comfort.
• If inexperienced users are setting their goals
  high and changing temperature settings to match
  can we intervene so they continue saving energy.

39
Conclusions cont.

• A basic cognitive strategy works well.
• Product-integrated feedback can yield substantial
  energy savings.
• Various psychological strategies can be
  successfully used through a programmable
  appliance interface.

40
Conclusions

• Goal-setting is a highly successful intervention
  to create effective feedback but success is
  influenced by personality factors.
• Using intrinsic control processes such as
  anchoring can be effective and may be less
  sensitive to personality and other grouping
  factors.

41
Conclusions and Recommendations

• Feedback alone isnt optimal.
• Feedback plus a goal saves energy.
• Different kinds of goals work better with
  different personalities.
• Social strategies can work, but not with a
  goal-feedback strategy.

42
Recommendations

• Use a goal-setting intervention with consequence
  feedback.
• Avoid directing attention to the self.
• Set default settings as low as possible.
• Keep tasks simple when using feedback.

43
Recommendations for Effective Feedback

• Use goal-setting but be aware of group
  differences.
• Supplement with basic intrinsic influences (e.g.
  anchoring) where possible.
• Use existing goal if possible.
• High frequency (e.g. per act).
• Credibility
• Do not draw attention to the self.

44
Feedback effects are

• Damaged by cues that direct attention to high
  levels of the goal hierarchy.
• Helped by cues that direct attention to
  task-learning processes.
• Helped by keeping the task easy.
• Helped by goal-setting.

45
The End
The End