LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (September 2005)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Fri, 16 Sep 2005 08:35:24 -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: Missing values analysis
Comments: To: kylie.lange@flinders.edu.au
In-Reply-To:  <432A775E.5000805@flinders.edu.au>
Content-Type: text/plain; charset="iso-8859-1"

Yes, Kylie, I apparently misunderstood. I thought he was trying to AVOID multiple values imputed, when apparently he was asking it was possible to obtain slightly different imputed values (possibly through the inclusion of a stochastic term). However, you put things right in the sense that SPSS's MVA does not do multiple or stochastic imputation.

Hector

> -----Original Message----- > From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf Of Kylie Lange > Sent: Friday, September 16, 2005 4:42 AM > To: SPSSX-L@LISTSERV.UGA.EDU > Subject: Re: Missing values analysis > > Hi Bozena, > > I think Hector may have slightly misunderstood what you are > trying to achieve, but indirectly he has answered your > question. My understanding (and Hector's as well from his > reply) is that SPSS' imputation methods are completely > deterministic - ie, the two cases with the same values on the > observed variables will have the same imputed value, and > additionally each time you run the imputation they will have > the same imputed value. For multiple imputation (where a > random error component is included in deriving the imputed > value) the results will differ slightly, as you say. For that > you will need a program that specifically implements a > multiple imputation method. > > For a list of multiple imputation software see > http://www.multiple-imputation.com/. > > Hopefully this helps, and others will chime in if I'm off > track with anything here. > > Cheers, > Kylie. > > > On 16/09/2005 2:27 AM, Hector Maletta wrote: > > >Bozena, > >It might be that the second run takes into account the > values imputed > >in the first run, and this may be the cause of the > difference. To avoid > >that, start every time with the same untouched file, or copy the old > >variables under new names before imputing valid values for > old missing values. > > > >Hector > > > > > > > >>-----Original Message----- > >>From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf > >>Of Zdaniuk, Bozena > >>Sent: Thursday, September 15, 2005 1:38 PM > >>To: SPSSX-L@LISTSERV.UGA.EDU > >>Subject: Missing values analysis > >> > >>Hello, I am entering the unfamiliar territory so please > forgive if my > >>question is naďve or parts of it incorrect. > >>Question: Is there a way in MVA in spss to have multiple > imputation of > >>missing data? In other words, is it possible to run, let's > say, a few > >>EM runs with the same predicted and predicting variables in > the syntax > >>and get slightly different imputed values on each run? > >>Thanks. > >>Bozena > >> > >>Bozena Zdaniuk, Ph.D. > >> > >>University of Pittsburgh > >> > >>UCSUR, 6th Fl. > >> > >>121 University Place > >> > >>Pittsburgh, PA 15260 > >> > >>Ph.: 412-624-5736 > >> > >>Fax: 412-624-4810 > >> > >>email: bozena@pitt.edu > >> > >> > >>-----Original Message----- > >>From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > On Behalf > >>Of Martin Sherman > >>Sent: Thursday, September 15, 2005 10:53 AM > >>To: SPSSX-L@LISTSERV.UGA.EDU > >>Subject: Re: converting a missing time value into a real time value > >> > >>Hector: Thanks for the response. I tried using q1=9:00. and got an > >>error message that the : was creating a problem. I will > >>wait to see if > >>anyone else responds. thanks again. martin > >> > >> > >> > >> > >>>>>"Hector Maletta" <hmaletta@fibertel.com.ar> 09/15/05 10:49 AM >>> > >>>>> > >>>>> > >>If you have only one case in that situation, you can issue > the command > >>as > >>follows: > >> > >>IF (SYSMIS(Q1))q1=9:00. > >> > >>But you may have other cases with the time missing, and you may not > >>want all of them assigned 9 hours. If it is so, you better allocate > >>the new value based on the relevant case's ID: > >> > >>If (ID=93869)q1=9:00. > >> > >>I have not considered possible further complications > arising from the > >>fact that your variable is a date variable (in hours). The simple > >>statement q=9:00 may not work, but I don't have time to go > into that > >>now. > >> > >>Hector > >> > >> > >> > >>>-----Original Message----- > >>>From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] > >>> > >>> > >>On Behalf > >> > >> > >>>Of Martin Sherman > >>>Sent: Thursday, September 15, 2005 11:43 AM > >>>To: SPSSX-L@LISTSERV.UGA.EDU > >>>Subject: converting a missing time value into a real time value > >>> > >>>Dear list, I have a time variable (HH.MM.SS) where a > subject has it > >>>missing (just a dot in the data field). I want to change > >>> > >>> > >>it to 9:00. > >> > >> > >>>I have tried the following but it doesn't work. > >>>if q1 eq 52724937 @55time_ eq 9:00. > >>>execute. > >>>and I have tried the following and it doesnot work. > >>>Is there a way I should be represented 9:00. Advice appreciated. > >>> > >>>if q1 eq 52724937 @55btime_ eq 9. > >>>execute. > >>> > >>>martin sherman > >>> > >>>__________ Informacisn de NOD32 1.1217 (20050914) __________ > >>> > >>>Este mensaje ha sido analizado con NOD32 Antivirus System > >>>http://www.nod32.com > >>> > >>> > >>> > >>> > >>__________ Información de NOD32 1.1217 (20050914) __________ > >> > >>Este mensaje ha sido analizado con NOD32 Antivirus System > >>http://www.nod32.com > >> > > __________ Información de NOD32 1.1218 (20050915) __________ > > Este mensaje ha sido analizado con NOD32 Antivirus System > http://www.nod32.com > >


Back to: Top of message | Previous page | Main SPSSX-L page