Methods in Experimental Psychology Flashcards
A word that means there’s always simple and a complex explanation in science
Parsimonious
What does a lot of bad science start with?
The complex explanation of the parsimonious model
bearded man followed by ducks
Imprinting
Ethologists like to explain behaviour with instincts
He made the point you have aggressive behaviour and that’s an instinct and that instinct causes aggression, and that demonstrates the aggressive instinct. Done.
Konrad lorenz
method of ____: Don’t like, based on beliefs basically, and you won’t win arguments if you ask someone about it because it’s an argument about their beliefs
method of tenacity
What did Popper say about theories?
If you want to put a theory down as bad, you have to be able to falsify it with evidence
what theory wins between two competing ones?
Simpler one wins and theory with most evidence to back it up
- Richer countries focused on this before the 1980s, but now applied gets more funding
- wanted to know the basic understandings about certain things in the world. Which good applied research is then based on
fundamental research
- Used to be used mostly by poor countries/ countries that underfunded research
- Richer countries focus on this over fundamental today
- solving problem without any real theory
applied research
what type of research?
- correlational (links b/w variables, can help generate hypothesis)
- descriptive and observational
- Case histories/ studies
- Surveys , interviews- content analysis, meta-analysis
non-experimental
_____ approach to science
Top-down, theory driven
“Theory testing”
deductive
____ approach to science
Bottom up, Observation driven
“Theory building”
inductive
_____ research that uses a large number of subjects
nomothetic
______ research that uses a single case/ subject and is then repeated to test for validity
Idiographic
_____ research
- starts with theory, then addresses the theory
theoretical
_____ research
more common in applied sciences (don’t care as much about theory but trying to solve a problem
- might have to go back to test theories
A- theoretical (data-driven)
______
- science based on historical data
- how things change over time
diachronic
______
- science based on here and now
Synchronic
Is qualitative data or quantitative data more dependable/ scientifically valid?
Quantitative data
Explain Small n research
Why large number of subjects is good, why small number is good
A lot of scientists say if you really want a good experiment, more subjects the better
Principle behind this? → your sample should be representative of population, larger sample, better representative
Problem with this: produces science to an average
Comparing average of control, with average of experimental group
Participants contribute to average, but individuals are lost
Maybe sometimes we should focus on individual case more
Gordon - we study personality, and tests, based on data of people in groups of 1000 or more. Study the person.
Why don’t outliers do what the general pop do.
Type of research:
count steps, words, Donald trump is prime subject(vocab limit) Maybe he
has dementia, as vocab since 90s has narrowed
content analysis
Type of research:
Vitamin d - heart health
Correlational data by itself to see that maybe a correlation, then test in an experimental setting
Correlational
Type of research ______
- take bunch of studies answering same question, put together, make a large
study from all independent studies. See patterns emerging
St john’s wort, better than placebo/ some antidepressants
Did meta-analysis - indeed effect is there
meta-analysis
Type of research____
used in hospital a lot
Some is good, quantitative, experimental
But some is not.
H.M. amnesia case. Phineas gage - rod in brain -
Rare cases especially are useful, bc you get from this case you wouldn’t otherwise bc it’s rare and or unethical to promote
- surveys and interviews
case studies
non-experimental research, ______: non-intrusive, make sure animals not aware researcher is there.
Naturalistic observations:
Start studying animals, tribes, live with for awhile and hope they forget about you, think of you as part of your group
Problem with this approach, it’s hard to argue your presence as an outsider is not going to influence behaviour. But it will still probably get interesting data
participant observation
what is ecological validity
an experiment has ecological validity if it reflect real-life situations or the data that would be obtained in real life settings
Define the “detection” part of psychophysics
s it there or not; a fluke or not; disease or not
And there are different levels of detection with this
Define the “discrimmination” part of psychophysics
Comparing something; choice or many different choice and you make a decision based on these choices
Define the “identification” part of psychophysics
requires a name for stimulus or stimuli
commiting to saying you understand what it is
define the “scale” part of psychophysics
How much of it”; gradients
SDT will apply to the first 3 processing types
Tell me beyond a threshold if its there or not. We all know its there, but how much of it is there is what i’m asking
To be clear on what the threshold is
What are the 4 basic elements of signal detection theory (SDT)
- signals
- response
- noise
- response bias
describe the “signals” part of SDT?
signals coming from the stimulus; the object; the target
describe the “response” part of SDT?
how do you know something is going on; the action or decision made
describe the “noise” part of SDT?
he “uncertainty” factor, the interference (intrinsic or extrinsic)
describe the “response bias” part of SDT?
the bias from the decision maker, the responder (receiver)
Can be “liberal or conservative”
Describe the evolution of Group-selection theory
thesis (early years): group selection theory
antithesis (later years): kin selection theory
synthesis (modern) : multi-level selection
What is the method of intuition based on?
what is it used for?
a hunch or gut feeling
sometimes used to create a hypothesis
What is the method of faith based on?
common beliefs "everyone knows that" from authority (not always experts) and we don't question it
What is the method of authority
looking at research vs doing research
What is the rational method?
What is it useful for?
using logical reasoning for thought experiments
essential in planning of research designs and useful on a scientific panel
what is the main method of verification of theories?
falsibility
Who died with hid theory that dogs didnt come from wolves?
Kuhn
When is small sample sized research important
when studying differences in people, we don’t wnt a group average
data that sparks interest in a question by having outliers that may or may not be flukes
ancedotal data
the ability to reflect real-life situations or data that would eb obtained in real-life settings
ecological validility
abductive appraoche where you use prior konwledge to make an educated guess. Often used by family practitioners
bayesian approach
accuracy is measured on ROC curve by ____
d’
observers bias is measured on the ROC curve by the ____
criterion or c or beta
when d ‘ is ___ it is harder to detect or discrimminate
when d’ is small
When is SDT and ROC curves irrelevent
when the curves of noise and stimulus do not overlap
a person who says yes is ___
liberal
a person who says no easily is ______
conservative
d’ equation for detectability
d’ = Z(hits) - Z (false alarms)
d’ equation for discrimmination
d’ = (Z (hit) - Z (flase alarms))/ 2^0.5`
what are discrimmination curves on ROC
stimulus 1 vs stimulus 2
what are the detection curves on ROC
noise vs (noise + simulus)
the correct rejection rate (true negative rate TNR)
specificity
the hit rate (True positive rate (TPR)
sensitivity
Type 1 errors are also called
false alarms
Type 2 errors are also called
misses
low type 1 error indicates
high specificity
high type 1 error indicates
low specificity
low type 2 errors indicate
high sensitvity
high type 2 errors indicate
low sensitivity
validility = constant error = ____
accuracy
validility = variable error
precision
repeatability vs reproducibility
repeatability= same conditions and short time range reproducability = different conditinos same outcome over long time range
symptoms and diagnositic criteria ___ validility
content
correlatinos and reponse to treatment ____ validility
concurrent
diagnositic stability over time ____vailidility
predictive
can it discrimminate between disorders ____ validility
discrimminate validility
the validility of a diagnosis
diagnostic validility
association between a predictor and an outcome variabe is an _______ study
oberservational
detection/ discrimmination between “yes” or “no” –> more than association, is a ____ study
diagnostic
accuracy is a balance of ___ and ___
sensitivity and specificity
while diagnosing:
if prevalance is low tests need to be _____
specific
while diagnosing
if prevalance is high tests need to be ______
sensitive
while diagnosing what is prior probability
the base rate, the probability of this person getting a disease based on their background
while diagnosing what is the predictive value
the liklihood of a positive or negative tests
the act of giving a misleading impression
a distortion
what is the distortion rule
procedures used to make observations should not introduce distortions
what can distortions come from
insturments
oberserver/experimenters
sampling procedures
environment
a set of laws/traditions that form a scientific tradition
paradigm
a collection of hypotheses about a specific a
theory
a specific implementation of a theory
a model
a general “fact” that is accepted; not always tested
principle
a generally accepted process or pattern
a rule
substantially verified theory
a law
a testable statement about the statement about the relationship between variables
a hypothesis
the steps fo defining a theory
- defining the scope
- reviewing the literature
- formulating the theory
- establishing predictive validility
- testing the theory empirically
variables that cannot be observed directly and not observed initially.
inferred from the data as a link between other variables
intervening variables
describe strong inferences
developing several alternative theories and testing all prefictions; works well in highly controlled studies
describe severe experimental testing
error detection and correction
progress from the identification of errors
ostensive definitions
phenomena to be obersved shuld be carefully describes (graphically, photographically, ect) and examples given
rate the scales of measurement from least information yielded to most information yielded
nominal –> ordinal –> interval –> ratio
what are the quantitative scales of measurement? What are the qualitative scales of measurement?
quantitative: interval and ratio scales
qualitative: nominal and ordinal scales
what are the differences between the interval and ratio scales?
the interval scales is less mathematical, has no absolute zero
the ratio scale has an absolute zero and 70= 0.5 x 140
what is a quantitative independent variable
a treatment that differs in frequency, amount, or degree
what is a qualitative independent variable?
the different kinds of treatment
what is a quantitative dependent variable
the score or duration observed
what is a qualitative dependent variable
the special procedure needed as a response observed
a threatening variable with obscuring factors that has an impact on the dependent variable; you have no control over this. We assume it will not impact data
an extraneous variable
an extraneous variable that can inadvertently affect another experimental variable. this WILL impact data
a confounding variable
what are demand characteristics
the participants own experiences changing their behaviour during an experiement
what types of subjects do researchers want
faithful subjects
what is a nocebo effect
the opposite of placebo (treatment has positive effect based on participants positive expectations), treatment has more negative effect because of the expectations of the treatment
what is placebo effect
treatment has positive effect based on participants positive expectations
what is the unrelated-experiment technique
same participants, two studies that the participants do not believe are related, where we give the independent variable to the subjects in one experiment, and give the dependent variable to subjects in another experiment
explain experimental-expectancy control groups
3 groups of researchers lead 3 different groups into believing three different results of the independent variable, one with the experimental outcome, one with another experimental outcome, one with no effect, to determine the effect of subject expectation
the ideal method of sampling
random sampling
what is stratified sampling? random or non-random?
random; sample within segments of the population ex: age, gender, ect
what is clustered sampling? random or non-random?
random; sampling from naturally occurring units of individuals ex: students in a class
what is proportionate sampling? random or non-random?
random; sampling that avoids over-representation in one area or segment of the population
what is haphazard sampling? random or non-random?
non-random; might look random but isn’t ex: grabbing fish from a fish tank
what is convinence sampling? random or non-random?
non-random; drawing samples from population close to you
what is volunteer-sampling? random or non-random?
non-random; sampling that is a self-selection process; participants are more likely more outgoing or possess common traits that made them volunteer
what is systematic sampling? random or non-random?
non-random, every nth element chosen, not random
what is sequential sampling? random or non-random?
non-random; gradual, one at a time selection; might run out of good particpants quicker and change requirements of subjects
what is quota sampling? random or non-random?
non-random; stratified sampling without randomness
what is purposful/selective sampling? random or non-random?
non-random; based on predetermined criteria ex: exceptional individuals ex: dogs
random sampling vs random assignment
random sampling: randomly selecting subjects for a population to put into samples Random assignement is where subjects in a sample are then randomly assigened a condition
