belief bias
we are more likely to believe an argument with incorrect reasoning it has a logical claim
Simpson’s Paradox
the results of 1 trial may be different than the overall data
base rate fallacy
you have to take the number of people in each category within the entire population into account
construct
concept that cannot be directly observed
measure
method used to gather data
variable
results
operalization
turning a construct into a measurement
steps to operalization
being precise about what you are measuring, choose an appropriate method, define variables
nominal
discrete categories
ordinal
ranked categories
interval
ranking with known intervals
ratio
ranking with known intervals and a meaningful 0
continuous
for any data point there can be an intermediate value
continuous measurements are
interval and ratio
discrete
there are 2 data points without an intermediate value in between them
discrete measurements are
nominal, ordinal, interval, ratio
Likert scale
measuring preferences on a scale
test-retest reliability
getting similar results across multiple repeats of an experiment
interrater reliability
2 different people get consistent measurements
internal consistency
do different parts of the measurement that perform the same function give similar answers
predictor variable
doing the explaining
outcome variable
being explained
internal validity
extent to which you can draw causal relationships between variables
external validity
extent to which you would see the same results in the general population
construct validity
are you measuring what you think you’re measuring
face validity
do the questions appear to be measuring that they are intended for
reliability
how consistent scores are
validity
how accurate the data are
mean
average
median
middle value
mode
number that occurs most often
positive skew
fewer values at high numbers, median closer to 25th percentile, mean is less than median
negative skew
more values at high numbers, median closer to 75th percentile, mean is greater than median
kurtosis
how pointy a graph is
range
greatest value-smallest value
IQR
75th percentile-25th percentile
mean standard deviation
degrees away from the mean
variance
measure that takes outliers into account and magnifies them
z-score
standardizing data to compare two unrelated constructs and interpret results in terms of standard deviation