exam II Flashcards
reliability
consistency of a measure; increases as number of items/observations increases
2 components of all measures
true score
measurement error
assess reliability with…
pearson r coefficient
test-retest reliability
measuring same individuals at two points
issues with test-retest reliability?
artificially inflated correlation, some variables are meant to change
internal consistency reliability
consistency among items within a measure, uses responses at only one time point
split-half reliability
correlates scores from one half of measure with scores on other half
Spearman-brown split half reliability coefficient (reliability corrected)
cronbach’s alpha reliability
data on individual items!
correlating each item to every other item in the scale
α = average inter-item correlation
item-total correlations
data on individual items!
correlating each item score with the total score
helps eliminate items that are less internally consistent
interrater reliability
extent to which raters agree in their observations
cohen’s kappa
operational definition is key!
construct validity
is the operational definition adequate?
does the test measure what it is supposed to measure?
face validity
measure appears “on the face of it” to measure what it is supposed to
not very sophisticated
content validity
comparing content of measure with reality/definition of construct
predictive validity
does the measure predict future behavior?
concurrent validity
examines relationship between scores on a measure and criterion behavior measured at the same time
convergent validity
scores on the measure correlate well with scores on a another measure of the same construct
discriminant validity
measure is not related to variables in which it should not be related
can discriminate between the measure and other potentially related variables
qualitative approach
observation of behavior in natural setting or descriptions of world/participants
interviews, focus groups, open ended questions
quantitative approach
specific behavior can be counted
statistical analysis
surveys/observations with coding schemes
naturalistic observation issues
ethics of concealment
nonparticipant observer vs participant observer
naturalistic observation limitations
not always appropriate for well-defined hypotheses
population/time/resources/location difficult
systematic observation
observation of several specific behaviors in specific setting
behavior quantified with coding scheme
natural or lab setting
systematic observation coding system
system for rating behaviors of interest, usually for frequency or degree
establish interrater reliability (cohen’s kappa)
systematic observation limitations
equipment, reactivity, long periods of time better for data
case study
detailed description of a single person/small group of persons OR a single setting
H.M.
partial retrograde amnesia
unable to recall recent past
phineas gage
damage to frontal lobe
emotion/personality/goals
types of archival research
statistical records, survey archives, communication records
content analysis
coding system to quantify information in record
survey 3 general types of quesitons
demographics/facts, attitudes/beliefs, behaviors
“yea-saying” or “nay-saying” response sets
respondents employ a response set to agree or disagree with all statements
simplicity
phrase questions to be more simple / therefore understandable
loaded questions
avoid emotionally charged words, insinuating ideas through question or not being specific
negative wording
confusing wording that can sound like opposite question
semantic rating scale
bad _ _ _ _ x _ _ good
labeling response alternatives can…
can influence responding with frequency presented
larger sample size is better for…
generalizing and detecting an effect that actually exists (reducing type II error rate)
probability sampling
each member of the population has a specifiable probability of being chosen
nonprobability sampling
do NOT know probability of any member being chosen
simple random sampling
each member of the population has an equal chance of being selected for the sample
stratified random sampling
population is divided into subgroups (strata)
randomly select members from each stratum
cluster sampling
identify clusters of individuals, clusters are randomly chosen, all individuals in cluster are included in the sample
haphazard “convenience” sampling
selecting a sample from environment in any way that is convenient
purposive sampling
identifying a criterion, conveniently sample from population
quota sampling
sample reflects the numerical composition of the population
sampling frame
the actual population of individuals or clusters that a random sample will be drawn from
continuous measurement scales
interval and ratio
allow for means and standard deviations
measures of central tendency
mean, median, mode
variability measures
range and standard deviation
comparing group percentages measurement
nominal IV, nominal/ordinal DVs
inferential statistic = chi square
correlating individual scores
measurement
interval/ratio IV, interval/ratio DVs
individuals measured on two variables
inferential statistic = pearson r or multiple regression
comparing group means measurement
nominal IV, interval or ratio DV
compare mean response of participants in two or more groups
inferential statistic = t test or ANOVA
correlation coefficient
a statistic that describes how strongly variables are related to one another
pearson r coefficient
-1.00 to 1.00
detects only linear relationship
measure of effect
small v medium v large effect
small (r ≈ |.10|)
medium (r ≈ |.30|)
large (r ≥ |.50|)
r2
percentage of shared variance between two variables
multiple correlation/regression
when you have more than one IV (predictor) predicting a single DV (criterion)
partial correlation
correlation between two variables of interest, with the influence of third variable removed from original correlation
statistically controlling third variable
