Question regarding Regression analysis

[devoidx]

Hi guys, I have 3 categorical age variables (age0to17, age18to40 and age41plus). for each variable 1 is yes and 2 is No. I put these 3 variables into a regression analysis as independent variables. My dependent variable is whether they had a certain operations (1 is yes, 2 is no). However, i always get missing values in the regression results for the third age variable (regardless of the order of putting in the variables into the regression, the third one turns out missing-as in there is a "." across the board in the regression results table). I don't understand why and how i can go about fixing this. any help? thanks

[devoidx]

Forgot to mention that it says "redundant" for both the yes and no values of the third age variables

[peter link]

You need to exclude one of the categories - this will be your reference
category.

peter

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
devoidx
Sent: Tuesday, June 02, 2015 9:26 AM
To: [hidden email]
Subject: [SPSSX-L] Question regarding Regression analysis

Hi guys, I have 3 categorical age variables (age0to17, age18to40 and
age41plus). for each variable 1 is yes and 2 is No. I put these 3 variables
into a regression analysis as independent variables. My dependent variable
is whether they had a certain operations (1 is yes, 2 is no). However, i
always get missing values in the regression results for the third age
variable (regardless of the order of putting in the variables into the
regression, the third one turns out missing-as in there is a "." across the
board in the regression results table). I don't understand why and how i can
go about fixing this. any help? thanks



--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/Question-regarding-Regression-
analysis-tp5729719.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

[Rich Ulrich]

"Redundant."

With a 2-category item like Sex, you have 1 d.f. and you can only
use one dummy variable, instead of having separate predictors for
Male and Female.  With a 3-category item like yours, you have 2 d.f.
and you can only use two dummy variables to encode it without
redundancy.  That's why you got the message that you report in your
followup message. 

Is there a reason not to use the actual age (linear)?  For three groups,
I would probably chose to apply my available 2 d.f.  in a linear contrast
(1,2,3)  and a non-linear contrast (0,1,0) -- unless there was an a-priori
reason to look at one age group for special emphasis. 

--
Rich Ulrich

> Date: Tue, 2 Jun 2015 09:26:04 -0700

> From: [hidden email]
> Subject: Question regarding Regression analysis
> To: [hidden email]
>
> Hi guys, I have 3 categorical age variables (age0to17, age18to40 and
> age41plus). for each variable 1 is yes and 2 is No. I put these 3 variables
> into a regression analysis as independent variables. My dependent variable
> is whether they had a certain operations (1 is yes, 2 is no). However, i
> always get missing values in the regression results for the third age
> variable (regardless of the order of putting in the variables into the
> regression, the third one turns out missing-as in there is a "." across the
> board in the regression results table). I don't understand why and how i can
> go about fixing this. any help? thanks
>

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

[devoidx]

Thanks Rich, Reason I'm not looking at it as a linear age is because the purpose is to see whether specific age groups are more likely to have a certain operation when compared to all other groups together. I didn't quite understand the last part of your response. Is there a way to accomplish why I'm trying to do? the regression has other independent variables other than age that i have put into it but my only issue is figuring out the age situation.

Thanks.

[Rich Ulrich]

If you want to look at adjusted group means, then it seems to me that
the sensible model is the ANOVA.  Enter the other variables from the
Regression as covariates.  Look at contrasts if you want.

--
Rich Ulrich

> Date: Tue, 2 Jun 2015 10:31:17 -0700

> From: [hidden email]
> Subject: Re: Question regarding Regression analysis
> To: [hidden email]
>
> Thanks Rich, Reason I'm not looking at it as a linear age is because the
> purpose is to see whether specific age groups are more likely to have a
> certain operation when compared to all other groups together. I didn't quite
> understand the last part of your response. Is there a way to accomplish why
> I'm trying to do? the regression has other independent variables other than
> age that i have put into it but my only issue is figuring out the age
> situation.
>

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

[David Marso]

"my only issue is figuring out the age situation."
Seems possible that your additional issue(s) is not comprehending the fundamental basics of regression analysis an linear models (in particular the notions of LINEAR DEPENDENCE, RANK, INVERTABILITY etc...).
 
Consider the following:
data list free/ v1 v2 v3.
begin data
1 0 0  0 1 0  0 0 1
END DATA.
REGRESSION DEPENDENT v3 / METHOD ENTER v1 v2.

devoidx wrote

Thanks Rich, Reason I'm not looking at it as a linear age is because the purpose is to see whether specific age groups are more likely to have a certain operation when compared to all other groups together. I didn't quite understand the last part of your response. Is there a way to accomplish why I'm trying to do? the regression has other independent variables other than age that i have put into it but my only issue is figuring out the age situation.

Thanks.