COGSCI 151 Midterm Exam #1 WITH CORRECT
ACTUAL QUESTIONS AND CORRECTLY WELL
DEFINED ANSWERS LATEST ALREADY GRADED
A+
Save
Practice questions for this set
Learn 1 /7 Study with Learn
Terms in this set (127)
signal detection theory models how people decide whether a signal is
present, given a noisy stimulus → decision
making under uncertainty
what trade-off does signal false alarms (detecting something that isn't
detection theory formalize actually there) → false positive
misses (not detecting something that is there) →
false negative
, perceptual component how well can you distinguish signals from noise
judgement component how sure do you have to be before you report a
signal
two sources of SDT variability sensitivity (d') → the quality of the internal
evidence, how well signals are separated from
noise
decision criterion (c) → the bias/threshold for
saying yes and making a decision
measuring sensitivity with d' z(Hit Rate) - z(False Alarm Rate)
measures how well the observer can discriminate
signal from noise
what does a larger d' mean greater separation between signal and noise
distributions (i.e. better at distinguishing signals)
measuring bias with c c = -0.5 x [z(Hit Rate) + z(False Alarm Rate)]
measures the observer's decision threshold
sensitivity on graphs high sensitivity means wider graphs, low
sensitivity means lower graphs
moving criterion to the right on a reduces hits and reduces false alarms
graph
ACTUAL QUESTIONS AND CORRECTLY WELL
DEFINED ANSWERS LATEST ALREADY GRADED
A+
Save
Practice questions for this set
Learn 1 /7 Study with Learn
Terms in this set (127)
signal detection theory models how people decide whether a signal is
present, given a noisy stimulus → decision
making under uncertainty
what trade-off does signal false alarms (detecting something that isn't
detection theory formalize actually there) → false positive
misses (not detecting something that is there) →
false negative
, perceptual component how well can you distinguish signals from noise
judgement component how sure do you have to be before you report a
signal
two sources of SDT variability sensitivity (d') → the quality of the internal
evidence, how well signals are separated from
noise
decision criterion (c) → the bias/threshold for
saying yes and making a decision
measuring sensitivity with d' z(Hit Rate) - z(False Alarm Rate)
measures how well the observer can discriminate
signal from noise
what does a larger d' mean greater separation between signal and noise
distributions (i.e. better at distinguishing signals)
measuring bias with c c = -0.5 x [z(Hit Rate) + z(False Alarm Rate)]
measures the observer's decision threshold
sensitivity on graphs high sensitivity means wider graphs, low
sensitivity means lower graphs
moving criterion to the right on a reduces hits and reduces false alarms
graph
