LECTURE – Chapter 1
Psychophysics: the study of quantitative relationships between physical events and psychological experiences.
Making the link between physical events and psychological experiences.
Abs thresh: detection threshold. Minimum physical intensity required to detect stimulus.
How do we measure abs thresh?
Method of adjustment: observer adjusts the stimulus until he can barely detect the stimulus. Not really
precise. Depends on internal factors sometimes.
Method of limits: Stimuli of diff intensities presented in ascending and descending order. Keeps saying yes
until he says no (for descending order). Starts with one that is clearly perceptible and then go down. Vice
versa. Go back up. Go back down. Repeat a few times then take average. Problem is subject knows that the
order is ascending/descending.
Method of constant stimuli: stimuli of diff intensities presented in random order. Multiple times (same
amount of times per intensity though). Graph it into a psychometric function. Functions aren’t step functions
but s-shaped because we are not ideal detectors. Abs thresh is usually the when he detects it 50% of the time.
All these methods are subject to observer bias. Some are more willing to say yes (usually young observers),
some are less willing and wait until they’re absolutely sure (older subjects usually)
One way to avoid this is to give a “forced choice”. You have certain answers already presented and tell them
to choose one (like MC)
Signal Detection Theory: measures sensitivity as well as your criterion level (how much your biased more to
say yes/no). Collection of psychophysical procedure and analytical techniques for diverse experimental tasks.
Why do we use signal detection theory instead of just forced choice? Sometimes we want to measure their
level of bias.
SDT has a signal (stimuli) not just light, sound, tec. But new faces within old faces, words that are
grammatically correct, etc. Always detected against background activity (noise) -> sometimes you say yes,
sometimes no. Variability can come from outside/internal factors. Your neurons also don’t fire exactly the
same everytime. This is called noise.
In SDT, there are sometimes catch trials (where you don’t actually present anything and you ask the subject if
there was a stimulus or not).
If you present a stimulus and he says yes, that’s called a hit.
If you present a stimulus and he says no, that’s called a miss.
When the stimulus is not presented but the observer answers yes, that’s called a false alarm.
When no stimulus is presented and the subject answers no, that’s called a correct rejection. There are many trials. One question and one answer = one trial.
The final step is to transform raw data into response rates (calculate the proportion of hits, misses, etc.)
By knowing the proportion of hits/misses and correct rejections/false alarms, we determine the criterion of
The sensitivity is the difference between the two distributions.
Metric d’ (d-prime) – distance btwn 2 distributions. The sensitivity is the difference.
D-prime: low when the two conditions overlap a lot. The more sensitive you are, the higher the d-prime (two
distributions that are really far). When d-prime is zero it means you can’t dissociate btwn sensation and no
If we didn’t have catch trials, someone who was super willing to say yes would be calculated as having