In: Statistics and Probability
You want to determine if there is a linear relationship between mental health and physical health. You have several people assessed by a psychologist and a medical doctor. The psychologist assess mental health (ment) on a scale of 1-30 (higher numbers mean better mental health) and the medical doctors rated their overall physical health (phys) from 1-30 (higher numbers mean better health). You obtain the data below. (note: there is more information than you need)
Person | MENT | PHYS | + - |
Rank MENT |
Rank PHYS |
MENT*PHYS | MENT2 | PHYS2 |
1 | 16 | 17 | + | 6 | 7 | 272 | 256 | 289 |
2 | 17 | 16 | - | 5 | 8 | 272 | 289 | 256 |
3 | 12 | 20 | + | 9 | 5 | 240 | 144 | 400 |
4 | 18 | 23 | + | 4 | 3 | 414 | 324 | 529 |
5 | 19 | 25 | + | 3 | 2 | 475 | 361 | 625 |
6 | 20 | 21 | + | 2 | 4 | 420 | 400 | 441 |
7 | 15 | 19 | + | 7 | 6 | 285 | 225 | 361 |
8 | 14 | 15 | + | 8 | 9 | 210 | 196 | 225 |
9 | 23 | 28 | + | 1 | 1 | 644 | 529 | 784 |
10 | 11 | 14 | + | 10 | 10 | 154 | 121 | 196 |
MEAN | 16.5 | 19.8 | 5.5 | 5.5 | 338.6 | 284.5 | 410.6 | |
Variance | 13.61111 | 20.62222 | 231 | 231 | 863867.26 | 609935.62 | 1273662.67 |
Determine the equation for the least squares regression line predicting mental health from physical health.
ΣY =
ΣX =
ΣY2 =
ΣX2 =
ΣXY =
SSXY =
SSX =
SSY =
N =
b =
mean of X =
mean of Y =
a =
type your final regression line in this box
What distance above and below the regression line would capture approximately 96% of the data points
If someone's physical health score was 15 what would you predict for their mental health score?
What is the correlation between mental health ad physical health
How much variance in mental health can be explained by physical health?
Explain this correlation in plain terms that someone who knew nothing about statistics would understand
Person | PHYS (X) | MENT (Y) | X^2 | XY | Y^2 |
1 | 17 | 16 | 289 | 272 | 256 |
2 | 16 | 17 | 256 | 272 | 289 |
3 | 20 | 12 | 400 | 240 | 144 |
4 | 23 | 18 | 529 | 414 | 324 |
5 | 25 | 19 | 625 | 475 | 361 |
6 | 21 | 20 | 441 | 420 | 400 |
7 | 19 | 15 | 361 | 285 | 225 |
8 | 15 | 14 | 225 | 210 | 196 |
9 | 28 | 23 | 784 | 644 | 529 |
10 | 14 | 11 | 196 | 154 | 121 |
SUM | 198 | 165 | 4106 | 3386 | 2845 |
Observation | Predicted Y | Residuals | Resdual ^2 |
1 | 14.70474138 | 1.295258621 | 1.677694894 |
2 | 14.06357759 | 2.936422414 | 8.622576592 |
3 | 16.62823276 | -4.628232759 | 21.42053847 |
4 | 18.55172414 | -0.551724138 | 0.304399524 |
5 | 19.83405172 | -0.834051724 | 0.695642279 |
6 | 17.26939655 | 2.730603448 | 7.456195192 |
7 | 15.98706897 | -0.987068966 | 0.974305143 |
8 | 13.42241379 | 0.577586207 | 0.333605826 |
9 | 21.7575431 | 1.242456897 | 1.54369914 |
10 | 12.78125 | -1.78125 | 3.172851563 |
SUM | 46.20150862 |
Interpretation of correlation: The correlation coefficient tells the strength of association between two values that vary with each other. In this case, the physical and mental health ratings by the doctors are assessed independently, and then we try to ascertain if one can be predicted from the other. Correlation is measured from -1 to +1, where -1 would indicate a strong negative linear relation i.e. one decreases as the other increases, 0 would mean no correlation and +1 would mean a strong (perfect) positive linear relationship.
Link to excel on which values were obtained: https://drive.google.com/file/d/1iXtbYBXIUeS5BKz1zo6jl3j_zYD0x8fA/view?usp=sharing