Signal Detection Theory & Application

Question 1

Definition of signal detection theory

Answer 1

Model of how humans separate “signal” –the relevant, important information from “noise” - all stimuli, whether relevant or not

Question 2

Assumptions of signal detection theory

Answer 2

1 - the world can be modeled as signal present or absent
2 - world is assumed to contain noise that is random and normally distributed
3 - signal increases the mean of the noise distribution

Question 3

Sensitivity

Answer 3

How well an operator is at separating a signal from noise (bottom-up)

Question 4

Response criteria,bias

Answer 4

Bias of the operator to say yes vs no (top-down)

Question 5

Response criteria risk

Answer 5

Beta < 1 is risky

Beta > 1 is conservative

Question 6

ROC curve

Answer 6

• Beta is the slope of the curve

- P(A) = d’ at a point (area under a curve)

Question 7

Optimal beta

Answer 7

Set based on signal-noise ratio multiplied by the costs and values

Question 8

Sluggish beta

Answer 8

The beta used by humans instead of the optimal beta

Question 9

Implications for design using SDT

Answer 9

• understanding the effects that changes to the system will have
• understand how to adjust a system based on how costly certain results are