Lecture Notes: Signal Detection Theory

Key Concepts

• Noise Threshold: Represents baseline noise present during the task.
• Noise Distribution: A graphical representation of background noise.
• Signal Distribution: A graph that is shifted to the right relative to the noise distribution; indicates the presence of a signal.
• D Prime (d'): The difference between the means of the noise and signal distributions.
  • High d': Easier task, signal distribution far from noise.
  • Low d': Harder task, signal distribution close to noise.
• Stimulus Intensity (x-axis): Represents how easily a stimulus can be distinguished from the background noise.

Strategy Variables

• C: Represents the individual's strategy (response criterion).
  • B, D, C, Beta: Different strategies are used to determine threshold choices.

Strategy Types

1. B Strategy

   • Threshold choice based on an arbitrary value.
   • Example: Threshold set at 2; responses above are "yes," below are "no."
   • Hit Probability: Area above the threshold.
   • False Alarm Probability: Area below the threshold.

2. D Strategy

   • Threshold relative to signal distribution (d')
   • Formula: d' - B
   • Example: Threshold set at 2, d' = 1, threshold = 1
   • Response "yes" if above threshold, "no" if below.

3. C Strategy (Ideal Observer)

   • Minimizes misses and false alarms.
   • Formula: B - (d'/2)
   • Example: 2 - 1/2 = 1.5
   • "Yes" if above 1.5, "no" if below.
   • When C = 0: Ideal observer.
   • When C < 1: Liberal strategy (less often "no").
   • When C > 1: Conservative strategy (more often "no").

4. Beta Strategy

   • Threshold equals the ratio of signal height to noise height.
   • Formula: ln(beta) = d' * C
   • Example calculation: 1 * 1.5 = 1.5

Summary

• Participants adjust their threshold based on their strategy.
• The choice of strategy affects detection ability and accuracy.
• Each strategy is expressed through different methods and equations.