Using a continuous variable with both negative and positive values and zero in regression

[saba.haq]

Hi,
I am doing regression analysis with dichotomous out come variable but continuous predictor variables. One of the variables I have has negative, positive and zero values like this -2.45, -3.46, 1, 0, 2.45, 3.56,, 4.76......
I need to know if I can use it as continuous variable in regression?? or do I need convert these values. If I need to convert these how do I do that.

But if can use it as continuous variable (which be great) what will be interpretation of the results??? I am really confused. I have never came across such a situation before and have not seen anything on internet or in books. It will be a huge help if someone can help me sort this out.


Thanking in advance.

[MaxJasper]

Continuous variables (SPSS Scale variables) can have any numerical values.

 

Outcome will be {dies, lives} for {-3.73, +4.6}

 

I am doing regression analysis with dichotomous out come variable but continuous predictor variables. One of the variables I have has negative, positive and zero values like this -2.45, -3.46, 1, 0, 2.45, 3.56,, 4.76......
I need to know if I can use it as continuous variable in regression?? or do I need convert these values. If I need to convert these how do I do that.

But if can use it as continuous variable (which be great) what will be interpretation of the results??? I am really confused. I have never came across such a situation before and have not seen anything on internet or in books. It will be a huge help if someone can help me sort this out.

[Bruce Weaver]

There is no reason why a continuous predictor variable cannot have both negative and positive values.  (E.g., mean-centered variables, which are quite commonly used, will have both negative & positive values.)

Logistic regression is often used when one has a dichotomous outcome variable.  Is that what you're using?  If you are, the Exp(B) value for your continuous predictor variable gives the odds ratio associated with a one unit increase in that predictor.

For more info on interpreting odds ratios, see this page from the folks at UCLA:

  http://www.ats.ucla.edu/stat/mult_pkg/faq/general/odds_ratio.htm

HTH.

saba.haq wrote

Hi,
I am doing regression analysis with dichotomous out come variable but continuous predictor variables. One of the variables I have has negative, positive and zero values like this -2.45, -3.46, 1, 0, 2.45, 3.56,, 4.76......
I need to know if I can use it as continuous variable in regression?? or do I need convert these values. If I need to convert these how do I do that.

But if can use it as continuous variable (which be great) what will be interpretation of the results??? I am really confused. I have never came across such a situation before and have not seen anything on internet or in books. It will be a huge help if someone can help me sort this out.


Thanking in advance.

[Maguin, Eugene]

You must be doing logistic regression since you have a dichotomous dependent variable. The negative numbers are no different from positive numbers. The regression coefficient is the log odds of a one unit change in the predictor.  Nothing is changed because your predictor has some negative values.

Gene Maguin

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of saba.haq
Sent: Monday, August 18, 2014 12:25 PM
To: [hidden email]
Subject: Using a continuous variable with both negative and positive values and zero in regression

Hi,
I am doing regression analysis with dichotomous out come variable but continuous predictor variables. One of the variables I have has negative, positive and zero values like this -2.45, -3.46, 1, 0, 2.45, 3.56,, 4.76......
I need to know if I can use it as continuous variable in regression?? or do I need convert these values. If I need to convert these how do I do that.

But if can use it as continuous variable (which be great) what will be interpretation of the results??? I am really confused. I have never came across such a situation before and have not seen anything on internet or in books. It will be a huge help if someone can help me sort this out.


Thanking in advance.



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Using-a-continuous-variable-with-both-negative-and-positive-values-and-zero-in-regression-tp5726993.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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

[Art Kendall]

There is no reason why you cannot have a continuous variable with positive and negative values.

What is the substative meaning of your DV?


What are your IVs?

How many cases do you have in each of the two groups defined by the dichotomy?

If you have a variable that is necessarily dichotomous, and you have another set of continuous variables would you be interested in "what differentiates (discriminates between) cases with the two different values?"

[saba.haq]

DV is presence and absence of depression and

IVs are attachment styles, emotional regulation, personality factors
It s the attachment style which has negative and positive values

It is a survey with 213 cases of depression and 1500 non cases

If I understood it correctly I am interested in looking at what differentiates (discriminates between) cases with the two different values (

so in that case do i have to make the IV dichotomous?

will be grateful if you could advise on that

[Art Kendall]

How was depression diagnosed?  most test result in continuous test scores?

"I am interested in looking at what differentiates (discriminates between) cases with the two different values "

Then one way to look at your data is to do do a DISCRIMINANT.

"so in that case do i have to make the IV dichotomous? "
I see no reason for that.

Are you really stuck with depression as a dichotomy? Do you have a way to get a less coarse measure?

[Jon K Peck]

I think it is time to suggest that you need to consult with your local statistical consulting resources in more depth.  Judging from the questions you are asking, you do not understand regression and related issues well enough to succeed in this analysis with just the advice you can get from this list.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
phone: [phone]




From:        "saba.haq" <[hidden email]>
To:        [hidden email]
Date:        08/19/2014 05:12 AM
Subject:        Re: [SPSSX-L] Using a continuous variable with both negative and positive values and zero in regression
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

DV is presence and absence of depression and

IVs are attachment styles, emotional regulation, personality factors
It s the attachment style which has negative and positive values

It is a survey with 213 cases of depression and 1500 non cases

If I understood it correctly I am interested in looking at what
differentiates (discriminates between) cases with the two different values (

so in that case do i have to make the IV dichotomous?

will be grateful if you could advise on that



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Using-a-continuous-variable-with-both-negative-and-positive-values-and-zero-in-regression-tp5726993p5727008.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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