What is a partial correlation?
Download as PPTX, PDF25 likes23,456 views
Partial correlation estimates the relationship between two variables while removing the influence of a third variable. It is a way to determine the correlation between two variables when controlling for a third. For example, a researcher may want to know the correlation between height and weight but also wants to control for gender, which can influence bone and muscle structure. Using the data sample provided, the correlation between height and weight was 0.825 but decreased to 0.770 when controlling for gender, showing gender partially explains the relationship between height and weight.
What is a partial correlation?
- 1. What is a Partial Correlation?
- 2. Welcome to this brief explanation on a Partial Correlation!
- 3. Partial correlation is an expression of analyses of co-variance (ANCOVA) applied to questions of prediction and relationship
- 4. Partial correlation is an expression of analyses of co-variance (ANCOVA) applied to questions of prediction and relationship Prediction & Relationship
- 5. . . . rather than questions of differences addressed by ANCOVA.
- 6. . . . rather than questions of differences addressed by ANCOVA.
- 7. Partial correlation estimates the relationship between two variables while removing the influence of a third variable from the relationship.
- 8. Partial correlation estimates the relationship between two variables while removing the influence of a third variable from the relationship. Relationship between a guy and a girl
- 9. Partial correlation estimates the relationship between two variables while removing the influence of a third variable from the relationship. Taking out the effect of video games on that relationship
- 10. For example, a researcher might want to know the relationship between height and weight.
- 11. For example, a researcher might want to know the relationship between height and weight. &
- 12. However, she is aware that people’s bone and muscle structures vary according to gender.
- 13. However, she is aware that people’s bone and muscle structures vary according to gender.
- 14. She can calculate a partial correlation between height and weight while removing (holding constant, eliminating) the effect of gender on the correlation.
- 15. She can calculate a partial correlation between height and weight while removing (holding constant, eliminating) the effect of gender on the correlation. &
- 16. She can calculate a partial correlation between height and weight while removing (holding constant, eliminating) the effect of gender on the correlation. &
- 17. She can calculate a partial correlation between height and weight while removing (holding constant, eliminating) the effect of gender on the correlation. & The Effect of Gender
- 18. Here’s the data set:
- 19. Here’s the data set: Individual Height (inches) Weight (pounds) Gender (1 – male, 2 – female) A 73 240 1 B 70 210 1 C 69 180 1 D 68 160 1 E 70 150 2 F 68 140 2 G 67 135 2 H 62 120 2
- 20. Here’s the data set: Individual Height (inches) Weight (pounds) Gender (1 – male, 2 – female) A 73 240 1 B 70 210 1 C 69 180 1 D 68 160 1 E 70 150 2 F 68 140 2 G 67 135 2 H 62 120 2
- 21. First let’s see what the correlation between height and weight is and then we will see what the correlation would be if we controlled for gender
- 22. Correlation between Height and Weight
- 23. Correlation between Height and Weight Individual Height (inches) Weight (pounds) Gender (1 – male, 2 – female) A 73 240 1 B 70 210 1 C 69 180 1 D 68 160 1 E 70 150 2 F 68 140 2 G 67 135 2 H 62 120 2
- 24. Correlation between Height and Weight
- 25. Correlation between Height and Weight & = .825
- 26. However, when controlling for gender the correlation between height and weight is:
- 27. However, when controlling for gender the correlation between height and weight is: & controlling for
- 28. However, when controlling for gender the correlation between height and weight is: & controlling for = .770
- 29. While still large, the relationship between height and weight is accounted for in part by gender.
- 30. While still large, the relationship between height and weight is accounted for in part by gender.