Stats Final Exam Flashcards
what is an example of a continuous variable
age, weight, height
what does correlation describe
how variables will co-vary; whether they behave similarly or differently
what symbol expressions correlation
pearsons r
what r value indicates no relationship between X and Y
0
what is X
the predictor
what is Y
the outcome
what r values are perfectly correlated
1 or -1
what is the expected correlation between variables
0
r <= .10 is
trivial
r < .30 is
small
r< .50 is
medium
r >= .50 is
large
what is a linear relationship
the best way to summarize the trend in the data is with a straight line
what is a quadratic relationship
the best way to summarize data is with a curved line
what are the 3 elements of causality
- correlation
- temporal precedence
- ruling out alternative explanations
the only reliable method to determine causality is with…
experimentation
most correlations are not implying…
causality
what does OLS stand for
ordinary least-squared regression
what is OLS
the smallest distance between Y and Y(hat)
what is Y(hat)
predicted value of the constant when x has certain qualities
what does it mean for Y(hat) when a predictor is non-significant
that Y(hat) doesnt change as the X value changes
what does it mean for Y(hat) when a predictor is significant
that Y(hat) changes as X value changes
what do you do when you’re asked to estimate an out-of-range value
decline! point out the model (OLS) does not allow this
what is NOIR with examples
Nominal (gender)
Ordinal (military rank)
Interval (temperature C)
Ratio (weight, age, height)
continuous data relies on what
the mean
what does parametric mean
they have many underlying assumptions
what does non-parametric mean
they have fewer underlying assumptions
is chi-squared parametric or non-parametric
non-parametric
if data is continuous, what will likely make an appearance
t-tests
what is the basic idea of t-tests
difference/error
chi-squared tests are a calculation of
what we expected the answer to be, and what the answer was
what is needed for an average
mean and standard deviation
what does categorical data not have
the necessary elements of an average (mean and SD)
categorical data can not be…
averaged!
what are 2 reasons you may use categorical data
- high variance implies different issues, and converting to categories solves this problem
- sometimes group size is too small to facilitate a normal analysis, so we collapse data
you cannot go from categorical data to…
continuous data
what are the two flavours of chi-square tests
- goodness of fit
- tests of independence
what is goodness of fit
is if data from 1 variable meets expected proportions
what is tests of independence
describe if there is a non-random association between 2 variables
what are 2 basic assumptions of chi-square
- independence of observations
- size of expected frequencies
what is independence of observations (chi-sqaure)
persons can only give data for one cell (ex. name) theres only one answer
what are some examples of GOF tests
- are men/women equally likely to be lawyers
- are Canadian university professors more likely to be a non-minority
- are left-handed people more likely to be divorced
what must be known for GOF
only makes sense is proportions are known; need a baseline
what are all GOF tests based on
general population
what is fe
frequency of expected
tests for independence uses….
frequency data to evaluate the relationship between two variables
what does effect size describe
the magnitude of difference between groups; extent to which the null hypothesis is incorrect
what is cramer’s V
a statistic to calculate the effect size of chi-square
what type of effect size to be want
large
what changes the standards for small, medium, and large effect sizes
degrees of freedom
