Psychophysics & Signal Detection Theory

Question 1

Absolute thresholds

Answer 1

the minimum physical intensity of a stimulus presented that can be sensed, always measured at a percentage (e.g., likelihood of seeing the smallest presentation of light 50% of the time)

Question 2

Difference thresholds (a.k.a. JND)

Answer 2

the minimum difference between two stimuli intensities that can be noticed, also given as a likelihood

(e.g., 50% of the time people can tell the difference between weight of object A and weight of object B)

Question 3

Method of Constant stimuli

Answer 3

a selected set of stimuli intensities are presented by the researcher

*ideally the threshold falls somewhere within the set of intensities

Question 4

Method of Limits

Answer 4

the researcher presents increasing and decreasing stimuli intensities from a relative maximum and minimum level, and the threshold point is found from the average

Question 5

Method of Adjustment

Answer 5

the participant is allowed to adjust the intensity until they find their own threshold

• this method is quick to conduct, but is vulnerable to participant bias

Question 6

Weber’s Law

Answer 6

there is a relationship between a stimulus’ original intensity and the just noticeable difference to reach the next psychological step; this relationship is a constant

Question 7

Fechner’s Law

Answer 7

–> Continuing from Weber’s law (since Fechner was Weber’s student)

Fechner pointed out the mathematical formula that captures this relationship:
P = k log S

• P = psychological (perceived) brightness
• k = the constant
• S = the physical stimulus intensity (luminance level of a light)

Question 8

Steven’s Law

Answer 8

P = k S^n
* highlighted that the relationship between perceived and physical is not always logarithmic, as Fechner’s law suggested

• the relationship between the physical stimulus and its psychological sensation can be exaggerated, veridical, or compressed –> treat “n” value as a slope to determine which

Question 9

Exaggerated stimuli

Answer 9

n > 1

• can only be predicted by steven’s law
• depicted by a sharp slope
  example: electric shock

Question 10

Compressed stimuli

Answer 10

n < 1

• can be predicted by weber, fechner, and steven’s laws
• depicted by a low slope
  example: brightness

Question 11

Veridical (true) stimuli

Answer 11

n = 1

• 1:1 slope
• only predicted by steven’s law
  example: apparent length

Question 12

Psychophysical scaling and magnitude estimation

Answer 12

From Steven’s Law, we know that not all stimuli and the psychological sensation of them have a 1:1 relationship

*The estimation of the psychological sensation of stimuli is dependent on the stimulus itself (e.g., doubling the physical intensity of a light does not double its perceived brightness - it’s less than doubled!)

–> different “n” slopes to determine perception qualities of a stimulus

Question 13

Signal Detection Theory (SDT)

Answer 13

quantifies the ability to discern between information-bearing patterns (signal) and random patterns that distract from the information (noise)

• -> used to separate sensitivity from response tendencies
• -> helps us understand how we make decisions under conditions of uncertainty, such as near threshold

Question 14

What are the two parts to SDT?

Answer 14

1) d’ (sensitivity index)

2) criterion of observer (response bias)

Question 15

Response bias (aka response criterion)

Answer 15

the black line on an SDT graph aka your response

Question 16

What is to the LEFT of the response criterion line?

Answer 16

Your response NO

Question 17

What is to the RIGHT of the response criterion line?

Answer 17

Your response YES

Question 18

How would you move the response criterion to say YES more often?

Answer 18

you move the line LEFT so that the yes area to the right is bigger

Question 19

How would you move the response criterion to say NO more often?

Answer 19

you move the line RIGHT so that the no area to the left is bigger

Question 20

Describe all the response patterns and where they’d fall on an SDT graph

Answer 20

1) False Alarm: When there’s no signal, and you say “YES” (right of the line)
2) Correct Rejection: When there’s no signal, and you say “NO” (left of the line)
3) Hit: When there’s a signal, and you say “YES” (right of the line)
4) Miss: When there’s a signal, and you say “NO” (left of the line)

Question 21

Sensitivity (d’)

Answer 21

distance between the signal and noise curves; higher sensitivity means it’s easier to distinguish signal from noise

(e.g.)
high sensitivity = fire alarm vs phone ringing
low sensitivity = phone ringing on TV vs your actual phone ringing

Question 22

ROC curve

Answer 22

a visual depiction of various sensitivity levels and its relationship with hit/miss/false alarm/correct rejection

• 0 sensitivity: you have equal chance of being correct as being incorrect; line is 1:1
• higher sensitivity: the curve gets pulled closer and closer to the top left corner, shifting the correct/incorrect likelihoods