dealing closely selected models

[Deepak]

Dear Darryl,
I had run single season analysis for my data set with a combination of six covariates. 1 continous and the rest were categorical. Most top selected models were associated with a continious covariate "threat". Now the issue is the 70 models each model varies from the previous model in delta AIC value with 0.1 - 1.0 difference, if this is the case even if the other covariates had some effect on the parameter estimates they will eventually go to the back in the line of selected models, how do I deal with it? Should I do model averaging? would you please clarify it for me.



Regards,
Deepak