by Sang, Qingzi

1

IS2350 Human Factor Final Exam. Presentation

• by Sang, Qingzi
• Fall, 2003

2
Presentation Content

• The presentation is part of the final examination
  of IS2350 Human Factor.
• It will cover the content from page 17 to 43 of
  Chapter 2 in Engineering Psychology and Human
  Performance.

3
Topics

• Signal Detection Theory
• The ROC Curve
• Application of Signal Detection Theory
• Vigilance

4
Signal Detection Theory
State of the World
Signal
Noise
Hit
False Alarm
Yes
Response
No
Correct rejection
Miss
5
Signal Detection Theory

• Two stages of information processing in the task
  of detection
• Sensory evidence (X) is aggregated concerning the
  presences or absence of the signal
• A decision is made about whether this evidence
  indicates a signal or not (if XgtXc)

6
Signal Detection Theory

• Why False Alarm and Miss?
• False Alarm Say yes when theres no signal
• Miss Say no when signal presents.

A
X
B
Time
7
Signal Detection Theory
Correct rejection
PX(N or S)
Hit
False alarm
Miss
Xc
X
P(H)P(M)1 and P(FA)P(CR)1
8
Signal Detection Theory
More yes, more hit and also more false alarm
More no, more correct rejection and also more miss
9
Signal Detection Theory

• Optimal Beta where beta should be set
• Expanded Optimal Beta with benefit and cost

10
Signal Detection Theory

• Sluggish Beta

Human adjust their beta in response to changes in
payoffs and probabilities
Obtained ß
probabilities
payoffs
Optimal ß
11
Signal Detection Theory

• Sensitivity

High d
Low d
12
Receiver Operating Characteristics
O
C
R
20 10
P(H)
0.95
Low ß
0.70
0.60
15 15
0.20
ß1
10 20
0.05
0.30
0.40
0.80
15 15
Low d ß1
High ß
P(FA)
13
Application of Signal Detection Theory

• Benefits of SDT
• It provides the ability to compare sensitivity
  and therefore the quality of performance between
  conditions or between operators that may
  different in response bias.
• By partitioning performance into bias and
  sensitivity components, it provides a diagnostic
  too that recommends different corrective actions,
  depending on whether change in performance
  results form a loss of sensitivity or a shift in
  response bias.

14
Application of SDT

• Medical Diagnosis

salience of abnormality and symptoms
Sensitivity
training of physician
signal probability
Response Bias
payoffs
15
Application of SDT
How to reduce the bias in lineup

• Tell witness the suspect may not be in the lineup
• Increase the similarity of people in lineup
• Use blank lineup control
• Use mock witness control

16
Vigilance

• What is vigilance task?
• An operator is required to detect signals over a
  long period of time, and signal are intermittent,
  unpredictable, and infrequent.
• Operators often show lower vigilance levels than
  desirable
• Vigilance decrement
• it can arise either as a result of a decrease in
  sensitivity or as a shift to a more conservative
  criterion

17
Vigilance

• Vigilance Paradigms
• Free-response VS. Inspection
• Successive VS. Simultaneous
• Sensory VS. Cognitive

18
Vigilance

• Factors affect sensitivity in a vigilance task
• Targets signal strength reduced sensitivity
• Uncertain about the time or location of target
  signal sensitivity
• In inspection task, event rate increase,
  sensitivity
• The sensitivity level is higher for simultaneous
  tasks than for successive tasks
• When observers are well trained --- sensitivity
  decrement
• Sensitivity increment sometimes occurs in
  simultaneous paradigms with cognitive but not
  sensory stimuli

19
Vigilance

• Factors affects bias
• Higher target probability decreasing beta
• Payoffs affect response bias as in the signal
  detection task.
• Increased beta values are evident when signal
  strength is reduced

20
Vigilance

• Sensitivity Loss Fatigue and Sustained Demand
  Theory
• sustained demand over time leads to the
  sensitivity decrement and that factors demanding
  greater mental resources will lower sensitivity
  level

21
Vigilance

• Combat the loss increasing sensitivity

• Show target example (reduce memory load)
• Increase target salience
• Vary event rate
• Train observers

22
Vigilance

• Combat the loss shift in response criterion

• Instructions
• Knowledge of Results
• False Signals
• Confidence levels

23
Thanks to the cooperation of Baby Vivian