How to set the significance level alpha?

[Ruth]

Hi,

I added the alpha=0.01 in the proc logistic, and the selection method is automatic. In doing this, I hope SAS can only keep those variables with significance below 0.01. But the final set of variables have a few with significance larger than 0.01.

I checked and found alpha only sets the confidence interval of the estimates.

Is there any option in regression analysis that can control the overall significance level for the variable selection?

Thanks in advance.

[ieva]

Hi,

Significance level of variables is not the only thing that should be looked at when making regression. There are a lot more like the basic R square, Durbin-Watson test and lot more. It is possible to make regressions where variables are significant, but model is not correct. And – sometimes - regressions where one or few variables are less significant, but still important.

So, I don’t know if there is a possibility of option to keep only significant variables, but most likely there is none because this wouldn’t be the best way to make regressions.

Regards,

Ieva

[data_null__ · Accepted solution]

I think these are the OPTIONS you seek.

SLENTRY=value

SLE=value

specifies the significance level of the score chi-square for entering an effect into the model in the FORWARD or STEPWISE method. Values of the SLENTRY= option should be between 0 and 1, inclusive. By default, SLENTRY=0.05. The SLENTRY= option has no effect when SELECTION=NONE, SELECTION=BACKWARD, or SELECTION=SCORE.

SLSTAY=value

SLS=value

specifies the significance level of the Wald chi-square for an effect to stay in the model in a backward elimination step. Values of the SLSTAY= option should be between 0 and 1, inclusive. By default, SLSTAY=0.05. The SLSTAY= option has no effect when SELECTION=NONE, SELECTION=FORWARD, or SELECTION=SCORE.