When you say the model is non-significant, does that mean the model is no good?
A model that is non-significant is a model that has no explanatory power whatsoever.
For instance, say you are wondering if tomorrow is going to rain or is going to be sunny. There are ways to predict that and weather forecasting agencies are using them.
Now, I tell you: here’s my model… I flip a coin. If it is heads it is going to be sunny, if it is tails, it is going to rain.
I may be correct sometimes, but my model is non-significant. Whenever I am correct, it is pure coincidence rather than ‘the fruit of a good model’.
In essence we prefer saying: non-significant, because especially in a situation with 2 outcomes, the model will inevitably be correct sometimes. Thus, it doesn’t make sense to call it: wrong. It is just.. not relevant or not significant.
Hope this helps!
The 365 Team