0
$\begingroup$

I am modelling a linear regression of two retail prices. Correlation shows a positive relationship. When I regressed it with each other, it showed a positive coefficient. However, when I added a trend variable to, it was significant but the regression coefficient becomes negative. How did this happen?[This is without trend and the next is when I added a trend][1]

This is with trend

Improve this question
$\endgroup$
2
  • $\begingroup$ That often happens when the explanatory and objective variables follow a similar trend. Due to the similar trend, they have a positive correlation. Including the trend means you control the influence of the trend. The coefficient sign may change as a result. $\endgroup$
    – Kota Mori
    Commented Oct 15, 2022 at 12:03
  • $\begingroup$ @KotaMori then is the regression with trend still valid? How will I interpret this? Do you know any existing literature I can read about this? $\endgroup$
    – ash
    Commented Oct 15, 2022 at 12:08

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.