late afternoon spss entertainment
Locked
late afternoon spss entertainment
 
|
Let’s say. 1.
logistic regression mdata14b with ager/enter ager/print summary.
2.
logistic regression mdata14b with gaf/enter ager/print summary.
3.
logistic regression mdata14b with gaf/enter gaf/print summary.
I had always assumed that command syntax was parsed against the published documentation; thus ‘variable’ was a required word.
(Everybody has contrary examples, I know.) Nice to know it’s not—less typing. But isn’t example 2 surprising? I would have expected that the union of the variable lists for the enter subcommands would be either a subset or a perfect subset of the ‘variables=’
list. Thus example 2 should have pitched itself back in my face as if to say “what a crappy editor you were”. Word: True. But what do those coefficients represent?
Gene Maguin |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Re: late afternoon spss entertainment
Administrator
|
Good morning Gene. I believe the issue is that your various models are not all being estimated with the same cases due to missing data on some of the variables (and listwise deletion). See below for an example using one of the sample datasets that comes with SPSS. HTH.
* Change path to where sample files are stored on your computer. GET FILE = "C:\SPSSdata\survey_sample.sav". FREQUENCIES sex. DESCRIPTIVES age educ. COMPUTE male = sex EQ 1. * Here are Gene's first two examples: * logistic regression mdata14b with ager/enter ager/print summary. * logistic regression mdata14b with gaf/enter ager/print summary. * In my examples, I will make the following substitutions: - male in place of mdata14b (the DV) - age in place of ager - educ in place of gaf. logistic regression male with age/enter age/print summary. logistic regression male with educ/enter age/print summary. * As in Gene's examples, the coefficient tables are not the same. * But pay attention to the Case Processing Summary, and notice * that the number of cases included in the analysis is not the * same for the two models. The first model includes everyone * with valid data for male and age; the second includes everyone * with valid data for those two variables AND educ--so a smaller N. * If I set a filter for having valid data on all 3 variables, and * then estimate the two models, I should get the same results * from both of them. COMPUTE All3 = NMISS(male,age,educ) EQ 0. FILTER BY All3. logistic regression variables = male with age/method enter age/print summary. logistic regression male with educ/enter age/print summary. * As expected, when the two models are estimated using the * same cases, the coefficients and SEs are the same. USE ALL. FILTER OFF.
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
Re: late afternoon spss entertainment
|
|
I agree. There was missing data. Thanks. I was so stunned (and amused) that it worked at all.
-----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Bruce Weaver Sent: Friday, August 21, 2015 9:17 AM To: [hidden email] Subject: Re: late afternoon spss entertainment Good morning Gene. I believe the issue is that your various models are not all being estimated with the same cases due to missing data on some of the variables (and listwise deletion). See below for an example using one of the sample datasets that comes with SPSS. HTH. * Change path to where sample files are stored on your computer. GET FILE = "C:\SPSSdata\survey_sample.sav". FREQUENCIES sex. DESCRIPTIVES age educ. COMPUTE male = sex EQ 1. * Here are Gene's first two examples: * logistic regression mdata14b with ager/enter ager/print summary. * logistic regression mdata14b with gaf/enter ager/print summary. * In my examples, I will make the following substitutions: - male in place of mdata14b (the DV) - age in place of ager - educ in place of gaf. logistic regression male with age/enter age/print summary. logistic regression male with educ/enter age/print summary. * As in Gene's examples, the coefficient tables are not the same. * But pay attention to the Case Processing Summary, and notice * that the number of cases included in the analysis is not the * same for the two models. The first model includes everyone * with valid data for male and age; the second includes everyone * with valid data for those two variables AND educ--so a smaller N. * If I set a filter for having valid data on all 3 variables, and * then estimate the two models, I should get the same results * from both of them. COMPUTE All3 = NMISS(male,age,educ) EQ 0. FILTER BY All3. logistic regression variables = male with age/method enter age/print summary. logistic regression male with educ/enter age/print summary. * As expected, when the two models are estimated using the * same cases, the coefficients and SEs are the same. USE ALL. FILTER OFF. Maguin, Eugene wrote > Let's say. > > > 1. logistic regression mdata14b with ager/enter ager/print summary. > > Variables in the Equation > > > > B > > S.E. > > Wald > > df > > Sig. > > Exp(B) > > Step 1a > > ager > > .226 > > .099 > > 5.176 > > 1 > > .023 > > 1.254 > > Constant > > -2.651 > > 1.409 > > 3.540 > > 1 > > .060 > > .071 > > a. Variable(s) entered on step 1: ager. > > > > 2. logistic regression mdata14b with gaf/enter ager/print summary. > > Variables in the Equation > > > > B > > S.E. > > Wald > > df > > Sig. > > Exp(B) > > Step 1a > > ager > > .210 > > .107 > > 3.863 > > 1 > > .049 > > 1.234 > > Constant > > -2.549 > > 1.510 > > 2.848 > > 1 > > .091 > > .078 > > a. Variable(s) entered on step 1: ager. > > > > 3. logistic regression mdata14b with gaf/enter gaf/print summary. > > Variables in the Equation > > > > B > > S.E. > > Wald > > df > > Sig. > > Exp(B) > > Step 1a > > gaf > > -.025 > > .018 > > 1.886 > > 1 > > .170 > > .975 > > Constant > > 1.512 > > .837 > > 3.262 > > 1 > > .071 > > 4.537 > > a. Variable(s) entered on step 1: gaf. > > I had always assumed that command syntax was parsed against the > published documentation; thus 'variable' was a required word. > (Everybody has contrary examples, I know.) Nice to know it's not-less > typing. But isn't example 2 surprising? I would have expected that the > union of the variable lists for the enter subcommands would be either > a subset or a perfect subset of the 'variables=' list. Thus example 2 > should have pitched itself back in my face as if to say "what a crappy editor you were". Word: True. > But what do those coefficients represent? > Gene Maguin > > > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (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 [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://spssx-discussion.1045642.n5.nabble.com/late-afternoon-spss-entertainment-tp5730488p5730493.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 |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |