[R] why does R stepAIC keep unsignificant variables?

[David Winsemius]

On Oct 8, 2012, at 5:43 PM, liang.che at us.pwc.com wrote:

> Ran a bunch of variables in R and the final result of StepAIC is as below: 
> Why are the first 5 variables kept in the stepwise result??  Are the last 
> 4 variables finally chosen after Stepwise?  Thanks
> 
> Coefficients:
> Estimate Std. Error t value Pr(>|t|)
> (Intercept) 1.315e-01 2.687e-01 0.490 0.63611
> Core_CPI__ 1.290e-02 7.496e-03 1.721 0.11927
> GDP_change -3.482e-03 2.075e-03 -1.678 0.12767
> Unemployment 1.209e-02 6.970e-03 1.735 0.11685
> interest -5.580e-03 3.923e-03 -1.422 0.18863
> housing 6.692e-04 5.812e-04 1.151 0.27928
> S_P 7.636e-05 3.967e-05 1.925 0.08641 .
> d1 2.087e-02 6.102e-03 3.421 0.00762 **
> d2 -2.059e-02 7.331e-03 -2.808 0.02043 *
> d3 -2.769e-02 6.268e-03 -4.418 0.00168 **
> ---
> Signif. codes: 0 ?***? 0.001 ?**? 0.01 ?*? 0.05 ?.? 0.1 ? ? 1
> Residual standard error: 0.008362 on 9 degrees of freedom
> Multiple R-squared: 0.9534, Adjusted R-squared: 0.9069
> F-statistic: 20.48 on 9 and 9 DF, p-value: 5.866e-05

install.package(fortunes)
requie(fortune)
fortune(182)

-- 

David Winsemius, MD
Alameda, CA, USA