SPSSX Discussion

Correspondence analysis - factor analysis

Classic

List

Threaded

7 messages Options

Samir Omerovic

Correspondence analysis - factor analysis

Dear all,

I have two question regarding some analysis. Let me briefly state my case. I have some 2000
respondents whom I asked to evaluate three (3) companies. I have 24 characteristics like 'for
business people', 'flexible and fair' etc. and each of respondent gave an answer for each of three
companies for each characteristics on the scale from 1 to 5 where 1 means 'not appropriate for the
company' and 5 means 'totally appropriate for the company'.

Now I would like to run factor analysis but I have problem because I have to include one person
three times, for one did answer for each of three companies.

I was suggested by a friend to use Correspondence analysis and to use Means instead of Frequencies
since I obviously do not have frequencies as such. Is this valid to do? I must say that I tried this
approach and the results are interpretable. I am just suspicious about the whole deal with using
Means.

Now the question is should I use Factor analysis or Correspondence.

Thanks in advance

Samir

GfK BH

Centar za istrazivanje trzista i javnog mnijenja d.o.o.

Samir Omerović

Researcher

71 000 Sarajevo, Skenderija 44

Bosna i Hercegovina

Tel: +387 33 550 300

Fax: +387 33 444 226

www.gfk.ba

www.gfk.com

-----------------------------------------

This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are
not the intended recipient (or have received this e-mail in error) please notify the sender
immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the
material in this e-mail is strictly forbidden.

Ova e-poruka sadrzi povjerljive i/ili povlastene podatke. Ako niste osoba naznacena kao primalac
molimo odmah upozoriti posiljaoca i unistiti primljenu e-poruku. Neautorizirano kopiranje,
objavljivanje ili distribucija sadrzaja e-poruke striktno je zabranjeno.

---------------------------------------

Samir Omerovic

Re: Correspondence analysis - factor analysis

Hi Jan,

OK. The purpose of the analysis is to get the image of these three companies, to answer the question
"Which characteristics/statements best describe each of companies?" Simple as that.

Samir

-----Original Message-----
From: Spousta Jan [mailto:[hidden email]]
Sent: Thursday, February 15, 2007 12:41 PM
To: Samir Omerović
Subject: RE: Correspondence analysis - factor analysis

Hi Samir,

Could you please tell us also what is the purpose of your analyzis? What does your boss / your
customer expect to get?

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 12:08 PM
To: [hidden email]
Subject: Correspondence analysis - factor analysis

Dear all,

I have two question regarding some analysis. Let me briefly state my case. I have some 2000
respondents whom I asked to evaluate three (3) companies. I have 24 characteristics like 'for
business people', 'flexible and fair' etc. and each of respondent gave an answer for each of three
companies for each characteristics on the scale from 1 to 5 where 1 means 'not appropriate for the
company' and 5 means 'totally appropriate for the company'.

Now I would like to run factor analysis but I have problem because I have to include one person
three times, for one did answer for each of three companies.

I was suggested by a friend to use Correspondence analysis and to use Means instead of Frequencies
since I obviously do not have frequencies as such. Is this valid to do? I must say that I tried this
approach and the results are interpretable. I am just suspicious about the whole deal with using
Means.

Now the question is should I use Factor analysis or Correspondence.

Thanks in advance

Samir

GfK BH

Centar za istrazivanje trzista i javnog mnijenja d.o.o.

Samir Omerović

Researcher

71 000 Sarajevo, Skenderija 44

Bosna i Hercegovina

Tel: +387 33 550 300

Fax: +387 33 444 226

www.gfk.ba

www.gfk.com

-----------------------------------------

This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are
not the intended recipient (or have received this e-mail in error) please notify the sender
immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the
material in this e-mail is strictly forbidden.

Ova e-poruka sadrzi povjerljive i/ili povlastene podatke. Ako niste osoba naznacena kao primalac
molimo odmah upozoriti posiljaoca i unistiti primljenu e-poruku. Neautorizirano kopiranje,
objavljivanje ili distribucija sadrzaja e-poruke striktno je zabranjeno.

---------------------------------------

Spousta Jan

Re: Correspondence analysis - factor analysis

In reply to this post by Samir Omerovic

Then you do not need factor nor correspondence analysis. Compute averages of answers (use DESCRIPTIVES) and present them in graphs and/or tables.

An example:

For business people:
1|-------C1-|----C2----|----------|--C3------|5

Flexible and fair:
1|---------C2----------|-C1----C3-|----------|5

Etc.

And then interpret it in words (which are weaknesses and strenghts of each company, which is their overall performance, which groups of respondents prefer what - use ANOVA here if appropriate - etc.)

(I know that they are all ordinal but I do not care about it and few people does; Diana please forgive me.)

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 12:50 PM
To: [hidden email]
Subject: Re: Correspondence analysis - factor analysis

Hi Jan,

OK. The purpose of the analysis is to get the image of these three companies, to answer the question "Which characteristics/statements best describe each of companies?" Simple as that.

Samir

-----Original Message-----
From: Spousta Jan [mailto:[hidden email]]
Sent: Thursday, February 15, 2007 12:41 PM
To: Samir Omerović
Subject: RE: Correspondence analysis - factor analysis

Hi Samir,

Could you please tell us also what is the purpose of your analyzis? What does your boss / your customer expect to get?

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 12:08 PM
To: [hidden email]
Subject: Correspondence analysis - factor analysis

Dear all,

I have two question regarding some analysis. Let me briefly state my case. I have some 2000 respondents whom I asked to evaluate three (3) companies. I have 24 characteristics like 'for business people', 'flexible and fair' etc. and each of respondent gave an answer for each of three companies for each characteristics on the scale from 1 to 5 where 1 means 'not appropriate for the company' and 5 means 'totally appropriate for the company'.

Now I would like to run factor analysis but I have problem because I have to include one person three times, for one did answer for each of three companies.

I was suggested by a friend to use Correspondence analysis and to use Means instead of Frequencies since I obviously do not have frequencies as such. Is this valid to do? I must say that I tried this approach and the results are interpretable. I am just suspicious about the whole deal with using Means.

Now the question is should I use Factor analysis or Correspondence.

Thanks in advance

Samir

GfK BH

Centar za istrazivanje trzista i javnog mnijenja d.o.o.

Samir Omerović

Researcher

71 000 Sarajevo, Skenderija 44

Bosna i Hercegovina

Tel: +387 33 550 300

Fax: +387 33 444 226

www.gfk.ba

www.gfk.com

-----------------------------------------

This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.

Ova e-poruka sadrzi povjerljive i/ili povlastene podatke. Ako niste osoba naznacena kao primalac molimo odmah upozoriti posiljaoca i unistiti primljenu e-poruku. Neautorizirano kopiranje, objavljivanje ili distribucija sadrzaja e-poruke striktno je zabranjeno.

---------------------------------------

T.S. Lim

Re: Correspondence analysis - factor analysis

In reply to this post by Samir Omerovic

Dear Samir:

For your problem, I would use biplot.

http://www.jstatsoft.org/v07/i05/

During my years with Research International, I always used biplot. You may
want to check BrandMap too.

Regards,
T.S. Lim

> Hi Jan,
>
> OK. The purpose of the analysis is to get the image of these three
> companies, to answer the question
> "Which characteristics/statements best describe each of companies?" Simple
> as that.
>
> Samir
>
>
>
> -----Original Message-----
> From: Spousta Jan [mailto:[hidden email]]
> Sent: Thursday, February 15, 2007 12:41 PM
> To: Samir Omeroviæ
> Subject: RE: Correspondence analysis - factor analysis
>
> Hi Samir,
>
> Could you please tell us also what is the purpose of your analyzis? What
> does your boss / your
> customer expect to get?
>
> Jan
>
> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
> Samir Omeroviæ
> Sent: Thursday, February 15, 2007 12:08 PM
> To: [hidden email]
> Subject: Correspondence analysis - factor analysis
>
> Dear all,
>
>
>
> I have two question regarding some analysis. Let me briefly state my case.
> I have some 2000
> respondents whom I asked to evaluate three (3) companies. I have 24
> characteristics like 'for
> business people', 'flexible and fair' etc. and each of respondent gave an
> answer for each of three
> companies for each characteristics on the scale from 1 to 5 where 1 means
> 'not appropriate for the
> company' and 5 means 'totally appropriate for the company'.
>
>
>
> Now I would like to run factor analysis but I have problem because I have
> to include one person
> three times, for one did answer for each of three companies.
>
> I was suggested by a friend to use Correspondence analysis and to use
> Means instead of Frequencies
> since I obviously do not have frequencies as such. Is this valid to do? I
> must say that I tried this
> approach and the results are interpretable. I am just suspicious about the
> whole deal with using
> Means.
>
>
>
> Now the question is should I use Factor analysis or Correspondence.
>
>
>
> Thanks in advance
>
>
>
> Samir
>
>
>
>
>
> GfK BH
>
> Centar za istrazivanje trzista i javnog mnijenja d.o.o.
>
>
>
> Samir Omeroviæ
>
> Researcher
>
>
>
> 71 000 Sarajevo, Skenderija 44
>
> Bosna i Hercegovina
>
>
>
>
>
> Tel: +387 33 550 300
>
> Fax: +387 33 444 226
>
> www.gfk.ba
>
> www.gfk.com
>
>
>
>
>
>
>
> -----------------------------------------
>
> This e-mail (and any attachment/s) contains confidential and/or privileged
> information. If you are
> not the intended recipient (or have received this e-mail in error) please
> notify the sender
> immediately and destroy this e-mail. Any unauthorized copying,
> disclosure or distribution of the
> material in this e-mail is strictly forbidden.
>
> Ova e-poruka sadrzi povjerljive i/ili povlastene podatke. Ako niste osoba
> naznacena kao primalac
> molimo odmah upozoriti posiljaoca i unistiti primljenu e-poruku.
> Neautorizirano kopiranje,
> objavljivanje ili distribucija sadrzaja e-poruke striktno je zabranjeno.
>
> ---------------------------------------
>

Samir Omerovic

Re: Correspondence analysis - factor analysis

In reply to this post by Spousta Jan

Jan,

I have calculated the averages of answers. That is the first thing I did. I was hoping I can do
something else with those averages rather than just interpret it in words.

Users of one of each of these companies give better answers for their company. I want to see what is
C1 (company 1) users perception of each of these three companies and so on. I also have some
non-users and want to see their perception also. I do not get much difference in results for users
of C1, C2 and C3 when it comes to their companies.
Just trying to find something else than Averages I guess.

Samir

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Spousta Jan
Sent: Thursday, February 15, 2007 1:06 PM
To: [hidden email]
Subject: Re: Correspondence analysis - factor analysis

Then you do not need factor nor correspondence analysis. Compute averages of answers (use
DESCRIPTIVES) and present them in graphs and/or tables.

An example:

For business people:
1|-------C1-|----C2----|----------|--C3------|5

Flexible and fair:
1|---------C2----------|-C1----C3-|----------|5

Etc.

And then interpret it in words (which are weaknesses and strenghts of each company, which is their
overall performance, which groups of respondents prefer what - use ANOVA here if appropriate - etc.)

(I know that they are all ordinal but I do not care about it and few people does; Diana please
forgive me.)

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 12:50 PM
To: [hidden email]
Subject: Re: Correspondence analysis - factor analysis

Hi Jan,

OK. The purpose of the analysis is to get the image of these three companies, to answer the question
"Which characteristics/statements best describe each of companies?" Simple as that.

Samir

-----Original Message-----
From: Spousta Jan [mailto:[hidden email]]
Sent: Thursday, February 15, 2007 12:41 PM
To: Samir Omerović
Subject: RE: Correspondence analysis - factor analysis

Hi Samir,

Could you please tell us also what is the purpose of your analyzis? What does your boss / your
customer expect to get?

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 12:08 PM
To: [hidden email]
Subject: Correspondence analysis - factor analysis

Dear all,

I have two question regarding some analysis. Let me briefly state my case. I have some 2000
respondents whom I asked to evaluate three (3) companies. I have 24 characteristics like 'for
business people', 'flexible and fair' etc. and each of respondent gave an answer for each of three
companies for each characteristics on the scale from 1 to 5 where 1 means 'not appropriate for the
company' and 5 means 'totally appropriate for the company'.

Now I would like to run factor analysis but I have problem because I have to include one person
three times, for one did answer for each of three companies.

I was suggested by a friend to use Correspondence analysis and to use Means instead of Frequencies
since I obviously do not have frequencies as such. Is this valid to do? I must say that I tried this
approach and the results are interpretable. I am just suspicious about the whole deal with using
Means.

Now the question is should I use Factor analysis or Correspondence.

Thanks in advance

Samir

GfK BH

Centar za istrazivanje trzista i javnog mnijenja d.o.o.

Samir Omerović

Researcher

71 000 Sarajevo, Skenderija 44

Bosna i Hercegovina

Tel: +387 33 550 300

Fax: +387 33 444 226

www.gfk.ba

www.gfk.com

-----------------------------------------

This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are
not the intended recipient (or have received this e-mail in error) please notify the sender
immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the
material in this e-mail is strictly forbidden.

Ova e-poruka sadrzi povjerljive i/ili povlastene podatke. Ako niste osoba naznacena kao primalac
molimo odmah upozoriti posiljaoca i unistiti primljenu e-poruku. Neautorizirano kopiranje,
objavljivanje ili distribucija sadrzaja e-poruke striktno je zabranjeno.

---------------------------------------

Spousta Jan

Re: Correspondence analysis - factor analysis

In reply to this post by Samir Omerovic

Samir,

Then split the file in four groups - users C1, users C2, users C3, others - and compare the means in and between groups (MEANS will help you to calculate it easily). Use ANOVAs and paired sample t-tests for testing differences (or their nonparametric counterparts if you wish to be exact).

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 1:54 PM
To: [hidden email]
Subject: Re: Correspondence analysis - factor analysis

Jan,

I have calculated the averages of answers. That is the first thing I did. I was hoping I can do something else with those averages rather than just interpret it in words.

Users of one of each of these companies give better answers for their company. I want to see what is
C1 (company 1) users perception of each of these three companies and so on. I also have some non-users and want to see their perception also. I do not get much difference in results for users of C1, C2 and C3 when it comes to their companies.
Just trying to find something else than Averages I guess.

Samir

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Spousta Jan
Sent: Thursday, February 15, 2007 1:06 PM
To: [hidden email]
Subject: Re: Correspondence analysis - factor analysis

Then you do not need factor nor correspondence analysis. Compute averages of answers (use
DESCRIPTIVES) and present them in graphs and/or tables.

An example:

For business people:
1|-------C1-|----C2----|----------|--C3------|5

Flexible and fair:
1|---------C2----------|-C1----C3-|----------|5

Etc.

And then interpret it in words (which are weaknesses and strenghts of each company, which is their overall performance, which groups of respondents prefer what - use ANOVA here if appropriate - etc.)

(I know that they are all ordinal but I do not care about it and few people does; Diana please forgive me.)

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 12:50 PM
To: [hidden email]
Subject: Re: Correspondence analysis - factor analysis

Hi Jan,

OK. The purpose of the analysis is to get the image of these three companies, to answer the question "Which characteristics/statements best describe each of companies?" Simple as that.

Samir

-----Original Message-----
From: Spousta Jan [mailto:[hidden email]]
Sent: Thursday, February 15, 2007 12:41 PM
To: Samir Omerović
Subject: RE: Correspondence analysis - factor analysis

Hi Samir,

Could you please tell us also what is the purpose of your analyzis? What does your boss / your customer expect to get?

Jan

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Samir Omerović
Sent: Thursday, February 15, 2007 12:08 PM
To: [hidden email]
Subject: Correspondence analysis - factor analysis

Dear all,

I have two question regarding some analysis. Let me briefly state my case. I have some 2000 respondents whom I asked to evaluate three (3) companies. I have 24 characteristics like 'for business people', 'flexible and fair' etc. and each of respondent gave an answer for each of three companies for each characteristics on the scale from 1 to 5 where 1 means 'not appropriate for the company' and 5 means 'totally appropriate for the company'.

Now I would like to run factor analysis but I have problem because I have to include one person three times, for one did answer for each of three companies.

I was suggested by a friend to use Correspondence analysis and to use Means instead of Frequencies since I obviously do not have frequencies as such. Is this valid to do? I must say that I tried this approach and the results are interpretable. I am just suspicious about the whole deal with using Means.

Now the question is should I use Factor analysis or Correspondence.

Thanks in advance

Samir

GfK BH

Centar za istrazivanje trzista i javnog mnijenja d.o.o.

Samir Omerović

Researcher

71 000 Sarajevo, Skenderija 44

Bosna i Hercegovina

Tel: +387 33 550 300

Fax: +387 33 444 226

www.gfk.ba

www.gfk.com

-----------------------------------------

This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.

Ova e-poruka sadrzi povjerljive i/ili povlastene podatke. Ako niste osoba naznacena kao primalac molimo odmah upozoriti posiljaoca i unistiti primljenu e-poruku. Neautorizirano kopiranje, objavljivanje ili distribucija sadrzaja e-poruke striktno je zabranjeno.

---------------------------------------

Kooij, A.J. van der

Re: Correspondence analysis - factor analysis

In reply to this post by T.S. Lim

The biplot can be obtained with Correspondence (plot option "rows and columns").

Using means is okay. Although a frequency table is the most common type of correspondence table, it is not the only type. A correspondence table is a table whose cells contain some measure of correspondence between the rows and columns. The measure can be any indication of similarity, association, interaction, etc. between the row and column variables.

Regards,

Anita van der Kooij
Data Theory Group
Leiden University

________________________________

From: SPSSX(r) Discussion on behalf of T.S. Lim
Sent: Thu 15/02/2007 13:20
To: [hidden email]
Subject: Re: Correspondence analysis - factor analysis

Dear Samir:

For your problem, I would use biplot.

http://www.jstatsoft.org/v07/i05/

During my years with Research International, I always used biplot. You may
want to check BrandMap too.

Regards,
T.S. Lim

**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
**********************************************************************