Date: Fri, 24 Oct 1997 09:14:08 -0500
Reply-To: "Nichols, David" <nichols@SPSS.COM>
Sender: "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From: "Nichols, David" <nichols@SPSS.COM>
Subject: Re: one more logistic question
When someone asks how to obtain something in an SPSS procedure,
I generally don't offer the option of computing it by hand, as
that's generally the thing they bought the software to avoid
having to do. Given the COMPUTE functionality and the MATRIX
language, the answer to the question "Can SPSS do X?" is always
yes, since someone sophisticated enough to program anything has
the available tools to do it. For most SPSS users, that's not
the most useful answer.
David Nichols
Senior Support Statistician
SPSS Inc.
nichols@spss.com
>----------
>From: Arnold Kester[SMTP:Arnold.Kester@UNIMAAS.NL]
>Sent: Friday, October 24, 1997 3:20 AM
>To: SPSSX-L@UGA.CC.UGA.EDU
>Subject: Re: one more logistic question
>
>David Nichols writes:
>>This is not available until Release 7.0. There, in syntax,
>>specify PRINT CI(.95) (or whatever coverage level you want).
>>In the dialogs, it's in the Options dialog. Note that the
>>level is set to .90 in the 7.0 release in the dialogs, which
>>was an error; it's at .95 in Release 7.5.
>>
>Actually, it isn't really hard to do by hand. Just calculate
>the (95%) interval B +- 1.96 * se(B), the transform both
>sides to OR's using the exp() function (hit the e button on
>your calculator). For other confidence levels change the 1.96
>to the appropriate normal quantile.
>
>>>----------
>>>From: Scott Roesch[SMTP:roesch@PROTOS.LIFESCI.UCLA.EDU]
>>>Sent: Tuesday, October 21, 1997 5:44 PM
>>>To: SPSSX-L@UGA.CC.UGA.EDU
>>>Subject: one more logistic question
>>>
>>>Any hints on how to obtain confidence intervals for the odds ratios in
>>>logistic regression?
>>>
>>>Scott R.
>>>
>
>--
>Arnold Kester
>Arnold.Kester@stat.unimaas.nl
>
| 
