generate t-stat

[Sartaj Hussain]

How can we modify this code to generate t-stat of the means instead of standard deviation (sd), and also for 4*4 sorts.

frame create means_and_sds int (idiovol_group mcap_group) ///
    float(mean_vw sd_vw mean_ew sd_ew)
forvalues iv = 1/5 {
    forvalues mc = 1/5 {
        summ vw_mean_q`mc'_idiovol_q`iv'_rt
        local mean_vw = r(mean)
        local sd_vw = r(sd)
        summ ew_mean_q`mc'_idiovol_q`iv'_rt
        local mean_ew = r(mean)
        local sd_ew = r(sd)
        frame post means_and_sds (`iv') (`mc') (`mean_vw') (`sd_vw') ///
            (`mean_ew') (`sd_ew')
    }
}
frame change means_and_sds
rename (mean* sd*) =_idiovol_
reshape wide *_idiovol_, i(mcap_group) j(idiovol_group)

[Clyde Schechter]

To change this for 4*4 binning, just change 1/5 to 1/4 in both -forvalues- commands. Nothing more is needed for that.

As for the t-statistics, do you want 2 separate t-statistics, one comparing mean_vw to 0 and the other comparing mean_ew to 0? Or do you want a single t-statistic comparing mean_ew to mean_vw?

[Sartaj Hussain]

I need this one:

I want 2 separate t-statistics, one comparing mean_vw to 0 and the other comparing mean_ew to 0.

[Clyde Schechter]

frame create means_and_sds int (idiovol_group mcap_group) ///
    float(mean_vw sd_vw t_vw mean_ew sd_ew t_ew)
forvalues iv = 1/4  {
    forvalues mc = 1/4 {
        summ vw_mean_q`mc'_idiovol_q`iv'_rt
        local mean_vw = r(mean)
        local sd_vw = r(sd)
        ttest vw_mean_q`mc'_idiovol_q`iv'_rt = 0
        local t_vw = r(t)
        summ ew_mean_q`mc'_idiovol_q`iv'_rt
        local mean_ew = r(mean)
        local sd_ew = r(sd)
        ttest ew_mean_q`mc'_idiovol_q`iv'_rt = 0
        local t_ew = r(t)
        frame post means_and_sds (`iv') (`mc') (`mean_vw') (`sd_vw') (`t_vw') ///
            (`mean_ew') (`sd_ew') (`t_ew')
    }
}
frame change means_and_sds
rename (mean* sd* t* ) =_idiovol_
reshape wide *_idiovol_, i(mcap_group) j(idiovol_group)

Changes in bold face.

I did not remove the standard deviations. If you really don't want them, you can always drop them at the end.

[Sartaj Hussain]

I tried the code #4 but there is some error that needs to be resolved. I append data example with a few variables and snapshot of error herewith.

* Example generated by -dataex-. For more info, type help dataex
clear
input float(mdate vw_mean_mcap_q1_idiovol_q1_rt ew_mean_mcap_q1_idiovol_q1_rt vw_mean_mcap_q1_idiovol_q2_rt ew_mean_mcap_q1_idiovol_q2_rt)
558  .009547775   .02479577   .02437658   .03475084
559   .10438219   .11858144   .08377855   .08805264
560   .08116962    .0792958   .07097481  .032742538
561   .04498115   .04408213   .04740855  .024079055
562 -.022288915  -.00587333   .03342637  .010281968
563  .019821817  .016998773  .011102415  .007421346
564  .067377895   .06323038  .016407823   .07976601
565   -.1059341   -.0785012    -.150866  -.13820857
566  -.07210954  -.05660603  -.06896405 -.067165166
567   .08386143   .08955546   .10053737   .10737565
568   .04357367   .05943481   .06094366    .0867588
569  .024111053   .02701717    .0080635   .05110602
570   .06358095    .0607609 -.031920392 -.016395649
571   -.0270472  -.02999136  -.03761415 -.011755468
572   .04246868   .04316803   .05046064   .06170107
573  .007294329    .0467165   .02188667  .022017315
574   .04165136   .04003866  .029694706   .04133752
575   .22977667    .2277255    .2143408   .20428786
576  -.22754773  -.22565822   -.2362953  -.27728713
577   .00794934  .011074592 -.070291035  -.05864974
578  -.15997925   -.1640525   -.1669928   -.1578676
579   .10637853   .11093468   .09823243   .11571355
580  -.02954163  -.03946456  -.05908868  -.06122958
581  -.16585436   -.1590568   -.1615779  -.15867306
582   .03693952   .04350401   .05151897   .05113066
583  .012166202  .004758508  .036111612   .03958639
584   -.1851966   -.1810432  -.21498574  -.20123994
585   -.3348244   -.3217562   -.3549802   -.3421502
586  -.10287447  -.08428375   -.1688749   -.1583635
587   .09367303   .07917035   .09050088   .08274078
588  -.06365843  -.06772166  -.07280084  -.05733358
589  -.08381806  -.07905384   -.0422506   -.0204582
590   .03771116   .03715475   .06624628   .04974006
591   .19414186   .19908004   .27371666   .24236186
592    .3274282    .3091991     .419429    .3985935
593  .033857517  .008383145  .002679941  .022902023
594   .11746235    .1435006   .10666234   .10002537
595   .12716314    .1357516    .1760132    .1691689
596    .1425892    .1412383   .07238307   .07065684
597  -.03785864   -.0527435  -.05946851  -.03268821
598    .0819361   .06925205   .10378387   .09575696
599   .15803707    .1544773   .09632215    .1232953
600 -.010999424 -.003957911 -.002746893   .02924245
601  .000514733 -.001124334 -.003368403  .003616981
602   .02170511  .030316364   .10812513   .11198892
603    .0783469   .08806104   .10514018   .12258244
604  -.07671385  -.07721204  -.06355686  -.06517138
605   .05854668   .06292336    .1078936   .11029316
606  .010707898  .032265536   .03949427   .04722167
607  .032694664  .018592536 -.009044237   .00068132
608   .04763261   .04564847   .11241443     .101737
609   .03009066   .02293751    .0008051  .008618193
610 -.030519104  -.04080753   -.1104302  -.11623478
611  .007439347  .007219307   .03413117  .026593776
612  -.13044187  -.12824793  -.09124906   -.0957571
613  -.05239417  -.05904739   -.0648646  -.07034237
614   .05894769    .0571517   .06899232   .07076477
615   .08696517   .09987694   .06393108    .0686882
616    -.040951  -.02531785  -.06234858  -.04622807
617 -.027015405  -.02894352 -.014230093  -.02519445
618   .02818784   .02450861   .07290703   .06457929
619  -.08527008  -.10849579  -.16345036   -.1631169
620 -.013742953 -.020762913  -.03907686 -.026884234
621  -.01020511 -.011751142   .02321391  .034331482
622  -.14280148  -.15280125  -.12470736  -.10832244
623  -.10006838   -.0965528  -.08936226  -.08807243
624   .14610061    .1563075   .17279357   .16941033
625   .08852811   .07399375   .05945755    .0521311
626  -.02549717 -.032774474  .001330905 -.003583591
627    .0459909   .06156145   .06713442   .04878448
628  -.04448419   -.0411814  -.07128345  -.05296824
629  .036024313   .03982674   .08572852   .06328414
630   .02592565   .03482852 -.002638496  -.02003376
631  .016543996  .016940258  .007570951   .04160224
632  .035490435   .05228644   .10778483   .08864895
633  .011696635  .030955903   .02460151   .03971419
634  .036192592   .04184807   .04925531   .06511546
635  .018727016  .017574824  .022316355  .030821254
636 -.063359216 -.068041295  .024449097  .010795265
637  -.09276952   -.1063733   -.0967764   -.0975683
638 -.016505292  -.04158242    -.120223  -.13017873
639  .016715733  .016578164   .08983716   .08676432
640  .018064087   .00524892 -.027863465 -.027553303
641  -.04754739  -.05412241  -.05063578  -.06187253
642  -.04007133  -.04559651 -.034037724 -.021941947
643 -.012221493 -.004349294  -.06043958  -.05565551
644   .05380223   .04430003   .05750297   .05579302
645    .0800121   .08847322   .10312444   .12573172
646   .07392643   .06666263    .0871317   .08839425
647   .08568712   .08508613   .04637758   .05119117
648 -.032748193 -.035004146  -.05185303  -.04392271
649  .008319468  .017399747   .02489616   .01859344
650    .1350643   .11971303   .12478655    .1241536
651   .10595305   .11478766   .09209468   .12088978
652   .13920607    .1494198   .20063406    .1954328
653    .1514034   .15727663    .1581262    .1759043
654   .02873998    .0507537 -.000114384   .01046246
655  .028436223  .028203806   .08201407   .08138157
656   .06895868   .07335494   .09946185   .11574554
657  .008545064  .007172796   .02306144   .01921513
end
format %tm mdate

[Clyde Schechter]

I should have asked for example data before writing the code in #4. I dug some up from one of our earlier threads, but it turns out that the variable names were different from what you show here. The following fixes the problem:

frame create means_and_sds int (idiovol_group mcap_group) ///
    float(mean_vw sd_vw t_vw mean_ew sd_ew t_ew)
forvalues iv = 1/4  {
    forvalues mc = 1/4 {
        summ vw_mean_mcap_q`mc'_idiovol_q`iv'_rt
        local mean_vw = r(mean)
        local sd_vw = r(sd)
        ttest vw_mean_mcap_q`mc'_idiovol_q`iv'_rt = 0
        local t_vw = r(t)
        summ ew_mean_mcap_q`mc'_idiovol_q`iv'_rt
        local mean_ew = r(mean)
        local sd_ew = r(sd)
        ttest ew_mean_mcap_q`mc'_idiovol_q`iv'_rt = 0
        local t_ew = r(t)
        frame post means_and_sds (`iv') (`mc') (`mean_vw') (`sd_vw') (`t_vw') ///
            (`mean_ew') (`sd_ew') (`t_ew')
    }
}
frame change means_and_sds
rename (mean* sd* t* ) =_idiovol_
reshape wide *_idiovol_, i(mcap_group) j(idiovol_group)

[Sartaj Hussain]

#6 works very well. Thanks indeed!