Correlation Coefficients

Question 1

XY Scatterplot

Answer 1

When both variables are continuous use scatterplot; predictor on x and outcome on y

Question 2

Trendline/ Line of Best Fit

Answer 2

Line that minimizes the sum of distances from each data point to the line; what straight line “fits” the data “best”; doesn’t matter the order participants are in

Question 3

Correlation Coefficient (r)

Answer 3

Quantifies the relationship between two continuous variables; small r=no evidence that variables are related to each other; large r= variables are likely to be related to each other; a small or large r depends on sample size (r<0.2=small; r>0.2=large)

Question 4

r

Answer 4

a number between -1 and +1

Question 5

Nature of relationship

Answer 5

Positive correlation: increase and decrease together (r>0)
Negative (inverse) correlation: increase in one is associated with decrease in other (r<0)
No relationship (r=0)

Question 6

Strength of relationship

Answer 6

How closely the points cluster around line of best fit; higher absolute value (close to +1 or -1) indicates higher strength; absolute value of r -> the greater the absolute value, the stronger the relationship; the spread of data around line of best fit

Question 7

P-value

Answer 7

A number between 0-1; the probability of finding the r from our sample if there is really no relationship between the variables; larger r=smaller p-value; large inferential stat. indicates that it’s likely there’s a real relationship; if p-value is small, it’s unlikely relationship is due to chance

Question 8

Large sample size

Answer 8

Allows you to detect a relationship of around 0.2; will have more power and be able to detect a weaker relationship

Question 9

Values of R

Answer 9

|r|<0.1 Strength= none
0.1<|r|<0.3 weak
0.3<|r|<0.5 moderate
0.5<|r|strong

Question 10

Not all relationships are linear

Answer 10

A correlation coefficient assumes a linear relationship between the variables; it doesn’t measure other kinds of relationships; data result in same value of r

Question 11

Types of Relationships

Answer 11

Associations between predictors and outcomes might be confounded by group differences; spurious relationships: are reliable but not real, relationship might be due to chance; bivariate (single) regression: outcome by predictor; multivariate (multiple) regression: outcome by multiple predictors; association between outcome and each predictor “controls for” the other predictors

Question 12

Correlation relationship

Answer 12

Describes a relationship between two measured variables; no causation

Question 13

Determining Causation

Answer 13

1. start with two equal groups
2. manipulate a predictor variable
3. measure outcome

Question 14

Correlation Analysis

Answer 14

can be extended to test for moderating variables (interactions); a third variable that affects strength of relationship; the two predictor variables interact; mediating variables: one predictor is a mechanism for another predictor