Re: binary logistic regression
Locked
 
|
Thanks for the compliment on the site. Nice
to know someone appreciates it.
I'm not a statistician, just a humble social
researcher. It's more than 30
years since I used any kind of regression or
factor analysis for serious analysis (See Hall & Ring 1974, Hall 1975 under
Publications) and I have no tutorials that go this far,
except a simple explanation of linear regression and
correlation. I'm copying this to the list as
there's bound to be someone who can be of more help (eg ViAnn Beadle, Jon
Peck, Bruce Weaver).
Are you trying to classify the most accident-prone
drivers from the other data or something else? There are some very good
sites with SPSS tutorials. Bruce Weaver has a very useful list on http://sites.google.com/a/lakeheadu.ca/bweaver/Home/statistics/spss and
his other site http://sites.google.com:80/a/lakeheadu.ca/bweaver/ even
mentions safe driving.
I once had a survey of teenagers in a comprehensive
school and the gender question was at the end of the questionnaire.
One class had a disruption and none of the pupils got as far as that question,
but I used discriminant analysis (amongst other procedures) to
estimate gender and got something which allocated 100% of the other pupils to
the correct group.
|
Administrator
|
Hi Khizar. You might find David Garson's online notes helpful. Here's the page for logistic regression: http://faculty.chass.ncsu.edu/garson/PA765/logistic.htm Good luck.
--
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/). |
|
Add to that -- the case study available from Help>Case Studies. Click on
Binary Logistic Regression. -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Bruce Weaver Sent: Monday, October 26, 2009 7:13 AM To: [hidden email] Subject: Re: binary logistic regression John F Hall-2 wrote: > > Thanks for the compliment on the site. Nice to know someone appreciates > it. > > I'm not a statistician, just a humble social researcher. It's more than > 30 years since I used any kind of regression or factor analysis for > serious analysis (See Hall & Ring 1974, Hall 1975 under Publications) and > I have no tutorials that go this far, except a simple explanation of > linear regression and correlation. I'm copying this to the list as > there's bound to be someone who can be of more help (eg ViAnn Beadle, Jon > Peck, Bruce Weaver). > > Are you trying to classify the most accident-prone drivers from the other > data or something else? There are some very good sites with SPSS > tutorials. Bruce Weaver has a very useful list on > http://sites.google.com/a/lakeheadu.ca/bweaver/Home/statistics/spss and > his other site http://sites.google.com:80/a/lakeheadu.ca/bweaver/ even > mentions safe driving. > > I once had a survey of teenagers in a comprehensive school and the gender > question was at the end of the questionnaire. One class had a disruption > and none of the pupils got as far as that question, but I used > discriminant analysis (amongst other procedures) to estimate gender and > got something which allocated 100% of the other pupils to the correct > group. > > ----- Original Message ----- > From: Engr.Khizar Azam > To: [hidden email] > Sent: Monday, October 26, 2009 10:06 AM > Subject: binary logistic regression > > > > Dear Dr John F Hall hi, > I am a faculty member and a researcher doing research on driver > accident problem in Pakistan, we do not have many people who can do > data analysis in SPSS usually they can do descriptives, while > searching on internet i visited your site and am glad that you have a > very informative site which will benefit thousands of academic as well > as research fellows i appreciate that. > sir, as mentioned i have collected data of 300 drivers and now want to > do binary logistic regression analysis in SPSS to identify the factors > so just was curious if you can help me out in this connection, i would > be grateful to you. > thanks > profound regards, > > Khizar Azam > Assistant professor > N.W.F.P U.E.T Peshawar > > > Hi Khizar. You might find David Garson's online notes helpful. Here's the page for logistic regression: http://faculty.chass.ncsu.edu/garson/PA765/logistic.htm Good luck. ----- -- 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://www.nabble.com/Re%3A-binary-logistic-regression-tp26057535p26059453.h tml 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 |
|
|
In reply to this post by John F Hall
How can you be sure you did not overfit the analysis to the data Nana Sent from my iPhone
|
|
|
Not quite sure what you mean by this, but not many
boys agreed with all the pro-women items in the scale, or read every issue of
"Jackie". It was only a fun exercise for the class to demonstrate a few
statistical procedures which might have been included in a second level course
had I not taken early retirement. The formulation of the research question
was always far more important: it's only laterthat we look for statistical
procedures one might use when looking for an answer.
|
|
|
In reply to this post by John F Hall
Dear list,
I need some advice on the following:
I have been asked to run a t-test for a project I am working on and I want to know if a t-test is actually the most appropriate test for the situation,
my supervisor is interested in finding out whether there is a difference between men and women on the following question "Were you using your assistive device when you fell?", with the reponse choice being yes or no.
My question is, would I be correct in thinking that a t-test is not appropriate for this type of question, because both variables are dichotomous?
Any advice would be greatly appreciated.
Stace |
Re: Question about running a t-test
|
|
Usually one would do a chi square test for this type of data. From: SPSSX(r) Discussion
[mailto:[hidden email]] On Behalf Of stace swayne Dear list, I need some advice on the following: I have been asked to run a t-test for a project I am working
on and I want to know if a t-test is actually the most appropriate test for the
situation, my supervisor is interested in finding out whether there is
a difference between men and women on the following question "Were you
using your assistive device when you fell?", with the reponse choice being
yes or no. My question is, would I be correct in thinking that a t-test
is not appropriate for this type of question, because both variables are dichotomous? Any advice would be greatly appreciated. Stace |
Re: Question about running a t-test
|
|
yes your apprehension is correct. t-test is used to see the difference between two means. you can check chi-square or p-test. Chi-square is more relevant where as p-test is used to check the difference between two proportions. you need to use p-test for two independent sample sets.
HTH.
On Fri, Nov 13, 2009 at 7:02 PM, David Wasserman <[hidden email]> wrote:
-- Samir Paul 382 Main Street, TORONTO ON M4C4X8 CANADA Phone: ( 001 ) 416 686 9958 |
Re: Question about running a t-test
|
|
In reply to this post by stace swayne
Hi Stace:
stace swayne wrote: > > I have been asked to run a t-test for a project I am working on and I > want to know if a t-test is actually the most appropriate test for the > situation, > > my supervisor is interested in finding out whether there is a > difference between men and women on the following question "Were you > using your assistive device when you fell?", with the reponse choice > being yes or no. > > My question is, would I be correct in thinking that a t-test is not > appropriate for this type of question, because both variables are > dichotomous? (chi-square) test. A t test needs a quantitative outcome (an normally distributed, besides). HTH, Marta García-Granero -- For miscellaneous SPSS related statistical stuff, visit: http://gjyp.nl/marta/ ===================== 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 |
Re: Question about running a t-test
|
|
While a chi-square test would be fairly common in this case it is perfectly okay to conduct a two-sample t-test if you have a large enough sample.
In this case you are asking whether the proportion of those using an assistive device differs between men and women. Simply code "no" as "0" and "yes" as "1." The means will then be the proportion of "yes" responses and the t-test will give you an appropriate answer, with sufficient sample size (n's > ~30).
Michael
**************************************************** Michael Granaas [hidden email] Assoc. Prof. Phone: 605 677 5295 Dept. of Psychology FAX: 605 677 3195 University of South Dakota 414 E. Clark St. Vermillion, SD 57069 ***************************************************** From: SPSSX(r) Discussion [[hidden email]] On Behalf Of García-Granero [[hidden email]] Sent: Saturday, November 14, 2009 4:31 AM To: [hidden email] Subject: Re: Question about running a t-test Hi Stace:
stace swayne wrote: > > I have been asked to run a t-test for a project I am working on and I > want to know if a t-test is actually the most appropriate test for the > situation, > > my supervisor is interested in finding out whether there is a > difference between men and women on the following question "Were you > using your assistive device when you fell?", with the reponse choice > being yes or no. > > My question is, would I be correct in thinking that a t-test is not > appropriate for this type of question, because both variables are > dichotomous? (chi-square) test. A t test needs a quantitative outcome (an normally distributed, besides). HTH, Marta García-Granero -- For miscellaneous SPSS related statistical stuff, visit: http://gjyp.nl/marta/ ===================== 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 |