interpretion of index generated by multiple correspondence analysis

[Mustafa Ozer]

Dear Statalist,

In order to create an attitude index towards domestic violence, I run the multiple correspondence analysis. Dichotomous variables in the data sets are coded as 0 and 1.

0 defined as No. 1 is defined as Yes.

Could you please help me to interpret the results of the index. Unfortunately, I could not find an online answer.
For instance, what does -3.067951 mean?

 mca Burningfood Neglectingchild argueWithhusnd refusesex wastemoney

predict F2
 tab F2

The results are the following:


rowscore |
(dim=1; |
standard |
norm.) | Freq. Percent Cum.
------------+-----------------------------------
-3.067951 | 114 3.06 3.06
-2.438291 | 11 0.30 3.35
-2.435728 | 2 0.05 3.41
-2.401917 | 5 0.13 3.54
-2.386881 | 7 0.19 3.73
-2.295974 | 20 0.54 4.27
-2.069015 | 128 3.44 7.70
-1.772257 | 3 0.08 7.78
-1.757221 | 3 0.08 7.86
-1.666314 | 8 0.21 8.08
-1.629941 | 4 0.11 8.19
-1.614905 | 9 0.24 8.43
-1.439355 | 34 0.91 9.34
-1.402981 | 97 2.60 11.94
-1.387945 | 25 0.67 12.61
-1.297038 | 177 4.75 17.36
-1.091188 | 5 0.13 17.50
-1.000281 | 1 0.03 17.53
-.985245 | 3 0.08 17.61
-.9488717 | 5 0.13 17.74
-.7733214 | 26 0.70 18.44
-.7582854 | 20 0.54 18.97
-.7219121 | 23 0.62 19.59
-.6673782 | 93 2.50 22.09
-.6310049 | 124 3.33 25.42
-.6159689 | 92 2.47 27.89
-.3192118 | 6 0.16 28.05
-.1443132 | 1 0.03 28.07
-.0922522 | 49 1.32 29.39
-.0800516 | 2 0.05 29.44
-.001345 | 127 3.41 32.85
.013691 | 131 3.52 36.37
.0500643 | 212 5.69 42.06
.6797243 | 2,159 57.94 100.00
------------+-----------------------------------
Total | 3,726 100.00

.
Thank you very much in advance,

Kind Regards,

[William Lisowski]

Have you reviewed mv.pdf, the [MV] Stata Multivariate Statistics Reference Manual included with Stata documentation, in particular the sections titled mca, mca postestimation commands, and mca postestimation plots? Specifically, you have apparently predicted rowscores (using the mca terminology) and the subsection headed predict within the mca section has a brief discussion. But all three sections could be useful to review if you have not already been able to do so.

[Mustafa Ozer]

Dear William,

I have read them but unfortunately i could not understand it.

Could you please help me with this? How should i interpret the index?

Regards

[Mustafa Ozer]

I have coppied stata commands used to create the autonomy index below. These variables were also coded as zero(=no) and 1(=yes). No means that she does not have the autonomy. and 1 means that she has the autonomy. Could you please help me how to interpret this index? and how could i create categories in the index?

Kind Regards,

Mustafa

  mca Shopping Budget officalworks

Multiple/Joint correspondence analysis         Number of obs      =      4639
                                               Total inertia      =  .1558493
    Method: Burt/adjusted inertias             Number of axes     =         1

                |   principal               cumul
      Dimension |    inertia     percent   percent
    ------------+----------------------------------
          dim 1 |    .1556137     99.85      99.85
    ------------+----------------------------------
          Total |    .1558493    100.00

Statistics for column categories in standard normalization

                 |          overall          |        dimension_1       
      Categories |    mass  quality   %inert |   coord   sqcorr  contrib
    -------------+---------------------------+---------------------------
    Shopping     |                           |                          
          others |   0.229    1.010    0.101 |   0.666    1.010    0.102
         herself |   0.104    1.010    0.222 |  -1.469    1.010    0.224
    -------------+---------------------------+---------------------------
    Budget       |                           |                          
          others |   0.288    0.978    0.049 |   0.407    0.978    0.048
         herself |   0.046    0.978    0.307 |  -2.566    0.978    0.300
    -------------+---------------------------+---------------------------
    officalworks |                           |                          
          others |   0.301    1.010    0.031 |   0.325    1.010    0.032
         herself |   0.033    1.010    0.291 |  -3.007    1.010    0.294
    ---------------------------------------------------------------------


.  predict autonmy_indx
(option rowscores assumed)
(6 missing values generated)
                                            

. tab  autonmy_indx

   rowscore |
    (dim=1; |
   standard |
     norm.) |      Freq.     Percent        Cum.
------------+-----------------------------------
  -3.039301 |        228        4.91        4.91
  -2.117997 |         27        0.58        5.49
  -1.756179 |        148        3.19        8.67
   -1.60096 |        271        5.83       14.51
  -.8348755 |         50        1.08       15.58
  -.7402133 |          2        0.04       15.63
  -.6796558 |        109        2.35       17.97
  -.3178381 |        800       17.22       35.19
    .474138 |          3        0.06       35.26
   .6034657 |      3,006       64.70       99.96
   .6417423 |          1        0.02       99.98
   .6945167 |          1        0.02      100.00
------------+-----------------------------------
      Total |      4,646      100.00

. sum  autonmy_indx

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
autonmy_indx |      4646    .0002751    .9996408  -3.039301   .6945167

[William Lisowski]

Replying to the question posed to me at #3 above, unfortunately, I'm not very familiar with MCA, so I probably understand it less than you do. My post at #2 was just to be sure that you were aware of all the sources of Stata documentation for the command. The discussion of MCA in [MV] suggests that a good text for gaining a better understanding MCA might be Rencher, A. C., and W. F. Christensen. 2012. Methods of Multivariate Analysis. 3rd ed. Hoboken, NJ: Wiley. From Wikipedia, another possible text is Greenacre, Michael (2007). Correspondence Analysis in Practice, Second Edition. London: Chapman & Hall/CRC.

[Mustafa Ozer]

thank you very much,

Kind Regards

[shruti shuvam]

Dear Mustafa,

May I know, did you find the answer to this question?
If so, kindly share this with me, I am also experiencing a similar issue.

Thank you in advance