How to determine the most significant factor in a model

I have a model with 18 factors fitted. Is there a way to determine which factors are the best in the model. If you remove one and calculate the increase in the deviance and do this separately for each factor (but only exclude one factor each time) will the factor whose exclusion results in the biggest increase in the deviance be the most significant factor in the model?



I know there are subjective considerations but I'm looking at this from a purely statistical point of view.



Thanks

How to determine the most significant factor in a model