How to make categories using standardized variable and its standard deviation

[zulfiqar ali shah]

I am interested in making categories(Large, small, medium) based on the log of [/CODE]market value of equity (z_ln_mve)
the variable (z-ln_mve) is standardized using commands:

gen z_ln_mve=0

replace z_ln_mve=ln_mve-mean(ln_mve)

I want the categories to be
small= less than or equal to _1 standard deviation in the market value variable (z_ln_mve)
medium= between greater than -1 and less than +1 standard deviation
large= greater than or equal to +1 standard deviation
the sample data is given below
where ln_mve= log of market value
and
z_ln_mve= standardized values of log of market values.

sum z_ln_mve

Variable | Obs Mean Std. Dev. Min Max
----------------------------------------------------------------------
z_ln_mve | 24,605 1.09e-06 2.304831 -6.974789 6.907081

* Example generated by -dataex-. To install: ssc install dataex
clear
input double ln_mve float z_ln_mve
17.565108747112916  -3.331611
 17.40258981761514   -3.49413
  17.8344416808965  -3.062278
17.688934097225264  -3.207786
16.811052813465977  -4.085667
17.639541341895686 -3.2571785
 17.57640830236685  -3.320312
 17.46979856730859 -3.4269214
17.856760203852197   -3.03996
17.865213339563255  -3.031507
18.012122858133953  -2.884597
 18.10107034415045   -2.79565
18.137627939884247  -2.759092
17.875263675416754  -3.021456
18.181522133441472  -2.715198
17.993046711073138  -2.903673
 17.86689542574624 -3.0298245
17.788252298427125  -3.108468
17.692942118622803  -3.203778
17.875263675416754  -3.021456
17.753096041442546  -3.143624
17.830963416520177 -3.0657566
17.788252298427125  -3.108468
17.764332114709475  -3.132388
 17.87025113359321  -3.026469
17.955306383090292 -2.9414136
18.029414355244015  -2.867306
 17.88356247823145 -3.0131576
 17.93980219655433  -2.956918
 17.93980219655433  -2.956918
17.891792977367967  -3.004927
17.875263675416754  -3.021456
 18.34004536068134  -2.556675
 18.86851544842704 -2.0282044
18.907854862987072  -1.988865
18.865424255857366 -2.0312958
18.922582669697317 -1.9741373
 18.92841359000811 -1.9683064
18.856717733197325 -2.0400023
 18.84667370651199 -2.0500462
 18.87773210353196  -2.018988
18.959889722110127 -1.9368303
18.869749254875966 -2.0269709
18.589030143179436   -2.30769
18.641506411047253 -2.2552135
 18.78419405288462  -2.112526
19.458357612511254 -1.4383624
20.288196825579668  -.6085232
19.792186287598817 -1.1045337
  19.4210515607744 -1.4756684
 19.36027150947268 -1.5364485
19.174546659547016 -1.7221733
19.179081814712408  -1.717638
 19.00127493827298  -1.895445
18.920240748252233 -1.9764793
19.044127302629125 -1.8525927
 19.05032246957805 -1.8463975
 18.74376374345456 -2.1529562
 18.75765285561523 -2.1390672
18.974431365376567 -1.9222887
18.615612457460994 -2.2811074
18.291778619711504 -2.6049414
18.386089299182746 -2.5106306
 18.46859052069449 -2.4281294
 18.43487946335218 -2.4618406
18.605224829099416  -2.291495
18.636069504450514 -2.2606504
18.648453563650236 -2.2482665
18.481399478987072 -2.4153206
   18.335848082746  -2.560872
 18.28072878352492  -2.615991
 18.28072878352492  -2.615991
 18.28072878352492  -2.615991
18.345267304662492 -2.5514526
 18.35563009169804   -2.54109
18.386089299182746 -2.5106306
18.258255927672863  -2.638464
 18.42049072590008  -2.476229
18.481399478987072 -2.4153206
 18.43487946335218 -2.4618406
18.258255927672863  -2.638464
18.386089299182746 -2.5106306
18.308127757713034  -2.588592
 18.39702923922108  -2.499691
 18.49941798448975  -2.397302
 18.43487946335218 -2.4618406
 18.49941798448975  -2.397302
 18.55160373766032 -2.3451164
 18.55160373766032 -2.3451164
18.481399478987072 -2.4153206
18.391076840693785  -2.505643
18.391076840693785  -2.505643
18.385088798849164  -2.511631
 18.33374281881054  -2.562977
 18.33374281881054  -2.562977
22.984427670946943  2.0877078
22.984427670946943  2.0877078
 22.99255779703019  2.0958378
22.833960674210953  1.9372407
 22.92828914500774  2.0315692
end

please provide me with the necessary commands for the operations.
Thank you

[Carlo Lazzaro]

Zulfiqar:
you may want to consider:

. sum z_ln_mve

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
    z_ln_mve |        100   -2.281381    1.132245  -4.085667   2.095838

. g wanted=0 if z_ln_mve<=r(sd)

. replace wanted=1 if z_ln_mve>-r(sd) & z_ln_mve<r(sd)

. replace wanted=2 if z_ln_mve>=r(sd)

. label define wanted 0 Small 1 Medium 2 Large

. label val wanted wanted

.

[Nick Cox]

I guess #1 did not mean what it said, although Carlo Lazzaro helpfully gave what was asked for.
z_ln_mve is already standardized. I imagine zulfiqar wants values less than mean - SD, more than mean + SD, and so on.

gen wanted = cond(z < -1, 1, cond(z >= 1, 3, 2)) if z < .

will map standardized values less than -1 to 1, those more than or equal to 1 to 3 and others to 2. Then define and apply value labels.

But why do this -- as it's throwing away useful detail? I don't understanding the enthusiasm for coarse binning of measured variables.

[zulfiqar ali shah]

thank you for the response.
It worked but in further steps I want to make interactions of the (z_ln_mve) and oil price returns (z_rBrent) in order to check the size moderated effect of (z_ln_mve) in explaining relationship of (z_rBrent) and dependent variable "X" which is problematic using the dummy variables. Please explain
The relationship is like:

X= B1(_ln_mve)+ B2(z_rBrent) +B3[(z_ln_mve)*(z_rBrent)]

data sample is:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(z_ln_mve z_rBrent)
 -3.331611     .0829995
  -3.49413    .06447985
 -3.062278    .05643204
 -3.207786   -.15744352
 -4.085667   .005509714
-3.2571785    .04209122
 -3.320312    -.0456385
-3.4269214    .09851485
  -3.03996   .010852398
 -3.031507   .033469226
 -2.884597    .09852401
  -2.79565    .06467648
 -2.759092    .10219084
 -3.021456    .04923337
 -2.715198   .070861176
 -2.903673   -.07484613
-3.0298245  -.036389347
 -3.108468    .03787261
 -3.203778   -.01562351
 -3.021456   -.11053196
 -3.143624   .064774975
-3.0657566   .009422956
 -3.108468  -.028123755
 -3.132388    .03367486
 -3.026469     .1007651
-2.9414136   .002490769
 -2.867306  -.027444176
-3.0131576     -.158669
 -2.956918   -.04007411
 -2.956918   .070972376
 -3.004927    .08868664
 -3.021456  -.018512225
 -2.556675   -.03268437
-2.0282044   .023707135
 -1.988865  -.000380628
-2.0312958    .03988143
-1.9741373  -.036056757
-1.9683064   -.01158681
-2.0400023   -.07136966
-2.0500462    -.0188323
 -2.018988    .01817641
-1.9368303    .05350817
-2.0269709    .05763538
  -2.30769   -.05003607
-2.2552135   .005026416
 -2.112526   .008478091
-1.4383624   .010767369
 -.6085232    -.0398224
-1.1045337     .0254831
-1.4756684  -.011384546
-1.5364485   .003571433
-1.7221733   .013021928
 -1.717638    .02730451
 -1.895445   -.05738145
-1.9764793  -.026357006
-1.8525927   -.08547599
-1.8463975    -.0969803
-2.1529562   -.20138346
-2.1390672   -.20111296
-1.9222887   -.07802203
-2.2811074     .1670413
-2.6049414   -.12641574
-2.5106306    .19277124
-2.4281294   -.01773908
-2.4618406   -.02981073
 -2.291495   -.19648337
-2.2606504   .037182726
-2.2482665   -.11217938
-2.4153206    .02500307
 -2.560872    -.1045272
 -2.615991   -.17880227
 -2.615991   -.06986643
 -2.615991   .035492297
-2.5514526    .09684266
  -2.54109    .19577536
-2.5106306   .032596823
 -2.638464   .000497532
 -2.476229    -.1563412
-2.4153206    .10313465
-2.4618406    .04274454
 -2.638464  -.014913678
-2.5106306    .04464634
 -2.588592    .11920808
 -2.499691   -.01920943
 -2.397302  -.001278018
-2.4618406   -.05022532
 -2.397302  -.020342527
-2.3451164   -.02713522
-2.3451164   -.04797211
-2.4153206    .09483209
 -2.505643 -.0044425996
 -2.505643    .09465432
 -2.511631    .06513956
 -2.562977   .035919346
 -2.562977    .05130757
 2.0877078     .0829995
 2.0877078    .06447985
 2.0958378    .05643204
 1.9372407   -.15744352
 2.0315692   .005509714
end

[zulfiqar ali shah]

Nick Cox Sir actually I am interested in Size moderated effect of oil prices on my dependent variable. I have already mentioned it in my reply to Carlo Lazzaro sir. Please just check it. You are absolutely right. It results in loss of many observations and the results are very confusing. What will be suitable so that would not effect the number of observations?

[zulfiqar ali shah]

#3
Nick Cox sir.
I want to make three different categories on the basis of (z_ln_mve) as small medium and large and the impact of (z_rBrent) on (X-variable) of different sizes.

[Nick Cox]

#4 #5 #6 make the context a little clearer. You can be interested in the interaction of predictors without needing to degrade predictors to categorical.

[zulfiqar ali shah]

Nick Cox I am interested in both. First I am looking the impact of overall interaction of predictors on the dependent variable.
like

(z_ln_mve)*(z_rBrent

than the impact of interaction of different categories based on

z_ln_mve

(size1, size2, size3) on the concerned variable like:

(size1)*(z_rBrent

(size2)*(z_rBrent

(size3)*(z_rBrent