logistic regression question

[Jordan Kennedy]

Hello all:

I have run a simple bivariate logistic regression SPSS 19 for Windows.
The predictor is significantly associated with the outcome--with a
fairly large Odds Ratio. However, the classification table listed does
not classify particularly well. I have pasted the output below. Any
help would be greatly appreciated.

                                        Classification Table(a)
                                Predicted
                        o_adhdalc
        Observed                .00     1.00    Percentage Correct
Step 1  o_adhdalc       .00     209     0       100.0
                1.00    74      0       .0
Overall Percentage                              73.9


                                Variables in the Equation

                B       S.E.    Wald    df      Sig.    Exp(B)
Step 1a p2_alcadhd_l    1.471   .298    24.377  1       .000
        Constant        -3.335  .512    42.450  1       .000    .036


Thank you!

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[lori.andersen]

Exp(1.471) = 4.35, that is your odds ratio

Is your independent variable categorical?
If not, a one unit increase in your IV would yield a 335% greater chance of membership in the "1" value of your DV.
If it is categorical, members of the IV group at the "1" level have a 335% greater chance of membership in the "1" value of the DV.

Lori

On Sun, Dec 4, 2011 at 7:49 PM, Jordan Kennedy [via SPSSX Discussion] <[hidden email]> wrote:

Hello all:

I have run a simple bivariate logistic regression SPSS 19 for Windows.
The predictor is significantly associated with the outcome--with a
fairly large Odds Ratio. However, the classification table listed does
not classify particularly well. I have pasted the output below. Any
help would be greatly appreciated.

                                        Classification Table(a)
                                Predicted
                        o_adhdalc
        Observed                .00     1.00    Percentage Correct
Step 1  o_adhdalc       .00     209     0       100.0
                1.00    74      0       .0
Overall Percentage                              73.9


                                Variables in the Equation

                B       S.E.    Wald    df      Sig.    Exp(B)
Step 1a p2_alcadhd_l    1.471   .298    24.377  1       .000
        Constant        -3.335  .512    42.450  1       .000    .036


Thank you!

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

---

If you reply to this email, your message will be added to the discussion below:

http://spssx-discussion.1045642.n5.nabble.com/logistic-regression-question-tp5047539p5047539.html

To start a new topic under SPSSX Discussion, email [hidden email]
To unsubscribe from SPSSX Discussion, click here.
NAML

--
Lori Andersen
Ph.D. student, Educational Policy, Planning & Leadership
College of William & Mary
Williamsburg, VA

[Rich Ulrich]

Sorry, I don't see the original post - I don't have any comment on Lori's
reply, but I can address the original question.  I have aligned the chart
(as my mail displays it) to show that all the individual were classified as "1".

Okay, you have a classification equation; you have an arbitrary cut-off
score that labels subjects into one group or the other.  Seeing the cutoff
place everyone in the same group is more common when the groups are
proportionately 90/10 instead of 74/26, but this shows it can happen here, too.

 - Look at the actual classification scores, and draw your own line
where it makes the fraction of errors about the same in either
direction.

--
Rich Ulrich

---

Date: Sun, 4 Dec 2011 18:06:48 -0800
From: [hidden email]
Subject: Re: logistic regression question
To: [hidden email]

Exp(1.471) = 4.35, that is your odds ratio

Is your independent variable categorical?
If not, a one unit increase in your IV would yield a 335% greater chance of membership in the "1" value of your DV.
If it is categorical, members of the IV group at the "1" level have a 335% greater chance of membership in the "1" value of the DV.

Lori

On Sun, Dec 4, 2011 at 7:49 PM, Jordan Kennedy [via SPSSX Discussion] <[hidden email]> wrote:

Hello all:

I have run a simple bivariate logistic regression SPSS 19 for Windows.
The predictor is significantly associated with the outcome--with a
fairly large Odds Ratio. However, the classification table listed does
not classify particularly well. I have pasted the output below. Any
help would be greatly appreciated.

                                        Classification Table(a)
                                Predicted
                             o_adhdalc
        Observed            .00     1.00    Percentage Correct
Step 1  o_adhdalc       .00     209     0       100.0
                                1.00    74      0       .0
Overall Percentage                              73.9


                                Variables in the Equation

                B       S.E.    Wald    df      Sig.    Exp(B)
Step 1a p2_alcadhd_l    1.471   .298    24.377  1       .000
        Constant             -3.335  .512    42.450  1       .000    .036


Thank you!
[...]