This is not an SPSS question
Locked
 
|
WARNING: Not an SPSS question. I’m looking for a small data set and step-by-step procedure to do a factor analysis by hand. It’s important that the procedure does not use matrix operations.
A colleague asked me, if I had a simple hand calculated example of factor analysis that he could use with his stats class. I suggested Harman, but he did want matrix operations. I remember working through an
analysis by hand (decades ago!) using centroid method, but I don’t remember whether I used matrices or not. Is there a simple examine online or in a textbook that I can recommend to my colleague? TIA, Stephen Salbod, Pace University, NYC
|
|
How many cases do you need, and how many variables? I have a data set on my website derived from a short self-completion questionnaire, but the variables are ordinal rather than scale. Otherwise you could always plunder a subset of the Quality of Life surveys. John Hall From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Salbod, Mr. Stephen WARNING: Not an SPSS question. I’m looking for a small data set and step-by-step procedure to do a factor analysis by hand. It’s important that the procedure does not use matrix operations. A colleague asked me, if I had a simple hand calculated example of factor analysis that he could use with his stats class. I suggested Harman, but he did want matrix operations. I remember working through an analysis by hand (decades ago!) using centroid method, but I don’t remember whether I used matrices or not. Is there a simple examine online or in a textbook that I can recommend to my colleague? TIA, Stephen Salbod, Pace University, NYC |
|
|
Dear John, Thank you for the offer. My primary concern right now is to get the recipe for a matrix free procedure. --Steve From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of John F Hall How many cases do you need, and how many variables? I have a data set on my website derived from a short self-completion questionnaire, but the variables are ordinal rather than scale. Otherwise
you could always plunder a subset of the Quality of Life surveys. John Hall From: SPSSX(r) Discussion [mailto:[hidden email]]
On Behalf Of Salbod, Mr. Stephen WARNING: Not an SPSS question. I’m looking for a small data set and step-by-step procedure to do a factor analysis by hand. It’s important that the procedure does not use matrix operations.
A colleague asked me, if I had a simple hand calculated example of factor analysis that he could use with his stats class. I suggested Harman, but he did want matrix operations. I remember working through an
analysis by hand (decades ago!) using centroid method, but I don’t remember whether I used matrices or not. Is there a simple examine online or in a textbook that I can recommend to my colleague? TIA, Stephen Salbod, Pace University, NYC
|
Administrator
|
In reply to this post by Salbod
"It's important that the procedure does not use matrix operations."
WHY?? Scalar expressions for such things are terribly cumbersome! You might sarch the web for algos to extract eigenvalues/vectors and convert that. OTOH: As soon as you begin to describe MVA to people you REALLY need matrices!!!!!!!
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
I was puzzled too!
I did ask him to consider matrices, but he reiterated the procedure must not use matrices. He wants to use it in a stats class and felt the matrix operations would consume too much time. --Steve -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso Sent: Wednesday, April 20, 2011 7:53 PM To: [hidden email] Subject: Re: This is not an SPSS question "It's important that the procedure does not use matrix operations." WHY?? Scalar expressions for such things are terribly cumbersome! You might sarch the web for algos to extract eigenvalues/vectors and convert that. OTOH: As soon as you begin to describe MVA to people you REALLY need matrices!!!!!!! Salbod, Mr. Stephen wrote: > > WARNING: Not an SPSS question. > > I'm looking for a small data set and step-by-step procedure to do a > factor analysis by hand. It's important that the procedure does not > use matrix operations. > > A colleague asked me, if I had a simple hand calculated example of > factor analysis that he could use with his stats class. I suggested > Harman, but he did want matrix operations. I remember working through > an analysis by hand (decades ago!) using centroid method, but I don't > remember whether I used matrices or not. Is there a simple examine > online or in a textbook that I can recommend to my colleague? > > TIA, Stephen Salbod, Pace University, NYC > -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/This-is-not-an-SPSS-question-tp4315660p4329613.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 |
|
|
Your colleague is misguided. He should take this opportunity to
introduce the basics of linear algebra. My 2 cents. Ryan On Thu, Apr 21, 2011 at 7:38 AM, Salbod, Mr. Stephen <[hidden email]> wrote: > I was puzzled too! > I did ask him to consider matrices, but he reiterated the procedure must not use matrices. He wants to use it in a stats class and felt the matrix operations would consume too much time. > > --Steve > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso > Sent: Wednesday, April 20, 2011 7:53 PM > To: [hidden email] > Subject: Re: This is not an SPSS question > > "It's important that the procedure does not use matrix operations." > WHY?? Scalar expressions for such things are terribly cumbersome! > You might sarch the web for algos to extract eigenvalues/vectors and convert that. > OTOH: As soon as you begin to describe MVA to people you REALLY need matrices!!!!!!! > > > Salbod, Mr. Stephen wrote: >> >> WARNING: Not an SPSS question. >> >> I'm looking for a small data set and step-by-step procedure to do a >> factor analysis by hand. It's important that the procedure does not >> use matrix operations. >> >> A colleague asked me, if I had a simple hand calculated example of >> factor analysis that he could use with his stats class. I suggested >> Harman, but he did want matrix operations. I remember working through >> an analysis by hand (decades ago!) using centroid method, but I don't >> remember whether I used matrices or not. Is there a simple examine >> online or in a textbook that I can recommend to my colleague? >> >> TIA, Stephen Salbod, Pace University, NYC >> > > > -- > View this message in context: http://spssx-discussion.1045642.n5.nabble.com/This-is-not-an-SPSS-question-tp4315660p4329613.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 > ===================== 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 |
|
|
RB: I hear you. I will gently suggest that he might find it easier, for himself and the students, to cover factor analysis using matrices. --Steve
-----Original Message----- From: R B [mailto:[hidden email]] Sent: Thursday, April 21, 2011 9:31 AM To: Salbod, Mr. Stephen Cc: [hidden email] Subject: Re: This is not an SPSS question Your colleague is misguided. He should take this opportunity to introduce the basics of linear algebra. My 2 cents. Ryan On Thu, Apr 21, 2011 at 7:38 AM, Salbod, Mr. Stephen <[hidden email]> wrote: > I was puzzled too! > I did ask him to consider matrices, but he reiterated the procedure must not use matrices. He wants to use it in a stats class and felt the matrix operations would consume too much time. > > --Steve > > -----Original Message----- > From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf > Of David Marso > Sent: Wednesday, April 20, 2011 7:53 PM > To: [hidden email] > Subject: Re: This is not an SPSS question > > "It's important that the procedure does not use matrix operations." > WHY?? Scalar expressions for such things are terribly cumbersome! > You might sarch the web for algos to extract eigenvalues/vectors and convert that. > OTOH: As soon as you begin to describe MVA to people you REALLY need matrices!!!!!!! > > > Salbod, Mr. Stephen wrote: >> >> WARNING: Not an SPSS question. >> >> I'm looking for a small data set and step-by-step procedure to do a >> factor analysis by hand. It's important that the procedure does not >> use matrix operations. >> >> A colleague asked me, if I had a simple hand calculated example of >> factor analysis that he could use with his stats class. I suggested >> Harman, but he did want matrix operations. I remember working through >> an analysis by hand (decades ago!) using centroid method, but I don't >> remember whether I used matrices or not. Is there a simple examine >> online or in a textbook that I can recommend to my colleague? >> >> TIA, Stephen Salbod, Pace University, NYC >> > > > -- > View this message in context: > http://spssx-discussion.1045642.n5.nabble.com/This-is-not-an-SPSS-ques > tion-tp4315660p4329613.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 > ===================== 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 Ryan
IMO, some knowledge of matrix algebra is helpful if one wants a deeper
understanding of many of the commonly used statistical tests in the social sciences. I can't imagine having learned about the underpinnings of MANOVA, for instance, without some knowledge of matrix algebra. And it would be quite difficult to appreciate how the random effects ("G") and residual ("R") variance covariance matrices play a role in linear mixed models without understanding the basics of matrix algebra. Matrix algebra shows very nicely how linear mixed models are an extension of general linear models. In my very humble opinion, a firm grasp of matrix algebra is generally more important than even understanding advanced calculus in the application of statistics in the social sciences. I'm quite interested in hearing if people agree/disagree with my comments. Ryan On Thu, Apr 21, 2011 at 9:42 AM, Jon K Peck <[hidden email]> wrote: > To put this another way, if there is no time to teach some matrix algebra, > how can there be time for the tedious exercise of running factor analysis by > hand? IMO that sort of thing went out before 1970. (I remember doing this > with regression around 1965 - the "revised Doolittle method IIRC" , and I > certainly didn't learn anything about regression from that exercise!) I do > not believe that drudging through all the calculations manually adds the > slightest insight into the statistical technique. > > (So, the customer is always right - except when he/she isn't.) > > Jon Peck > Senior Software Engineer, IBM > [hidden email] > 312-651-3435 > > > > From: R B <[hidden email]> > To: [hidden email] > Date: 04/21/2011 07:33 AM > Subject: Re: [SPSSX-L] This is not an SPSS question > Sent by: "SPSSX(r) Discussion" <[hidden email]> > ________________________________ > > > Your colleague is misguided. He should take this opportunity to > introduce the basics of linear algebra. > > My 2 cents. > > Ryan > > On Thu, Apr 21, 2011 at 7:38 AM, Salbod, Mr. Stephen <[hidden email]> > wrote: >> I was puzzled too! >> I did ask him to consider matrices, but he reiterated the procedure must >> not use matrices. He wants to use it in a stats class and felt the matrix >> operations would consume too much time. >> >> --Steve >> >> -----Original Message----- >> From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of >> David Marso >> Sent: Wednesday, April 20, 2011 7:53 PM >> To: [hidden email] >> Subject: Re: This is not an SPSS question >> >> "It's important that the procedure does not use matrix operations." >> WHY?? Scalar expressions for such things are terribly cumbersome! >> You might sarch the web for algos to extract eigenvalues/vectors and >> convert that. >> OTOH: As soon as you begin to describe MVA to people you REALLY need >> matrices!!!!!!! >> >> >> Salbod, Mr. Stephen wrote: >>> >>> WARNING: Not an SPSS question. >>> >>> I'm looking for a small data set and step-by-step procedure to do a >>> factor analysis by hand. It's important that the procedure does not >>> use matrix operations. >>> >>> A colleague asked me, if I had a simple hand calculated example of >>> factor analysis that he could use with his stats class. I suggested >>> Harman, but he did want matrix operations. I remember working through >>> an analysis by hand (decades ago!) using centroid method, but I don't >>> remember whether I used matrices or not. Is there a simple examine >>> online or in a textbook that I can recommend to my colleague? >>> >>> TIA, Stephen Salbod, Pace University, NYC >>> >> >> >> -- >> View this message in context: >> http://spssx-discussion.1045642.n5.nabble.com/This-is-not-an-SPSS-question-tp4315660p4329613.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 >> > > ===================== > 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 |
|
|
RB: You comments are right on. I'm in complete agreement with you.
Before I received your email my colleague was thrilled to have found an answer to his problem: Guilford, J. P. (1936). Psychometric methods. New York: McGraw-Hill Book Co. I picked the book up at the University library. There is a chapter in it on Factor Analysis (p 457-516) where a two-factor problem is worked out using the centroid method. I'm working through it now, to see what kind of insight can gained from hand calculations (I'm cheating a little, I'm using Excel). You gave me an idea, this fall, I'm going to touch base with my friend and demonstrate FA using SPSS's matrix procedures. --Steve -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of R B Sent: Thursday, April 21, 2011 11:21 PM To: [hidden email] Subject: Re: This is not an SPSS question IMO, some knowledge of matrix algebra is helpful if one wants a deeper understanding of many of the commonly used statistical tests in the social sciences. I can't imagine having learned about the underpinnings of MANOVA, for instance, without some knowledge of matrix algebra. And it would be quite difficult to appreciate how the random effects ("G") and residual ("R") variance covariance matrices play a role in linear mixed models without understanding the basics of matrix algebra. Matrix algebra shows very nicely how linear mixed models are an extension of general linear models. In my very humble opinion, a firm grasp of matrix algebra is generally more important than even understanding advanced calculus in the application of statistics in the social sciences. I'm quite interested in hearing if people agree/disagree with my comments. Ryan On Thu, Apr 21, 2011 at 9:42 AM, Jon K Peck <[hidden email]> wrote: > To put this another way, if there is no time to teach some matrix > algebra, how can there be time for the tedious exercise of running > factor analysis by hand? IMO that sort of thing went out before 1970. > (I remember doing this with regression around 1965 - the "revised > Doolittle method IIRC" , and I certainly didn't learn anything about > regression from that exercise!) I do not believe that drudging > through all the calculations manually adds the slightest insight into the statistical technique. > > (So, the customer is always right - except when he/she isn't.) > > Jon Peck > Senior Software Engineer, IBM > [hidden email] > 312-651-3435 > > > > From: R B <[hidden email]> > To: [hidden email] > Date: 04/21/2011 07:33 AM > Subject: Re: [SPSSX-L] This is not an SPSS question > Sent by: "SPSSX(r) Discussion" <[hidden email]> > ________________________________ > > > Your colleague is misguided. He should take this opportunity to > introduce the basics of linear algebra. > > My 2 cents. > > Ryan > > On Thu, Apr 21, 2011 at 7:38 AM, Salbod, Mr. Stephen > <[hidden email]> > wrote: >> I was puzzled too! >> I did ask him to consider matrices, but he reiterated the procedure >> must not use matrices. He wants to use it in a stats class and felt >> the matrix operations would consume too much time. >> >> --Steve >> >> -----Original Message----- >> From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf >> Of David Marso >> Sent: Wednesday, April 20, 2011 7:53 PM >> To: [hidden email] >> Subject: Re: This is not an SPSS question >> >> "It's important that the procedure does not use matrix operations." >> WHY?? Scalar expressions for such things are terribly cumbersome! >> You might sarch the web for algos to extract eigenvalues/vectors and >> convert that. >> OTOH: As soon as you begin to describe MVA to people you REALLY need >> matrices!!!!!!! >> >> >> Salbod, Mr. Stephen wrote: >>> >>> WARNING: Not an SPSS question. >>> >>> I'm looking for a small data set and step-by-step procedure to do a >>> factor analysis by hand. It's important that the procedure does not >>> use matrix operations. >>> >>> A colleague asked me, if I had a simple hand calculated example of >>> factor analysis that he could use with his stats class. I suggested >>> Harman, but he did want matrix operations. I remember working >>> through an analysis by hand (decades ago!) using centroid method, >>> but I don't remember whether I used matrices or not. Is there a >>> simple examine online or in a textbook that I can recommend to my colleague? >>> >>> TIA, Stephen Salbod, Pace University, NYC >>> >> >> >> -- >> View this message in context: >> http://spssx-discussion.1045642.n5.nabble.com/This-is-not-an-SPSS-que >> stion-tp4315660p4329613.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 >> > > ===================== > 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 ===================== 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: This is not an SPSS question
Administrator
|
In reply to this post by Ryan
I agree with you, Ryan. I think an introduction to matrix algebra is an essential part of a graduate level course in applied statistics, and have many times found myself wishing it had been part of the grad stats course I took many years ago (in psychology). It would have been far more useful than the intro to calculus we were given.
Cheers, Bruce
--
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/). |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |