Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Commands to send factor analysis to another application


From   John Luke Gallup <[email protected]>
To   [email protected]
Subject   Re: st: Commands to send factor analysis to another application
Date   Wed, 5 Dec 2012 13:39:16 -0800

Kerry,

You can use the command -frmttable- to create the table you describe.  Commands like -outreg-, -outreg2-, and -estout- don't work because -factor- does not save results in the e(b) and e(V) matrices used by these programs.  -frmttable- takes results from a generic Stata matrix and creates a formatted table with them.  It is distributed with the -outreg- package.

The following code pieces together a table like the one you describe:

sysuse auto, clear
factor price-foreign, pcf
rotate, promax
predict f1 f2 f3
alpha price-foreign, generate(factor1)
scalar alpha1 = r(alpha)
alpha rep78-foreign, generate(factor2)
scalar alpha2 = r(alpha)
alpha trunk-foreign, generate(factor3)
scalar alpha3 = r(alpha)

// make table of rotated factor loadings
frmttable, statmat(e(r_L)) varlabels

// insert blank row; need -blankrows- option from now on
mat row = J(1,`e(f)',.)
frmttable, statmat(row) blankrows append

// append eigenvalues>1
mat Ev = e(Ev)
mat Ev = Ev[1, 1..`e(f)']
frmttable, statmat(Ev) blankrows append

// cumulative? variance
frmttable, statmat(e(r_Ev)) rtitle(Variance) blankrows append

// Cronbach's alpha
mat alpha = (alpha1, alpha2, alpha3)
frmttable using fac_test, statmat(alpha) rtitle(Alpha) blankrows append

The resulting table looks like this:


                                       -----------------------------------------------------
                                                                 Factor1  Factor2  Factor3 
                                       -----------------------------------------------------
                                        Price                     1.13     0.35     -0.23  
                                        Mileage (mpg)             -0.57    0.29     -0.14  
                                        Repair Record 1978        0.17     1.03     0.28   
                                        Headroom (in.)            -0.22    0.11     1.07   
                                        Trunk space (cu. ft.)     0.18     0.10     0.84   
                                        Weight (lbs.)             0.60     -0.36    0.20   
                                        Length (in.)              0.48     -0.30    0.36   
                                        Turn Circle (ft.)         0.34     -0.55    0.19   
                                        Displacement (cu. in.)    0.59     -0.38    0.16   
                                        Gear Ratio                -0.41    0.55     -0.05  
                                        Car type                  0.13     0.92     -0.08  
                                                                                           
                                        Eigenvalues               6.79     1.32     1.02   
                                        Variance                  5.17     5.14     4.66   
                                        Alpha                     0.27     0.25     0.26   
                                       -----------------------------------------------------

John



On Dec 3, 2012, at 8:07 AM, "MacQuarrie, Kerry" <[email protected]> wrote:

> Hello Statalisters,
> 
> I am wondering if there are available commands (either native or user-written) to select from the results of factor analysis procedures and produce tables in another application (namely Excel).  I have used tabout, estout, and outreg commands for doing so with descriptive and regression results, but I'm stymied here.  I've also been reading an (somewhat older) Stata Journal article* on using the file command, but it's not clear that is exactly what I'm looking for either.
> 
> My basic code for the factor analysis takes the form of:
> 
> factor var1 var2 var3 ... vark, pcf
> rotate, promax
> predict f1 f2 f3
> alpha var1 var2...vark, generate(factor1)
> alpha var3 var4...vark, generate(factor2)
> alpha var5 var6...vark, generate(factor3)
> 
> I want to save a table that has the rotated factor loadings that comes from the second line of syntax--this is the core of the table--, but I want to append an output table with other pieces of information too.  These would include the eigenvalues and cumulative variance (but only for the retained factors, e.g. those with eigenvalues>1) from the first line and the Cronbach's alpha for each of the factors in lines 4-6.
> 
> Another complication is that the above factor analysis code will likely be modified into looped code, as I am running the factor analyses on multiple country datasets, each of which (so far) appear to have the same structure regarding these factors.
> 
> Any tips on where to start would be much, much appreciated.
> 
> Regards,
> Kerry
> 
> *Slaymaker, E. (2005) "Using the file command to produce formatted output for other applications." The Stata Journal 5(2): 239-247.
> 
> Kerry MacQuarrie
> 
> 
> 
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index