Correlational Methods (Bivariate) - Exam 3

Bivariate Correlation

measure of association between exactly two variables. Ranges from -1.00 to 1.00

Positive Linear Correlation

When one variable increases, and so does another

Negative Linear Correlation

One variable increases and the other decreases

Effect Size

describes the strength of an association. Correlation coefficient is representative of it

Correlation Coefficient

a statistical measure of the strength and direction of a linear relationship between 2 variables. Also called Pearson’s r

Very small or weak effect size for r.

Answer in this format: .01 , .10 and so on. NOT like this: .05 (or -.05) only give the first number but recognize they both count positive and negative

.05

small or weak r effect size

Answer in this format: .01 , .10 and so on. NOT like this: .05 (or -.05) only give the first number but recognize they both count positive and negative

.10

Moderate r effect size

Answer in this format: .01 , .10 and so on. NOT like this: .05 (or -.05) only give the first number but recognize they both count positive and negative

.20

fairly powerful effect of r

Answer in this format: .01 , .10 and so on. NOT like this: .05 (or -.05) only give the first number but recognize they both count positive and negative

.30

Unusually large in psychology - VERY powerful or too good to be true. Can be greater than the answer, but that is the lower limit

Answer in this format: .01 , .10 and so on. NOT like this: .05 (or -.05) only give the first number but recognize they both count positive and negative

.40

Ideal minimum r value for good reliability in reliability studies

.80

Factors influencing the magnitude of a correlation

Outliers, restriction of range, curvilinear relationships

Outlier

An extreme score that may not belong with the other data due to sampling issues or errors. Reduces the size of the real correlation

Restriction of Range

limits the magnitude of the correlation by having a range of scores that is more narrow than that within the populatio

Curvilinear relationships

A non-linear relationship between two variables.

Requirements to demonstrate causation

Covariation, temporal precedence, no alternate explanations (internal validity). Correlational studies do not meet these requirements

Covariation

the extent to which one variable moves in tandem with another

Temporal Precedence

the cause must occur before the associated effect

Linear Regression

choosing the best fitting line to describe the relationship between two variables

principle of least squares

a prediction line that minimizes the total squared difference between observed values and predicted values

Predictor Variable

Independent variable within a linear regression. It is displayed on the X axis

Criterion Variable

dependent variable within a linear regression that is displayed on the Y axis

Construct Validity

How well each variable is measure

Statistical Validity

How well the data supports the conclusion

Internal Validity

Can a causal inference be made from an association

External Validity

can the results be generalized to the whole population being studied?

Moderator

a variable that affects the strength or direction between the independent and dependent variables