Task 1 Flashcards
Cognitive psychology
Study of mental activity as an info processing problem
- Assumption:
We don’t directly perceive & act in the world
-> Our perception, thoughts, actions depend on internal transformations or
computations
Tasks are composed of set of mental operations
-> Which involve taking representation as an input and performing some sort of process on it
-> Thus producing new representation or output
2 key underlying concepts
- Info processing depends on internal representations
- Mental representations undergo transformations
Example:
You first smell garlic bread (sensation) → You think about holiday in Italy at your grandmother’s house (perception) → You pick up the bread and eat it (action)
Reaction Time
The time between stimulus presentation and the response
Operation def. :
Participant intends to be as fast as possible w/out making errors
- Trials are repeated several dozen times
-> To get reliable mean RT (mRT)
-Observed average RT should gain validity as a measure of what we aim it to reflect
-> The minimum amount of time needed by participant to produce correct response
Theoretical assumption:
RT is minimum amount of time needed by pp to produce correct response
-> PP may try go be faster but will then find out that errors occur more often
-> PP can also be perfectly accurate if they take much longer time
- Instruction & training are to avoid that
Speed accuracy tradeoff
Tendency, when performing task, for either speed or accuracy to be sacrificed in order to prioritize the other
- The faster → the less accurate
-effects on mean RT should always be inspected in context of accuracy Neg. relationship btw RT & accuracy
->One way to investigate this: rune same task multiple times & give diff. speed-accuracy instruction every time
Outliers with RT
-Inattentiveness can cause large RT
- That skew the mean RT, standard deviation and variance
- Outliers can also be extremely fast responses
Dealing with outliers
Medians: The median is “robust measure of central tendency”
- Meaning it is relatively insensitive to outliers
Trimmed means:
-Middle nr. of trials
- Thus: it deletes upper & lower 50% of distribution
- One may delete smaller proportion (e.g. fastest & slowest 10%) & then take mean of remaining ones
Standard deviation:
- Compute mean & SD for each partic. & condition
- Then delete RTs that deviate more than some nr ‘C’ of SDs from mean
Fixed Criterion
-Delete all RTs exceeding some value and compute mean of rest
Doing nothing
-Own choice whether to remove or not & whether it would make sense
Instructions to participants
-Should be consistent
- Give feedback
- Experiments shouldn’t be too long
Subtraction method
- Construct 2 tasks that differ ony in single component
- Measure RT in both taks
- Subtract RT
- will give duration of single differing component
-Subtracting RT as RT(C)-RT(B) = duration of response selection
-Subtraction RT as RT(B)-RT(A) = duration of stimulus discrimination
Donders´3 tasks
-Task A: Simple RT
-> 1 possible stimulus and 1 possible response
-> sensory time + motor time
-Task B: 2-choice RT
-> 2 possible stimuli; each demand different response
-> sensory time + discrimination time + response selection time + motor time
-Task C: Go-nogo task/ disjunctive RT
-> 2 stimuli and 1 possible response
-> sensory time + discrimination time + motor time
Pure insertion
(assumption of Donder´s 3 task)
-Duration of all processing stages remains same when extra stage is added or removed
-Assumption implies no interactions among the cognitive components of a task
Seriality
(assumption of Donder´s 3 task)
Processing stages are carried out in a strictly serial manner
Only in the case that total RT is equal to sum of durations of individual stages
Aim of Sternberg´s additive factor method
- Identify whether stages are serial or parallel
- Controlling for limitations of Donder’s subtraction method
- Discover processing stages
Interaction
-Effect of one factor modulates the effect of another
-Simple effect of factor A varies across levels of factor B
Diffusion model
-Provide theory that explains distribution of RTs of both correct responses & errors in 2- choice RT task
Basic assumption:
-Info accumulates continuously during time btw stimulus onset & response Accumulated info is represented by internal counter
- Counter is driven in opposite directions
- By tiny bits of info → Supporting the different outcomes (correct vs. incorrect)
- Over time: counter runs in a corridor btw 2 thresholds - As soon as upper or lower threshold is hit
→ decision has been made & response A or B is executed
Random walk
The whole process of various trials can be simulated to obtain simulated RT distributions
- Can be tuned in a way that the simulated RT mimics the actually obtained RT
distributions closely
➔ If the match is close: Assumptions of the model are valid & parameters of diffusion model are
psychologically meaningful
Sternberg´s additive factor model
In each trial, a set of letters is presented and the participant has to memorize them
- This set can contain 1-6 letters
After these letters have been shown, a single ‘probe’ letter is presented and the participant has to say whether it was shown in the set of letters before or not → then he has to press ‘yes’ or ‘no’
=>RT increases linearly with the number of items held in memory
4 mental processes
1. Encode: the participant must identify the visible target
2. Compare: the participant must compare the mental representation of the target with
the representation of the the items in memory
3. Decide: the participant must decide whether the target matches one of the memorized
items
4. Respond: the participant must respond appropriately for the decision made in step 3
-> Factors that influenced one stage (e.g. stimulus intensity) would be additive with those that influenced another stage (e.g. motor output)
Sternberg´additive factor model assumption
- Seriality, but not pure insertion
- If we use an experimental manipulation, the duration of one stage can be increased,
but the quality of the stage will stay the same
RT is simply the sum of the durations of all stages
- If two manipulations are administered and they don’t affect each other → additive
effects (Additive effect means, that they affect two separate stages)
- If two manipulations modify each other’s effect, they must affect a stage in common
→ interaction
