Date: Tue, 14 Jul 2009 18:00:36 -0300
Reply-To: Hector Maletta <hmaletta@fibertel.com.ar>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Hector Maletta <hmaletta@fibertel.com.ar>
Subject: Re: Regression Qualitatitive Factor
In-Reply-To: <24486961.post@talk.nabble.com>
Content-Type: text/plain; charset="us-ascii"
Jimjohn,
In fact, a binary variable valued at 0 and 1 can be rightly regarded as an
interval variable, where the unit of measurement is the distance between the
situation at 0 and the situation at 1. As only one interval is involved, and
the interval involved equals the unit of measurement, there is no problem of
comparing intervals (as in ordinal variables). (If other figures are used
instead of 0 and 1, e.g. 1 and 2, nothing changes).
However, some purists say that using dummies (binaries) may infringe on some
of the assumptions of linear regression. In particular, the error term may
not have a normal distribution.
Others also say that a binary independent variable itself cannot have a
normal distribution, but they are barking at the wrong tree: independent
variables need not have a normal distribution for linear regression to be
applicable.
Whatever the objections, dummies are usually used in regression (chiefly in
econometrics) without compunction. You can follow the crowd and do it too,
without losing any sleep because of it.
Hector
-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of
jimjohn
Sent: 14 July 2009 17:37
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Regression Qualitatitive Factor
Can someone please tell me if there is anyting special I need to do if im
running regression in SPSS where one of my independent variables is binary
(it is either 1 or 0). Can I just run the regression as I normally do?
Thanks in advance!
--
View this message in context:
http://www.nabble.com/Regression-Qualitatitive-Factor-tp24486961p24486961.ht
ml
Sent from the SPSSX Discussion mailing list archive at Nabble.com.
=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (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
LISTSERV@LISTSERV.UGA.EDU (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
| 
