Collecting Data / Data Measurements Flashcards
variability
- confounding variables
moderating variables
- influence IV and DV relationship
- cold meds fatigue and alcohol example
mediating variable
- correlation NOT causation
- fire rates and ice cream sales (there is a third mediating variable making these appear like they are causing each other but they are not)
rating scales
very poor poor okay good very good
Likert scale
strongly disagree disagree neutral agree strongly agree
semantic differentiation
ranking easy to hard
1-7 scale
accuracy
how accurately a participant performs a task
frequency
how often do we observe a behavior
latency
how long to develop a behavior
duration
how long does a behavior last
amplitude
how strongly is behavior exhibited
levels of measurement
nominal
ordinal
interval
ratio
nominal
name
not quantifiable
(gender, occupation, eye colour)
ordinal
order
rankings
interval
distance is meaningful
LSAT scores
ratio
has to have an absolute zero
income, mac grading scale
is a reliable measure always correct?
no
how to establish reliability
test re-test
split-half reliability
inter-rater reliability
test re-test
collect measures from participants at different times
split-half reliability
in a test with n similar questions, take score of a person from the first half and see how it correlates with the scoree of another person from the second half
inter-rater reliability
how similar are participants rating from one question to another
validity
measure what it says it measures
content validity
the degree to which it measures all dimensions of the construct
face validity
on the face of it, the measure seems valid
construct validity
does our measure tap into theoretical concepts that it is supposed to measure
criterion validity
does measure correlate well with other measures of some behavior
central tendency
mean median and mode
mean
average
median
middle
mode
occurs the most
range
min-max
measures of dispersion
- calculate the distance between each data point and the mean
- take the square of all of them
- take the sum of the squares and divide by (n-1)
scales
measures of the specific psychological characteristic
inventories
measure things like personality
research skills
- use catalog search, not discovery
- a more generic search for a book
- the more specific search for the article
- use filters
- check credentials of article
Linguistics and Language Behaviour Behaviour Abstracts
- primarily abstracts
- has good filters
Web of Science
- broader than LLBA
- a lot of citations and abstracts
Google Scholar
- broadest
- natural language search
- advanced search bar
- links to mac library
Purdue owl
APA owl
external validity
replicating a result in different settings with different populations
ecological validity
how natural and generalizable is the task, testing enviroment etc.
internal validity
did the independent variable cause the changes ew observe in the dependent variable
controlling the extraneous variables
- elimination
- constancy
- secondary variable as IV
- randomization
- repeated measure
- statistical control
problems with internal validity (7)
- participant reactivity
- demand characteristics
- participant sophistication
- social desirability
- experimenter bias
- range restriction
- order effects
participant reactivity
act differently knowing they are being observed
positive results due to “good subject” effect
negative results due to “bad subject” effect
solution: disguise hypothesis
demand characteristics
when paticipant is told how to behave
solution: hiding true topic
participant sophistication
because of previous knowledge participant may come to know the hypothesis
social desirability
act as they feel they should to look good
experimenter bias
when experimenter influences the results
solution: blind experiment, script interactions
range restriction
when not all of measurement scale is used
ceiling/floor effect
order effects
when one part of the study affects a later part of the study
practice and carry-over effects
problems with external validity (1)
self-selection
self-selection
when people volunteer themselves for an experiment
probability
specific probability a participant will end up in your sample
non-probability
it is impossible to say how likely anyone participant will end up in your sample
random sampling
every person in the sampling frame has an equal chance of being in the sample
systematic sampling
pick every nth person on the list
stratified sampling
say 20% of the population has X and we want 100 participants we would select 20 participants randomly from that 20% that has X
cluster sampling
randomly selecta small number of clusters and study all of the workers in those clusters
multistage sampling
i.e. select a school district, a school, a class
non-probability sampling
convenience, quota, referral
convenience sampling
test whoever is available
quota sampling
pick people from subgroups and include them in the order you approach them
referral sampling
have participant refer you to someone they know who is also part of the population you are studying
