In: Statistics and Probability
Consider the following small data set.
Subject | x | y |
1 | 7 | 21 |
2 | 17 | 15 |
3 | 15 | 24 |
4 | 5 | 19 |
5 | 14 | 30 |
Find the linear correlation coefficient.
?=
Given the following data set, let ? be the explanatory variable and ? be the response variable.
x | 7 | 6 | 3 | 8 | 2 | 5 | 3 |
y | 3 | 6 | 7 | 2 | 8 | 5 | 8 |
(a) If a least squares line was fitted to this data, what
percentage of the variation in the ? would be explained by the
regression line? (Enter your answer as a percent.)
ANSWER: %
(b) Compute the correlation coefficient: ?=
1)
correlation r='Sxy/(√Sxx*Syy) = | 0.1061 |
(please take care of number of decimal places)
2)
a)
SST=Syy= | 33.714286 | |
SSE =Syy-(Sxy)2/Sxx= | 3.708333 | |
SSR =(Sxy)2/Sxx = | 30.005952 |
Coefficient of determination R2=SSR/SST=0.8900 |
percentage of the variation in the ? would be explained by the regression line =89.0 %
b)
correlation r='Sxy/(√Sxx*Syy) = | -0.9434 |