Homework SourseIn a multiple regression only x1 and x2 are found to be sign
In a multiple regression, only x1 and x2 are found to be significant. However, insignificance can be an indication confounding variables. Given the correlation matrix below, does there appear to be confounding variables.
NOTE: Variables may be confounded when they are both correlated to the subject response and correlated with each other.
| Matrix | x2 | x3 | y |
|---|---|---|---|
| x1 | 0.17 | -0.70 | -0.6 |
| x2 | ___ | -0.10 | 0.89 |
| x3 | ___ | ___ | 0.91 |
Solution
x1 and x3
Confounding as x1 is correlated to x3 directly and both correlated to y
Confounding variables as they are correlated by pairs and also with y each.
: confounding as x2 and x3 are correlated and also with y each.
confounding as x2 and x3 are correlated and also with y each.
| x1, x2 and x3 |
