# Partial Correlation

HideShow resource information
• Created by: Chloe
• Created on: 30-04-15 14:01
• Partial Correlation
• Interpreting a correlation between two variables (X&Y) can be made difficult if both are correlated with a third variable (Y)
• This may give rise to a 'spurious' relationship between X&Y acting as a 'common cause' of changes in both
• A partial correkation procedure can remove (partial out) the effect of the third variable from the correlation between the two variables of interest
• Venn Diagrams: used to express relationships between sets - they represent variance as overlapping areas associated with two or more sources
• - Coeffiicient of determination = R^2 - Correlation coefficient equals the square root of the coefficient of determination
• Example: (a) shows that X account for 49% of the varience in Y (rxy =.7) (b) shows that Z accounts for 36% of the varience in Y (rzy = .6) (c) shows that 33% of the varience in Y is shared by X and Z (rxz = ?.33 = .57) leaving only 16% of the varuence uniquels attribituable to X (d) also indicated that only 3% of the variance in Y is uniquels attributable to Z. Almost all of it is due to its association with X