Chapter 8 Flashcards
Two variables are said to be associated when
they vary together
when one changes the other changes
2 things to do when the independent variable is it the columns
calculate percentages for each group
compare the percentages horizontally
3 characteristics of a bivariate association
1.) does an association exist?
2.)If an association exists: How strong is the association?
3.)What is the pattern or direction of the association?
the stronger the relationship…
the greater the change in conditional distributions
example of no association
example of a perfect association
How to calculate maximum difference
largest # - smallest # in a row
maximum difference for weak moderate and strong (used for Phi and V)
weak= 0-10
moderate=11-30
strong= more than 30
In positive relationships and EX
the variables vary in the same direction
as job satisfaction increases so does productivity
in negative relationships and EX
the variables vary in opposite directions
as one increases the other decreases
education decreases TV viewing increases
To examine associations in bivariate tables, follow the rule:
percentage DOWN
compare ACROSS
measures of association characterize the
strength of bivariate relationships
for nominal level variables there are 2 common measures of association (3)
chi square-base (phi or cramers V)
PRE: measure Lambda
What does phi do and what does it use
judges the strength of the relationship
2x2 tables only
What does cramers V do and what does it use
fixes the denominator problem by adjusting for tables size
uses anything over 2x2
Phi and V will or will not be equal
will be equal
range association for Phi and V
0= no association
1= perfect association
PRE
proportional reduction in error
PRE prediction 1
predicting the core of the dependent with no information from the independent
PRE prediction 2
predicting the score of the dependent with information from the independent
Lambda tells us the
improvement in predicting Y while taking X into account
What is E1
Prediction 1
What is E2
Prediction 2
What is lambda
Difference between E1 and E2
If the variables are associated we should make fewer errors using which prediction
prediction 2 should have less errors
Lambda gives the indication of the
strength of the relationship
lambda is asymmetrical meaning
the value will vary depending on which variable is the independent
What does lambda do that Phi and V don’t
predicts the proportional reduction in error
