Correlation

Question 1

Correlation

Answer 1

-Statistical technique used to measure and describe the relationship (strength, direction and shape) between co-variables

Question 2

Mean

Answer 2

• basic data model
• Data = model + error
• Error = variance (deviation = sum of squares)

Question 3

Co-variance

Answer 3

• two variables that vary together

- x-x y-y / N-1

Question 4

Pearson’s r

Answer 4

• Most common correlation coefficient
• results: 0.1 Small effect – 0.3 Medium – 0.5 Large
• assumptions: interval, normalative, no anomalies, homogenous
• spearmans = non parametric

Question 5

Correlation coefficient

Answer 5

-Mean and variance can impact the shape or pattern data. Close linear clusters = good correlation. Anomalies can cause error so plot data points first

Question 6

Caveats

Answer 6

• correlation does mean causation
• Both variables must show variation
• restricting range reduces reliability
• third variable = spurious relationship
• restricting range reduces reliability

Question 7

Confounding variables

Answer 7

-Explains the variability in a relationship

Question 8

Variance

Answer 8

-average error

Question 9

Standardized variance

Answer 9

• Compare strength and direction across different measures
• predictor on X axis
• Standardized measure of linear relationship
• 1 - +1
• Coefficient of determination (R2) = proportion of variance explained

Question 10

Partial correlatin

Answer 10

• Removes variance caused by third variable
• 1st half of table = 0 order (no control)
• 2nd half of table = controlled condition (does value reduce)