Respected Researchers
In gravity model, many researchers have involved dummies to check the impact e.g.( WTO, RTA, Region) in the form of 1 and 0.
In my gravity, I want to check the influence of one Belt One road countries ( 56 countries) taken as dummy '1' for if country i in OBOR otherwise '0'.
How to get a positive and significant sign for it.
Thank you in advance.
Dear Muhammad,
It sounds like you know how to create dummy variables. Multiple linear regression is a complex process. We refer students to the textbooks by Andy Field and Tabachnik and Fidel.
I would recommend using a backwards elimination method. This will remove items from the model with a high probability value one by one until the model stabilizes.
Dear Peter
Thank for your reply. Yes, i can use backwards technique but if my important variables are also in that list those who are not significant or showing opposite sign then what should I do. I can't remove them from my model because those are in demanding list.
It's not really clear what you are asking because you have not explained your model or what you are testing. But generally speaking, asking how to get a significant result from data you already have is not good research practice; your results are your results, and p-hacking the analysis to try to make non-significant results become significant is not good. See, e.g., Simmons et al. (2011), attached. Just report what happened in your study, don't try to change non-significant predictors into significant ones just to fit the outcome you wanted.
http://journals.sagepub.com/doi/abs/10.1177/0956797611417632
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression