range of standardized regression coefficients

[vilas]

What would be the range of standardized regression coefficients in theory? I rarely see standardized regression coefficients greater than 1 or -1. Then, are the coefficients unreliable? Does it mean that the data violate statistical assumptions (e.g., normal distribution)? Comments will be greatly appreciated. Thank you!

[statisticsdoc]

Standardized coefficients have a range between -1 and 1 algebraically.  Software packages sometimes generate coefficiebts slightly outside of this range (it can happen when the independent variables are extremely collinear).
Hth
Steve
www.statisticsdoc.com
------Original Message------
From: vilas
Sender: SPSSX(r) Discussion
To: [hidden email]
ReplyTo: vilas
Subject: range of standardized regression coefficients
Sent: Dec 15, 2009 1:51 PM

What would be the range of standardized regression coefficients in theory? I
rarely see standardized regression coefficients greater than 1 or -1. Then,
are the coefficients unreliable? Does it mean that the data violate
statistical assumptions (e.g., normal distribution)? Comments will be
greatly appreciated. Thank you!
--
View this message in context: http://old.nabble.com/range-of-standardized-regression-coefficients-tp26799741p26799741.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


Sent from my Verizon Wireless BlackBerry

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

[Bruce Weaver]

statisticsdoc wrote

Standardized coefficients have a range between -1 and 1 algebraically.  Software packages sometimes generate coefficiebts slightly outside of this range (it can happen when the independent variables are extremely collinear).
Hth
Steve
www.statisticsdoc.com

From that website:

"I am an Oxford-educated Ph.D. with over twenty years of  experience in statistical consulting and research design."

Also from that website:

 Education       B.A. (First Class Honours), University of Oxford
                     Ph.D. (Psychology; Minor in Quantitative Methods), University of Illinois

Perhaps I was reading things into it, Steve, but that first statement had me thinking your PhD was from Oxford.  I suppose "U of Illinois-educated PhD" just doesn't sound as good, does it.*  ;-)

* No offense to any U of Illinois alumni lurking on the list!  If it makes you feel any better, I'm a lowly U of Guelph and McMaster University educated BA & MSc, also in Psychology.  (That's the formal part of my education, at least.)

[statisticsdoc]

Well, over the years,  I have read some good social research from Guelph.
Seasons Greetings
SteveB
Sent from my Verizon Wireless BlackBerry

-----Original Message-----
From: Bruce Weaver <[hidden email]>
Date:         Tue, 15 Dec 2009 15:21:20
To: <[hidden email]>
Subject:      Re: range of standardized regression coefficients

statisticsdoc wrote:
>
> Standardized coefficients have a range between -1 and 1 algebraically.
> Software packages sometimes generate coefficiebts slightly outside of this
> range (it can happen when the independent variables are extremely
> collinear).
> Hth
> Steve
> www.statisticsdoc.com
>

From that website:

"I am an Oxford-educated Ph.D. with over twenty years of  experience in
statistical consulting and research design."

Also from that website:

 Education       B.A. (First Class Honours), University of Oxford
                     Ph.D. (Psychology; Minor in Quantitative Methods),
University of Illinois

Perhaps I was reading things into it, Steve, but that first statement had me
thinking your PhD was from Oxford.  I suppose "U of Illinois-educated PhD"
just doesn't sound as good, does it.*  ;-)

* No offense to any U of Illinois alumni lurking on the list!  If it makes
you feel any better, I'm a lowly U of Guelph and McMaster University
educated BA & MSc, also in Psychology.  (That's the formal part of my
education, at least.)



-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/
"When all else fails, RTFM."

NOTE:  My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.
--
View this message in context: http://old.nabble.com/range-of-standardized-regression-coefficients-tp26799741p26803514.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

[Mike Pritchard]

I'm trying to copy the variable definitions (including name, type, label and
values) from one file to another.

I don't have all of the syntax that originally set up the source file. I
thought I could use APPLY DICTIONARY with each variable identified, but this
doesn't let me change the names, and won't let me change other attributes if
the type or width might cause problems.

So now I think I have to create syntax for each variable (including RENAME,
TYPE, etc.)  For all 250 variables.

Is there a better way?

I'm using SPSS V16.

Thanks
Mike

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

[statisticsdoc]

Bruce,
BTW I am going to change my website over the break.  I can see how the wording of the landing page could be confusing.  Thank you for bringing this to my attention.
Best,
Stephen Brand
(Illini)
Sent from my Verizon Wireless BlackBerry

-----Original Message-----
From: Bruce Weaver <[hidden email]>
Date:         Tue, 15 Dec 2009 15:21:20
To: <[hidden email]>
Subject:      Re: range of standardized regression coefficients

statisticsdoc wrote:
>
> Standardized coefficients have a range between -1 and 1 algebraically.
> Software packages sometimes generate coefficiebts slightly outside of this
> range (it can happen when the independent variables are extremely
> collinear).
> Hth
> Steve
> www.statisticsdoc.com
>

From that website:

"I am an Oxford-educated Ph.D. with over twenty years of  experience in
statistical consulting and research design."

Also from that website:

 Education       B.A. (First Class Honours), University of Oxford
                     Ph.D. (Psychology; Minor in Quantitative Methods),
University of Illinois

Perhaps I was reading things into it, Steve, but that first statement had me
thinking your PhD was from Oxford.  I suppose "U of Illinois-educated PhD"
just doesn't sound as good, does it.*  ;-)

* No offense to any U of Illinois alumni lurking on the list!  If it makes
you feel any better, I'm a lowly U of Guelph and McMaster University
educated BA & MSc, also in Psychology.  (That's the formal part of my
education, at least.)



-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/
"When all else fails, RTFM."

NOTE:  My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.
--
View this message in context: http://old.nabble.com/range-of-standardized-regression-coefficients-tp26799741p26803514.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