1

I have got a OLS on overweight with year and excise on sugar as independent variables. Excise isn't significant in this regression, but if I do a Chow breakpoint test on my data (for the year the excise was abandoned), it tests significant. So it is not about the interpretation of the insignificant excise, but about the significant break. How should I interpret this?

TLDR: How should I interpret a significant break in an insignificant independent variable?

J. Doe
  • 11
  • 4
  • please provide more information to clarify your question – Antoine Jun 15 '16 at 12:06
  • What is it you want me to add? – J. Doe Jun 15 '16 at 12:49
  • This sounds like a particular example of http://stats.stackexchange.com/questions/28474/ : by adding some extra variables to the regression (which is [how the Chow Test works](https://en.wikipedia.org/wiki/Chow_test)), a formerly "insignificant" variable becomes significant. – whuber Jun 15 '16 at 12:51
  • Not really, I guess, because I searched for a structural break in 1993, where the excise was abandoned. The Wald test in Stata doesn't add new variables then, right? Or am I just making an enormous mistake? Also, my main question is how to interpret the break. The independent variable excise does not have a significant effect, but how do I interpret sec. the significant break? – J. Doe Jun 15 '16 at 13:01

0 Answers0