Multivariate linear regression analysis

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

Re: Multivariate linear regression analysis

Postby GerineL » Fri Aug 22, 2014 9:54 am

As I indicated a few times before, it would be great if you could explain your variables.
You still have not done that, so it is hard for me to say whether or not education could be a continuous variable or not.
So please explain for each variable what the question was, and what the participants could answer.

I don't think you need to worry about them, but you should report them in your final report.
p_s_
Posts: 11
Joined: Mon Aug 04, 2014 4:15 am

Re: Multivariate linear regression analysis

Postby p_s_ » Sat Aug 23, 2014 4:45 am

Thanks GerineL,
GerineL wrote:As I indicated a few times before, it would be great if you could explain your variables.
You still have not done that, so it is hard for me to say whether or not education could be a continuous variable or not.
So please explain for each variable what the question was, and what the participants could answer.
Sorry, I did not do that earlier.

Education, Question was "What is your highest qualification" Possible answers were some middle school, some high school, high school, some college, graduated college, graduate school

Income, question was, "What is your household income" Answers were 0-15999, 16000-25999, 26000-35999 and so on till 90000

Age, What is your age, Choices were 18-24, 25-30, 31-39,.. and so on.

Now, can you please advise which independent variables should be continuous and which ones categorized?

I appreciate your assistance and time.
GerineL wrote: I don't think you need to worry about them, but you should report them in your final report.
Those topics are new to me, so if they are not relevant for the analysis at hand, I would rather skip them. Well, then I need to understand them and interpret it to explain it well.
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

Re: Multivariate linear regression analysis

Postby GerineL » Mon Aug 25, 2014 9:09 am

Ok, this explains a lot :-).

So in fact, you don't have scale variables as you indicated earlier, rather, you have ordinal variables.
A scale variable for age would have been:
Please indictae your age: ... (and they would just put their age in years). Now, your variables are not continuous.
That's a pity, but there is nothing you can do about that now.

I am not entirely sure what you did earlier, but in fact, ordered logistic regression is the way to go.
There is an example here:
http://www.ats.ucla.edu/stat/spss/dae/ologit.htm

There are lots of other casestudies online as well, I would suggest you'd take a look at those, and see if you can try to apply them to your own data. Please come back if something is still not clear.

as to reporting SE etc.: I don't know what you are doing these analyses for, but in scientific reports (papers and so forth) it is common to report these.
p_s_
Posts: 11
Joined: Mon Aug 04, 2014 4:15 am

Re: Multivariate linear regression analysis

Postby p_s_ » Tue Aug 26, 2014 2:50 am

Thanks GerineL,
GerineL wrote:Ok, this explains a lot :-).

So in fact, you don't have scale variables as you indicated earlier, rather, you have ordinal variables.
A scale variable for age would have been:
Please indictae your age: ... (and they would just put their age in years). Now, your variables are not continuous.
That's a pity, but there is nothing you can do about that now.
Do you think education variable can be categorical?
GerineL wrote: I am not entirely sure what you did earlier,
I actually did logit model http://www.ats.ucla.edu/stat/spss/dae/logit.htm or http://core.ecu.edu/psyc/wuenschk/MV/Mu ... c-SPSS.PDF because I was trying to find factors which influence usage of preventive insurance(free flu vaccines, routine checkups etc.) by people having insurance.

GerineL wrote: but in fact, ordered logistic regression is the way to go.
There is an example here:
http://www.ats.ucla.edu/stat/spss/dae/ologit.htm

There are lots of other casestudies online as well, I would suggest you'd take a look at those, and see if you can try to apply them to your own data. Please come back if something is still not clear.
Thanks, now since I need to find which factors(age, income, education) influence this, I will look up ordered logistic regression
GerineL wrote: as to reporting SE etc.: I don't know what you are doing these analyses for, but in scientific reports (papers and so forth) it is common to report these.
I can report, but anything I write in a report needs to be known well by me to tackle questions about it, if they arise(which is likely). As you can see, I am a beginner so if I am not using some method, I felt I could skip it.

Thanks so much for helping me.
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

Re: Multivariate linear regression analysis

Postby GerineL » Tue Aug 26, 2014 7:50 am

Look at this site to see differences between different types of variables.
http://onlinestatbook.com/2/introductio ... ement.html (or wikipedia)

As to skipping the SE etc: They are actually quite relevant, it is very uncommon to not report them. Look at some case studies though, you probably have lots of them in which each of those aspects is explained.
p_s_
Posts: 11
Joined: Mon Aug 04, 2014 4:15 am

Re: Multivariate linear regression analysis

Postby p_s_ » Wed Aug 27, 2014 4:26 am

Thanks GerineL,
GerineL wrote:Look at this site to see differences between different types of variables.
http://onlinestatbook.com/2/introductio ... ement.html (or wikipedia)

Thanks, the book seems suitable for a beginner like me.

As to skipping the SE etc: They are actually quite relevant, it is very uncommon to not report them. Look at some case studies though, you probably have lots of them in which each of those aspects is explained.
I will look them up.

I appreciate all your assistance and time.
Iguanahulu
Posts: 1
Joined: Sat Mar 12, 2016 10:41 pm

Multivariate linear regression analysis

Postby Iguanahulu » Sun Mar 13, 2016 1:25 pm

Hi

I am trying to do Cox multivariate regression analysis in a project where I am trying to identify predictors of a particular disease using variables such age, X- ray findings, and other identified known risk factors, and I am wondering how to perform such an analysis using excel, or even online calculators.
I am not sure how to figure out the x and y of the formula , not sure where to start ? You can tell I never did such an advanced analysis, but this is what the reviewers wanted

Thanks
GerineL
Moderator
Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

Re: Multivariate linear regression analysis

Postby GerineL » Mon Mar 14, 2016 8:50 am

If you want to use excel, it might make sense to ask this question on an excel forum rather than an spss forum

Who is online

Users browsing this forum: No registered users and 2 guests

cron