binary logistic regression

[Johnny Amora]

Hi experts,

I have two questions regarding binary logistic
regression:

(1) Suppose the outcome variable is YES/NO (coded 1
and 0). What did the literature say about the
proportion of YES and NO in the sample? That is, are
there studies saying that the NO should be x% and the
YES should be (1-x)% of the sample?Please provide the
reference (books or published article).

(2) In the case of nominal predictor variable, are
there studies restricting the number of categories?
That is, can I use 7 or more categories. Assume that
the sample size is sufficient.

Thank you in advance.

Johnny


      ____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

=====================
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

[vlad simion]

Hi Johnny,

Considering that your summation of YES/NO responses, should add up to the
whole sample (100%), it would be normal that if the proportion of YES is x%,
than the proportion of NO to be (1-x)%. That is, if I correctly understood
your question.

Concerning your second question, it is recommended to have at least 10 cases
per predictor. I don't know of any restrictions regarding the number of
categories in the predictor, but you must be carefull when having a lot of
categories, not to end up with zero cell counts.

A good refference is Alan Agresti's book : "Categorical Data Analysis".

Hth,
Vlad

On Fri, Feb 15, 2008 at 2:47 PM, John Amora <[hidden email]> wrote:

> Hi experts,
>
> I have two questions regarding binary logistic
> regression:
>
> (1) Suppose the outcome variable is YES/NO (coded 1
> and 0). What did the literature say about the
> proportion of YES and NO in the sample? That is, are
> there studies saying that the NO should be x% and the
> YES should be (1-x)% of the sample?Please provide the
> reference (books or published article).
>
> (2) In the case of nominal predictor variable, are
> there studies restricting the number of categories?
> That is, can I use 7 or more categories. Assume that
> the sample size is sufficient.
>
> Thank you in advance.
>
> Johnny
>
>
>
>  ____________________________________________________________________________________
> Be a better friend, newshound, and
> know-it-all with Yahoo! Mobile.  Try it now.
> http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
>
> =====================
> 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
>

=====================
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

[Johnny Amora]

Vlad wrote:

> Considering that your summation of YES/NO responses,
> should add up to the whole sample (100%), it would
>be normal that if the proportion of YES is x%,
> than the proportion of NO to be (1-x)%.

My question is "what is x"? Is x equals 50%, 60%, 70%?
Are there studies about this?

Thanks!

> Concerning your second question, it is recommended
> to have at least 10 cases
> per predictor. I don't know of any restrictions
> regarding the number of
> categories in the predictor, but you must be
> carefull when having a lot of
> categories, not to end up with zero cell counts.
>
> A good refference is Alan Agresti's book :
> "Categorical Data Analysis".
>
> Hth,
> Vlad
>
> On Fri, Feb 15, 2008 at 2:47 PM, John Amora
> <[hidden email]> wrote:
>
> > Hi experts,
> >
> > I have two questions regarding binary logistic
> > regression:
> >
> > (1) Suppose the outcome variable is YES/NO (coded
> 1
> > and 0). What did the literature say about the
> > proportion of YES and NO in the sample? That is,
> are
> > there studies saying that the NO should be x% and
> the
> > YES should be (1-x)% of the sample?Please provide
> the
> > reference (books or published article).
> >
> > (2) In the case of nominal predictor variable, are
> > there studies restricting the number of
> categories?
> > That is, can I use 7 or more categories. Assume
> that
> > the sample size is sufficient.
> >
> > Thank you in advance.
> >
> > Johnny
> >
> >
> >
> >
>

____________________________________________________________________________________
> > Be a better friend, newshound, and
> > know-it-all with Yahoo! Mobile.  Try it now.
> >
>
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

> >
> > =====================
> > 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
> >
>

      ____________________________________________________________________________________
Looking for last minute shopping deals?
Find them fast with Yahoo! Search.  http://tools.search.yahoo.com/newsearch/category.php?category=shopping

=====================
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