Chapter 8 - Bivariate Correlational Research

Question 1

What do scatterplots display?

Answer 1

form - linear or non linear
direction - positive or negative
strength - none, weak or strong (closer the points, the stronger)

Question 2

define dichotomous data

Answer 2

data that only has 2 responses, ex. yes or no

can be best represented with a bar graph over a scatterplot. use error bars to show ow much variation there is in your sample. but different groups do not infer causality, only can infer when there is manipulation of the groups!

Question 3

how do we analyze assocation when one variable is categorical?

Answer 3

use r or a t-test (measuring difference between means)

Question 4

how to judge the construct validity of association claims?

Answer 4

look at both variables and ask about operationalization,s acle reliability and validity

Question 5

How to judge the statistical validity of an association claim?

Answer 5

• effect size along with the accuracy of the predictions ( large is not always the best)
• confidence interval or p-values (range in which we expect the data to fall in, who precise out point estimate is, don’t want to include 0 if you want there to be an effect, easier to get significant effect with large sample)
• replication (confounds can influence association, so need to replicate to rule them out)
• outliers (extreme outside variables, want to remove any before calculating the association)
• range restriction (sometimes have to restrict data to ensure you are looking at the proper population)
• curvilinear (sometimes data just will not be linear, quadratic)

Question 6

define bivariate correlation

Answer 6

an association that involves exactly two variables

Question 7

when is it best to use a bar graph over a scatterplot?

Answer 7

when one of the variables is categorical

Question 8

defien effect size

Answer 8

the magnitude/strenght of a relationship between two or more varaibles

Question 9

define statisically significant

Answer 9

when p < 0.05, that is that its unlikely the results came from the null-hypothesis

Question 10

define directionality problem

Answer 10

in a correlational study, the occurance of both variables being measured around the sam time, making it unclear which variables in the association can first

Question 11

define spurious assocation

Answer 11

a bivriate assocation that is attributatble only to systematic mean differences on subgroups within the sample, the original association is not present within the subgroups

Question 12

define a moderator

Answer 12

a variable that, depending on its level, changes the relationship between two other variables