PSYC 2018 - Quiz Flashcards
What is a variable?
Any characteristic of the organism, environment, or research situation that can vary
Well defined
ex= age, temperature
What is an abstract
abstract (constructs)
anxiety, depression
Types of Variables
Independent (IV)
Manipulated by researcher and has more than one level
Important IV in psychology is subject variable
can not be changed (logistically or ethically)
Dependent (DV)
Measure of behaviour (what the participant does)
Experimental designs (effect of iv on dv)
Non-experimental designs (association between iv and dv)
Confound
“Confuses” the study
Cannot be separated from the IV ( or accounted for)
when you get your results you don’t know what they mean = cause that you forgot a very important aspect
ex doing a hearing test of a reading test befor hand
What are the four Measuring behaviour
nominal
ordinal
interval
Ration
Nominal Scales
“Name”
Distinct categories (no specific order)
Least amount and least precise (just category)
inal scale study
”Is a woman more likely to give her phone number to a man if the man is accompanied by a dog?” (Gueguen & Coccotti, 2008)
Field study: “Attractive” man with or without dog approaches woman
Phone number given 35% vs 11%
Ordinal scales
Distinct categories in rank order
Can’t quantify magnitude of difference
Interval Scales
Most used in psychology research
Numerical ordered intervals so can compare but no clear 0 (no magnitude comparison)
For example
ex temperature
iq score
anxiety
does not mean absence of intelligence, temperature or anxiety, no such thing as 0
Ratio scales
Numerical ordered intervals with a true 0 so can have ratio comparisons.
For example
number of words recalled
height
reaction time
Mean, median, and mode
Measures of central tendency ( where’s the middle )
Mean
Average
Median
Middle value
Mode
Most common
Evaluating measures
Reliability
consistency of a measure
The relation between reliability and validity (dose it measure what it says it measures)
Validity = reliability
Reliability ≠ validity
Reliability And Validity
Who To Measure
Sample vs Population
E.g., Children with ADHD
Sampling
Taking proportion of population
Must be well-defined and unbiased
eg. student perceptions of university life
Representativeness can determine or limit inferences (consider if sample is biased)
Why Sample?
Economics
Time
Manageability and control
Types of sampling
Non-probability (not random)
Convenience or Haphazard
Quota
Purposive
Read the other two from the text
Snowball
Purposive (random)
Type of convenience sampling
Based on certain characteristic
Typical group of _______
Problems?
Researcher bias
Snowball
aka Network, Chain Referral, Network Sampling
Ask two participants and so on
Useful for small or stigmatized
Problems?
sample basis (lack of generalizability)
ethical issues (confidentiality, freedom to participate)
Probability (random)
Simple random
Systematic
Stratified random
Proportionate Stratified
Cluster
Stratified random
Population divided in subgroups
Equal random from each
Proportionate Stratified
Random from each subgroup based on proportion in population
Cluster
Clusters of individuals or naturally occurring groups
Sampling
The Validity of Research
Internal validity= truth in the study
External validity= truth in real life
Internal validity
Methodologically sound and confound-free
IV and DV
not due to uncontrolled factor
Is DV due to IV? Confounding variable? Both?
low internal validity
Internal validity means ruling out other variables
Threats to Internal Validity
History (Time-related variable)
Maturation (Time-related variable / Participant variable)
Selection Bias (Participant variable)
Attrition (Time-related variable)
Environmental variable- read in textbook
History
Unplanned, uncontrollable event outside the study that affects the DV
More likely in studies that continue over long period
any event that is out of your control that affects your independent variable
sometimes researchers cant do anything about it
Example
Does an anti-smoking campaign reduce smoking?
increase in cigarette tax
find decreased smoking
did anti smoking campaign work?
Example
Program for test anxiety in first year university students
Maturation
Natural processes in participants during study that affects DV \ (development, boredom)
More likely in studies that continue over long period
more experience over time, changes that occur over time with the individual
Example
Program to increase complexity of play between 4 and 9 years
find more complex play in 9 year olds
did program increase complexity
Example
Program for test anxiety in university students
Selection bias
Participants “selected” into groups in unequal way
Groups in a study must be equal in all ways except for IV
Example
Study on physical aggressiveness
group 1= football, rugby, hockey
group 2= chess players
Ulcers in “executive” monkeys (Brady et al., 1958)
“Executive” monkey vs. Control monkey
mild shocks to feet every 20 seconds
both had lever but only “executive” could stop shocks
found “executive” monkey had ulcers but control monkey did not
“Executive” monkey study widely reported
Closer examination of procedure revealed selection bias
“Executive” group = Learned lever press quickly (more sensitive to shock / higher levels of emotion
Weiss (1968) replicated with better controls
found opposite result
“executives” developed fewer ulcers than those with no control
Attrition
Participant mortality
Participant drops out
Problematic if pattern
particular characteristics more in experimental than control
Group left may not be equivalent to group that started out
Example
Weight loss program starts with 50 participants and ends with 30 ( each lost 5 lbs no side effects)
What about the 20 that left
maybe withdrew after losing 25
External Validity
Extent to which research generalizes to other contexts (real life )
Externally valid studies
generalize to other populations
generalize to environments
Other Populations
Why is it important that findings in a study generalize to other groups of people?
eg. different age education level ses
Analysis of participants in all empirical studies 2003 – 2007 (Arnett, 2008)
68% from U.S.
67% of U.S. and 80% of other countries were…?
67% of U.S. and 80% of other countries were…?
But it depends…
When can we generalize?
basic process like prescription memory and attention
When should we question generalizability?
process affected by culture = emotions associated with personal achievement
Other Environments
Generalizing results to different settings
particular criticism for lab studies
But it depends… eg= basic process in memory
What about real life situations?
Ecological validity (Neisser, 1976)
Cognitive psychology
eyewitness testimony, long term learning of school subject
Social psychology
eg= interpersonal attraction using speed dating
Other Times
Generalizing results to other points in time
For example
Studies on conformity by Solomon Asch in 1950s
conservative values dominant
conformity & obedience to authority valued in american society
Conformity
Organization or group (“Go along with it”)
Alter behaviour as result of group pressure
Asch (1950s)
“Perceptual judgment” study
which comparison line matchers standard ?
unknow to participant all others in group are confederates
37% on average 75% at least once
Asch (1955) and others
Social factors that influence conformity
Replication with brain imaging (Berns et al., 2005)
In general
Basic process (e.g., cognition) > Social processes
A Note of Caution about External Validity
Not all research with low external validity is “invalid”
For example
Laboratory studies on false memories (Roediger & McDermott, 1995)
pillow, nap, but not sleep
recall “sleep” not on the list
sometimes people remember with confidence something they did not experience
important and relevant for eyewitness testimony
even through lab production , eyewitness testimony
External validity increases if:
Generalize to other populations
Generalize to other environments
Generalize to other times
BUT … external validity of psychological knowledge not determined by single study
Accumulates over time, with replication and extension
internal validity more critical than external validity
A final note on Internal and External Validity
Goal of any research study is to maximize internal and external validity
But have to be balanced and usually is a trade-off
Purpose or goals help decide which validity more important
Variable
Anything that can be measured and can differ across entities or across time.
* Can be well-defined (e.g., age, temperature) or abstract (constructs; e.g.,
intelligence, excitement, aggression)
Successive Measurements
compare scores from two
successive measurements (test-retest reliability)
Simultaneous Measurements:
measuring via direct
observation – two or more observers simultaneously
record measurements (inter-rater reliability)
Internal Consistency: s
split the items in half and
score each group then look at the agreement between
the two scores (split-half reliability)
Face Validity
does the measurement technique look like it measures the variable that it
claims to measure?
Construct Validity
Based on many research studies that use the same measurement
procedure. It requires that the scores obtained from a measurement procedure behave
exactly the same as the variable itself.
Concurrent Validity:
are the scores from your new measure directly related to the
another, better established, measure?
Predictive Validity
does the measurement of your construct accurately predict
behaviour?
Convergent Validity
two methods of measuring the same construct produce strongly
related scores.
Divergent/Discriminant Validity
differentiate between two different constructs by
measuring both constructs and showing there is little or no relationship between the
two measurements
Scales of Measurement (NOIR)
- Nominal
+ Distinct categories without specific order (e.g., academic major) - Ordinal
+ Distinct categories in rank order (e.g., first place, second place, third place in
a race)
+ No consistent difference or cannot determine the magnitude of the difference
between categories - Interval
+ Numerically ordered intervals but no true 0 (e.g., temperature in degrees Celsius)
+ Can’t make ratios statements – can’t say that 40℃ is twice as hot as 20℃ - Ratio
+ Numerically ordered intervals with true 0 (e.g., temperate in degrees Kelvin)
Non-Probability Sampling
Convenience or haphazard sampling
* Available, meet some of the characteristics, recruited non-randomly
* E.g., an existing subject pool; mall surveys
Purposive sampling
* Specifically recruited
Quota sampling
* Trying to reach a specific proportion (e.g., 60% female) but not randomly
seeking out participants to match criteria (based on convenience)
Snowball sampling
* Based on referrals, networks of friends, other students
+ Correlational
describe relationships
Experimental
explain relationships
explain relationships
describe relationships (with an attempt to explain)
Nonexperimental
- describe relationships