Difference-in-differences

Clyde Schechter

Join Date: Apr 2014

Posts: 29954
#31

25 Apr 2021, 16:51

This looks slightly better, but it's still not right. We are only seeing a few values of event_time_var. That shouldn't be happening. We should see them all (except for -5 as a reference category).

Also, before we introduce the complications of a higher level interaction with age, let's get the basic model working. Then we can try adding in the extra features.

To that end, I need you to use -dataex- to post example data that meets the following constraints:

1. It contains observations from both 3 states where ps = 1 and 3 states where ps = 0, and contains the ps variable.
2. It contains a variable showing the year in which each state began pre-registration
3. It contains the year variable and the statefips variable.
4. The years included for any given state are a block of consecutive years, and for ps = 1 states include the year in which pre-registration began as well as some before and some after.
5. It contains the outcome variable (register).

Then I can try to figure out what's going wrong.
Comment

Cairone Federica

Join Date: Apr 2021
Posts: 31

#32

26 Apr 2021, 01:18

When I try to select what you tell me using together with the command -dataex-, the command -if- STATA shows me this message

input statement exceeds linesize limit. Try specifying fewer variables

Code:


	Code:
	* Example generated by -dataex-. To install: ssc install dataex
clear
input float ps byte statefip float event_time_var int year float(years_since_implement year_first_post_implement)
0 1 0 1988 -5 .
0 1 0 2008 -5 .
0 1 0 1988 -5 .
0 1 0 1984 -5 .
0 1 0 1988 -5 .
0 1 0 1990 -5 .
0 1 0 1988 -5 .
0 1 0 2006 -5 .
0 1 0 2008 -5 .
0 1 0 1980 -5 .
0 1 0 1998 -5 .
0 1 0 1988 -5 .
0 1 0 2010 -5 .
0 1 0 1996 -5 .
0 1 0 1992 -5 .
0 1 0 1982 -5 .
0 1 0 1980 -5 .
0 1 0 1982 -5 .
0 1 0 2000 -5 .
0 1 0 1980 -5 .
end
label values statefip STATEFIP
label def STATEFIP 1 "alabama", modify
label values event_time_var event_time_var
label def event_time_var 0 "-5", modify

I know it's not exhaustive, but I don't know how to show it to you. Unfortunately, the only thing I could do is put a screenshot of the Data Editor (Browse).

Click image for larger version

Name: Screen1.png
Views: 2
Size: 122.4 KB
ID: 1605857

Click image for larger version

Name: Scree2.png
Views: 1
Size: 130.1 KB
ID: 1605858

Attached Files

Comment

Cairone Federica

Join Date: Apr 2021
Posts: 31

#33

26 Apr 2021, 01:27

I give you some cross tabulation, maybe it can be useful .. in fact, they are not very convincing...

Code:

 tab ps years_since_implement

           | years_since_implement
        ps |        -5          3 |     Total
-----------+----------------------+----------
         0 | 1,213,442          0 | 1,213,442 
         1 |         0    495,035 |   495,035 
-----------+----------------------+----------
     Total | 1,213,442    495,035 | 1,708,477

Code:

tab ps  event_time_var

           |    event_time_var
        ps |        -5          3 |     Total
-----------+----------------------+----------
         0 | 1,213,442          0 | 1,213,442 
         1 |         0    495,035 |   495,035 
-----------+----------------------+----------
     Total | 1,213,442    495,035 | 1,708,477

Code:

tab ps year_first_post_implement

           |               year_first_post_implement
        ps |      1994       2008       2010       2012       2014 |     Total
-----------+-------------------------------------------------------+----------
         1 |    20,042     93,926    263,663     24,556     92,848 |   495,035 
-----------+-------------------------------------------------------+----------
     Total |    20,042     93,926    263,663     24,556     92,848 |   495,035

Comment

Clyde Schechter

Join Date: Apr 2014

Posts: 29954
#34

26 Apr 2021, 10:57

OK, well thanks for trying. Something is clearly wrong with the years_since_implement because we are getting only 3 and -5, and we should be getting all the numbers (or nearly all of them) in between as well. Now, years_since_implement is calculated from year_first_post_implement (which looks OK), and year, I worry that there is a problem with the year variable in your data. But what you have shown does not give any information about that variable.

Can you show the results of:

Code:

table year year_first_post_implement, c(min years_since_implement max years_since_implement)

Then run your regression model again and show the results of:

Code:

table year year_first_post_implement if e(sample), c(min years_since_implement max years_since_implement)
Comment

Cairone Federica

Join Date: Apr 2021
Posts: 31

#35

26 Apr 2021, 11:28

Yes! For sure. Here we are:

Code:

table year year_first_post_implement, c(min years_since_implement max years_si
> nce_implement)

----------------------------------------
survey    |  year_first_post_implement  
year      | 1994  2008  2010  2012  2014
----------+-----------------------------
     1982 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1984 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1986 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1988 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1990 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1992 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1994 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1996 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1998 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2000 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2002 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2004 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2006 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2008 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2010 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2012 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2014 |    3     3     3     3     3
          |    3     3     3     3     3
----------------------------------------

and then after the basic regression:

Code:

. table year year_first_post_implement if e(sample), c(min years_since_implement
>  max years_since_implement)

----------------------------------------
survey    |  year_first_post_implement  
year      | 1994  2008  2010  2012  2014
----------+-----------------------------
     1982 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1984 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1986 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1988 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1990 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1992 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1994 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1996 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     1998 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2000 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2002 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2004 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2006 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2008 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2010 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2012 |    3     3     3     3     3
          |    3     3     3     3     3
          | 
     2014 |    3     3     3     3     3
          |    3     3     3     3     3
----------------------------------------

.

I am sending you the commands that were originally used, but which for me are very complex and I cannot understand. Maybe they can help you:

Code:

*Generate Cohorts*
gen ps = 0
by statefip: egen max_pre=max(pre_reg)
by statefip: replace ps =max_pre

by statefip: egen pre_reg_y=min(year) if pre_reg==1
by statefip: egen target=min(pre_reg_y)
egen treated_year=group(pre_reg_y)
by statefip: egen cohort=max(treated_year)
replace cohort=0 if cohort==.

*Generate Leads and Lags*
forvalues kk = 0(1)5 {
by statefip: gen F`kk'=target-2*`kk'
by statefip: gen F`kk'_pre=0
by statefip: replace F`kk'_pre=1 if age18_24==1 & year==F`kk'
by statefip: gen Fold`kk'_pre=1 if age18_24==0 & year==F`kk'
}

forvalues kk = 1(1)3 {
by statefip: gen L`kk'=target+2*`kk'
by statefip: gen L`kk'_pre=0
by statefip: replace L`kk'_pre=1 if age18_24==1 & year==L`kk'
by statefip: gen Lold`kk'_pre=1 if age18_24==0 & year==L`kk'
}

by statefip: gen F5_last=0
by statefip: replace F5_last=1 if age18_24==1 & year<=target-10 & target!=.
by statefip: gen L3_last=0
by statefip: replace L3_last=1 if age18_24==1 & year>=target+6 & target!=.

*Generate event window*
gen eventwindow = 0
forvalues kk = 0(1)5 {
    replace eventwindow = 1 if F`kk'_pre == 1 | Fold`kk'_pre==1
    }
forvalues kk = 1(1)2 {
    replace eventwindow = 1 if L`kk'_pre == 1 | Lold`kk'_pre==1
    }

* Generate mean of omitted time
gen year_omitted=.
replace year_omitted=year if F1_pre==1
by state: egen max_year_omitted=max(year_omitted)
by state: egen register_young_m=mean(register) if (age18_24==1 & year==max_year_omitted)
by state: egen register_old_m=mean(register) if (age18_24==0 & year==max_year_omitted)
by state: egen max_register_young_m=max(register_young_m)
by state: egen max_register_old_m=max(register_old_m)
gen register_gap_m=max_register_old_m-max_register_young_m

by state: egen vote_young_m=mean(vote) if (age18_24==1 & year==max_year_omitted)
by state: egen vote_old_m=mean(vote) if (age18_24==0 & year==max_year_omitted)
by state: egen max_vote_young_m=max(vote_young_m)
by state: egen max_vote_old_m=max(vote_old_m)
gen vote_gap_m=max_vote_old_m-max_vote_young_m

by state: egen register_young_D=mean(register) if (age18_24==1 & pre_reg==0 & target!=.)
by state: egen register_old_D=mean(register) if (age18_24==0 & pre_reg==0 & target!=.)
by state: egen max_register_young_D=max(register_young_D)
by state: egen max_register_old_D=max(register_old_D)
gen register_gap_D=max_register_old_D-max_register_young_D

by state: egen vote_young_D=mean(vote) if (age18_24==1 & pre_reg==0 & target!=.)
by state: egen vote_old_D=mean(vote) if (age18_24==0 & pre_reg==0 & target!=.)
by state: egen max_vote_young_D=max(vote_young_D)
by state: egen max_vote_old_D=max(vote_old_D)
gen vote_gap_D=max_vote_old_D-max_vote_young_D

Comment

Clyde Schechter

Join Date: Apr 2014

Posts: 29954
#36

26 Apr 2021, 11:39

OK, so clearly the calculation of years_since_implement is going wrong. The code you show, however, does not include the code that creates that variable. In fact, in addition to being complicated, the code mostly creates variables that we are not currently using. Please show the code that creates the variable years_since_implement. The problem is almost certainly in that.
Comment

Cairone Federica

Join Date: Apr 2021
Posts: 31

#37

26 Apr 2021, 12:18

Yes. This is the code:

Code:

/*GEN EVENTS ON AGE 18-24 PREREG*/

so statefip

*Generate Cohorts*
gen treated_states = 0
by statefip: egen max_pre=max(pre_reg)
by statefip: replace treated_state=max_pre

by statefip: egen pre_reg_y=min(year) if pre_reg==1
by statefip: egen target=min(pre_reg_y)
egen treated_year=group(pre_reg_y)
by statefip: egen cohort=max(treated_year)
replace cohort=0 if cohort==.

Comment

Clyde Schechter

Join Date: Apr 2014

Posts: 29954
#38

26 Apr 2021, 13:05

But that code does not create the variable years_since_implement. The code creating years_since_implement is doing something wrong--so we need to focus on that.

I did a little experimenting with some made up data that looks like what I think you have. And there was a mistake in what I proposed for it. Here's corrected code:

Code:

gen years_since_implement = year - year_first_post_implement replace years_since_implement = max(min(years_since_implement, 3), -5) replace years_since_implement = -5 if ps == 0

I think this will fix the problems

Last edited by Clyde Schechter; 26 Apr 2021, 13:15.
Comment

Cairone Federica

Join Date: Apr 2021
Posts: 31

#39

26 Apr 2021, 13:28

Code:

. regress register i.ps##i.years_since_implement i.statefip i.year, cluster(stat
> efip) 
years_since_implement:  factor variables may not contain negative values
r(452);

Comment

Clyde Schechter

Join Date: Apr 2014
Posts: 29954

#40

26 Apr 2021, 13:31

Yes, I realized back in #29 that this would happen and proposed the following solution there:

Code:

gen event_time_var = years_since_implement + 5
label define event_time_var  0 "-5" 1 "-4" 2 "-3" 3 "-2" 4 "-1" 5 "0" 6 "1" 7 "2" 8 "3"
label values event_time_var event_time_var

regress register i.ps##i.event_time_var ... // ADDITIONAL COVARIATES AND OPTIONS AS APPROPRIATE

Comment

Cairone Federica

Join Date: Apr 2021
Posts: 31

#41

26 Apr 2021, 14:36

Okay, I'm driving you crazy! I'm sorry. Keep giving me the same mistake, maybe I'm doing something wrong.

Code:

/*EVENT-TIME DUMMY */
by statefip, sort: egen year_first_post_implement = min(cond(pre_reg, year, .))
gen years_since_implement = year - year_first_post_implement
gen event_time_var = years_since_implement + 5
label define event_time_var  0 "-5" 1 "-4" 2 "-3" 3 "-2" 4 "-1" 5 "0" 6 "1" 7 "2" 8 "3"
label values event_time_var event_time_var

Then, when I regress:

Code:

 regress register i.ps##i.event_time_var i.statefip i.year, cluster(statefip) 
event_time_var:  factor variables may not contain negative values

Comment

Clyde Schechter

Join Date: Apr 2014

Posts: 29954
#42

26 Apr 2021, 15:49

I think your rushing too fast, and copying code without looking at it to see where it goes in the sequence. You skipped two important commands. The full sequence for preparing event_time var, start to finish, looks like this:

Code:

gen years_since_implement = year - year_first_post_implement replace years_since_implement = max(min(years_since_implement, 3), -5) replace years_since_implement = -5 if ps == 0 gen event_time_var = years_since_implement + 5 label define event_time_var 0 "-5" 1 "-4" 2 "-3" 3 "-2" 4 "-1" 5 "0" 6 "1" 7 "2" 8 "3" label values event_time_var event_time_var

Now you should not run into a factor variable problem.
Comment

Cairone Federica

Join Date: Apr 2021
Posts: 31

#43

26 Apr 2021, 23:57

Sorry, you're right! I fixed it, but when I regress:

Code:

 regress register i.ps##i.event_time_var i.statefip i.year, cluster(statefip) 
note: 0b.ps#1.event_time_var identifies no observations in the sample
note: 0b.ps#3.event_time_var identifies no observations in the sample
note: 0b.ps#5.event_time_var identifies no observations in the sample
note: 0b.ps#7.event_time_var identifies no observations in the sample
note: 0b.ps#8.event_time_var identifies no observations in the sample
note: 1.ps#1.event_time_var omitted because of collinearity
note: 1.ps#3.event_time_var omitted because of collinearity
note: 1.ps#5.event_time_var omitted because of collinearity
note: 1.ps#7.event_time_var omitted because of collinearity
note: 1.ps#8.event_time_var omitted because of collinearity
note: 44.statefip omitted because of collinearity

Linear regression                               Number of obs     =  1,350,537
                                                F(20, 50)         =          .
                                                Prob > F          =          .
                                                R-squared         =     0.0189
                                                Root MSE          =     .41782

                               (Std. Err. adjusted for 51 clusters in statefip)
-------------------------------------------------------------------------------
              |               Robust
     register |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
         1.ps |  -.0101383   .0063279    -1.60   0.115    -.0228484    .0025717
              |
event_time_~r |
          -4  |   .0088175   .0156243     0.56   0.575    -.0225648    .0401998
          -2  |   .0158877   .0118842     1.34   0.187    -.0079824    .0397578
           0  |   .0052568    .017569     0.30   0.766    -.0300317    .0405453
           2  |   .0160359    .021538     0.74   0.460    -.0272244    .0592963
           3  |   .0075756   .0249932     0.30   0.763    -.0426248     .057776
              |
           ps#|
event_time_~r |
        0#-4  |          0  (empty)
        0#-2  |          0  (empty)
         0#0  |          0  (empty)
         0#2  |          0  (empty)
         0#3  |          0  (empty)
        1#-4  |          0  (omitted)
        1#-2  |          0  (omitted)
         1#0  |          0  (omitted)
         1#2  |          0  (omitted)
         1#3  |          0  (omitted)
              |
     statefip |
      alaska  |   .0308861   .0001915   161.29   0.000     .0305014    .0312707
     arizona  |  -.0927812   .0001679  -552.58   0.000    -.0931184   -.0924439
    arkansas  |  -.0868887    .000134  -648.28   0.000    -.0871579   -.0866194
  california  |  -.0277687    .001235   -22.48   0.000    -.0302493    -.025288
    colorado  |  -.0091602   .0035583    -2.57   0.013    -.0163072   -.0020133
 connecticut  |   .0009516   .0008769     1.09   0.283    -.0008096    .0027129
    delaware  |   -.028709   .0001999  -143.59   0.000    -.0291106   -.0283074
district o..  |   .0346767   .0003227   107.47   0.000     .0340287    .0353248
     florida  |  -.0324138   .0013894   -23.33   0.000    -.0352044   -.0296232
     georgia  |  -.0680808   .0004669  -145.81   0.000    -.0690186    -.067143
      hawaii  |  -.1191363    .010763   -11.07   0.000    -.1407543   -.0975182
       idaho  |  -.0602836   .0001294  -465.79   0.000    -.0605435   -.0600236
    illinois  |   .0047676   .0004448    10.72   0.000     .0038742    .0056609
     indiana  |  -.0617863   .0002381  -259.50   0.000    -.0622646   -.0613081
        iowa  |   .0002549   .0004214     0.60   0.548    -.0005916    .0011014
      kansas  |   -.053148   .0002981  -178.29   0.000    -.0537468   -.0525492
    kentucky  |  -.0440332   .0004061  -108.43   0.000    -.0448488   -.0432175
   louisiana  |    .023204    .003506     6.62   0.000     .0161621     .030246
       maine  |    .077071   .0020555    37.50   0.000     .0729425    .0811995
    maryland  |  -.0092022   .0003877   -23.74   0.000    -.0099809   -.0084236
massachuse~s  |    .030039   .0050252     5.98   0.000     .0199456    .0401325
    michigan  |   .0438619   .0007103    61.75   0.000     .0424353    .0452886
   minnesota  |    .078792   .0007331   107.48   0.000     .0773196    .0802644
 mississippi  |   .0312137   .0003136    99.52   0.000     .0305837    .0318437
    missouri  |   .0059264   .0003565    16.62   0.000     .0052104    .0066424
     montana  |  -.0111077   .0001853   -59.93   0.000      -.01148   -.0107354
    nebraska  |  -.0270129   .0002678  -100.86   0.000    -.0275508    -.026475
      nevada  |  -.1381774   .0005891  -234.57   0.000    -.1393606   -.1369943
new hampsh..  |  -.0388454   .0012625   -30.77   0.000    -.0413812   -.0363096
  new jersey  |  -.0087847    .000886    -9.91   0.000    -.0105644    -.007005
  new mexico  |  -.0722432   .0002705  -267.11   0.000    -.0727864      -.0717
    new york  |  -.0350971   .0005965   -58.84   0.000    -.0362952   -.0338991
north caro..  |  -.0457065   .0030626   -14.92   0.000     -.051858    -.039555
north dakota  |   .1224431   .0001484   825.07   0.000      .122145    .1227412
        ohio  |  -.0354814   .0005419   -65.47   0.000    -.0365698   -.0343929
    oklahoma  |  -.0589376   .0001088  -541.54   0.000    -.0591562    -.058719
      oregon  |   .0280558   .0011865    23.65   0.000     .0256727     .030439
pennsylvania  |  -.0779009   .0004892  -159.23   0.000    -.0788836   -.0769183
rhode island  |          0  (omitted)
south caro..  |   -.080478   .0003175  -253.45   0.000    -.0811158   -.0798403
south dakota  |   .0095469    .000185    51.60   0.000     .0091753    .0099185
   tennessee  |  -.0659307   .0002355  -279.96   0.000    -.0664038   -.0654577
       texas  |  -.0566178    .000174  -325.43   0.000    -.0569673   -.0562684
        utah  |  -.0566532    .000101  -561.05   0.000     -.056856   -.0564504
     vermont  |   .0083417    .000756    11.03   0.000     .0068232    .0098602
    virginia  |  -.0465189   .0003345  -139.09   0.000    -.0471907   -.0458472
  washington  |  -.0174756   .0005158   -33.88   0.000    -.0185116   -.0164396
west virgi..  |  -.0852678   .0001112  -767.10   0.000    -.0854911   -.0850446
   wisconsin  |   .0492858    .000379   130.03   0.000     .0485245    .0500471
     wyoming  |  -.0742549   .0005674  -130.88   0.000    -.0753945   -.0731153
              |
         year |
        1984  |   .0468768   .0046794    10.02   0.000     .0374778    .0562757
        1986  |   .0061268   .0045516     1.35   0.184    -.0030153     .015269
        1988  |   .0314728   .0059299     5.31   0.000     .0195622    .0433834
        1990  |    .002353   .0059247     0.40   0.693    -.0095471    .0142532
        1992  |   .0634687   .0061362    10.34   0.000     .0511439    .0757936
        1994  |   .0048732    .006623     0.74   0.465    -.0084294    .0181758
        1996  |   .0523482   .0073225     7.15   0.000     .0376406    .0670558
        1998  |   .0245392   .0077904     3.15   0.003     .0088917    .0401867
        2000  |   .0675903   .0077751     8.69   0.000     .0519736    .0832071
        2002  |   .0354962   .0086831     4.09   0.000     .0180556    .0529368
        2004  |   .0979956   .0076852    12.75   0.000     .0825594    .1134318
        2006  |   .0593739   .0088469     6.71   0.000     .0416044    .0771435
        2008  |   .1070713   .0090163    11.88   0.000     .0889616     .125181
        2010  |   .0636579    .008792     7.24   0.000     .0459987    .0813171
        2012  |   .1027669   .0095466    10.76   0.000     .0835919    .1219418
        2014  |   .0602205   .0098251     6.13   0.000     .0404863    .0799547
              |
        _cons |   .7464395   .0056862   131.27   0.000     .7350184    .7578606
-------------------------------------------------------------------------------

or in the other case:

Code:

regress register i.ps#i.event_time_var i.statefip i.year, cluster(statefip) 
note: 0b.ps#1.event_time_var identifies no observations in the sample
note: 0b.ps#3.event_time_var identifies no observations in the sample
note: 0b.ps#5.event_time_var identifies no observations in the sample
note: 0b.ps#7.event_time_var identifies no observations in the sample
note: 0b.ps#8.event_time_var identifies no observations in the sample
note: 1.ps#8.event_time_var omitted because of collinearity

Linear regression                               Number of obs     =  1,350,537
                                                F(20, 50)         =          .
                                                Prob > F          =          .
                                                R-squared         =     0.0189
                                                Root MSE          =     .41782

                               (Std. Err. adjusted for 51 clusters in statefip)
-------------------------------------------------------------------------------
              |               Robust
     register |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
--------------+----------------------------------------------------------------
           ps#|
event_time_~r |
        0#-4  |          0  (empty)
        0#-2  |          0  (empty)
         0#0  |          0  (empty)
         0#2  |          0  (empty)
         0#3  |          0  (empty)
        1#-5  |  -.0075756   .0249932    -0.30   0.763     -.057776    .0426248
        1#-4  |   .0012419   .0171135     0.07   0.942    -.0331315    .0356154
        1#-2  |   .0083122   .0178038     0.47   0.643    -.0274479    .0440722
         1#0  |  -.0023188   .0144806    -0.16   0.873    -.0314038    .0267663
         1#2  |   .0084604   .0119609     0.71   0.483    -.0155638    .0324845
         1#3  |          0  (omitted)
              |
     statefip |
      alaska  |   .0308861   .0001915   161.29   0.000     .0305014    .0312707
     arizona  |  -.0927812   .0001679  -552.58   0.000    -.0931184   -.0924439
    arkansas  |  -.0868887    .000134  -648.28   0.000    -.0871579   -.0866194
  california  |  -.0303314   .0202296    -1.50   0.140    -.0709638     .010301
    colorado  |   -.011723   .0227306    -0.52   0.608    -.0573787    .0339328
 connecticut  |   .0009516   .0008769     1.09   0.283    -.0008096    .0027129
    delaware  |  -.0312718   .0193946    -1.61   0.113     -.070227    .0076835
district o..  |    .032114   .0191758     1.67   0.100    -.0064018    .0706298
     florida  |  -.0349765   .0196953    -1.78   0.082    -.0745357    .0045826
     georgia  |  -.0680808   .0004669  -145.81   0.000    -.0690186    -.067143
      hawaii  |   -.121699   .0086999   -13.99   0.000    -.1391732   -.1042249
       idaho  |  -.0602836   .0001294  -465.79   0.000    -.0605435   -.0600236
    illinois  |   .0047676   .0004448    10.72   0.000     .0038742    .0056609
     indiana  |  -.0617863   .0002381  -259.50   0.000    -.0622646   -.0613081
        iowa  |   .0002549   .0004214     0.60   0.548    -.0005916    .0011014
      kansas  |   -.053148   .0002981  -178.29   0.000    -.0537468   -.0525492
    kentucky  |  -.0440332   .0004061  -108.43   0.000    -.0448488   -.0432175
   louisiana  |   .0206413   .0226309     0.91   0.366    -.0248143    .0660969
       maine  |   .0745082   .0214191     3.48   0.001     .0314866    .1175299
    maryland  |   -.011765   .0191605    -0.61   0.542      -.05025      .02672
massachuse~s  |   .0274763   .0236844     1.16   0.252    -.0200952    .0750477
    michigan  |   .0438619   .0007103    61.75   0.000     .0424353    .0452886
   minnesota  |    .078792   .0007331   107.48   0.000     .0773196    .0802644
 mississippi  |   .0312137   .0003136    99.52   0.000     .0305837    .0318437
    missouri  |   .0059264   .0003565    16.62   0.000     .0052104    .0066424
     montana  |  -.0111077   .0001853   -59.93   0.000      -.01148   -.0107354
    nebraska  |  -.0270129   .0002678  -100.86   0.000    -.0275508    -.026475
      nevada  |  -.1381774   .0005891  -234.57   0.000    -.1393606   -.1369943
new hampsh..  |  -.0388454   .0012625   -30.77   0.000    -.0413812   -.0363096
  new jersey  |  -.0087847    .000886    -9.91   0.000    -.0105644    -.007005
  new mexico  |  -.0722432   .0002705  -267.11   0.000    -.0727864      -.0717
    new york  |  -.0350971   .0005965   -58.84   0.000    -.0362952   -.0338991
north caro..  |  -.0482693   .0216882    -2.23   0.031    -.0918313   -.0047073
north dakota  |   .1224431   .0001484   825.07   0.000      .122145    .1227412
        ohio  |  -.0354814   .0005419   -65.47   0.000    -.0365698   -.0343929
    oklahoma  |  -.0589376   .0001088  -541.54   0.000    -.0591562    -.058719
      oregon  |   .0254931   .0183625     1.39   0.171     -.011389    .0623752
pennsylvania  |  -.0779009   .0004892  -159.23   0.000    -.0788836   -.0769183
rhode island  |  -.0025628   .0194303    -0.13   0.896    -.0415897    .0364642
south caro..  |   -.080478   .0003175  -253.45   0.000    -.0811158   -.0798403
south dakota  |   .0095469    .000185    51.60   0.000     .0091753    .0099185
   tennessee  |  -.0659307   .0002355  -279.96   0.000    -.0664038   -.0654577
       texas  |  -.0566178    .000174  -325.43   0.000    -.0569673   -.0562684
        utah  |  -.0566532    .000101  -561.05   0.000     -.056856   -.0564504
     vermont  |   .0083417    .000756    11.03   0.000     .0068232    .0098602
    virginia  |  -.0465189   .0003345  -139.09   0.000    -.0471907   -.0458472
  washington  |  -.0174756   .0005158   -33.88   0.000    -.0185116   -.0164396
west virgi..  |  -.0852678   .0001112  -767.10   0.000    -.0854911   -.0850446
   wisconsin  |   .0492858    .000379   130.03   0.000     .0485245    .0500471
     wyoming  |  -.0742549   .0005674  -130.88   0.000    -.0753945   -.0731153
              |
         year |
        1984  |   .0468768   .0046794    10.02   0.000     .0374778    .0562757
        1986  |   .0061268   .0045516     1.35   0.184    -.0030153     .015269
        1988  |   .0314728   .0059299     5.31   0.000     .0195622    .0433834
        1990  |    .002353   .0059247     0.40   0.693    -.0095471    .0142532
        1992  |   .0634687   .0061362    10.34   0.000     .0511439    .0757936
        1994  |   .0048732    .006623     0.74   0.465    -.0084294    .0181758
        1996  |   .0523482   .0073225     7.15   0.000     .0376406    .0670558
        1998  |   .0245392   .0077904     3.15   0.003     .0088917    .0401867
        2000  |   .0675903   .0077751     8.69   0.000     .0519736    .0832071
        2002  |   .0354962   .0086831     4.09   0.000     .0180556    .0529368
        2004  |   .0979956   .0076852    12.75   0.000     .0825594    .1134318
        2006  |   .0593739   .0088469     6.71   0.000     .0416044    .0771435
        2008  |   .1070713   .0090163    11.88   0.000     .0889616     .125181
        2010  |   .0636579    .008792     7.24   0.000     .0459987    .0813171
        2012  |   .1027669   .0095466    10.76   0.000     .0835919    .1219418
        2014  |   .0602205   .0098251     6.13   0.000     .0404863    .0799547
              |
        _cons |   .7464395   .0056862   131.27   0.000     .7350184    .7578606
-------------------------------------------------------------------------------

Comment

Clyde Schechter

Join Date: Apr 2014

Posts: 29954
#44

27 Apr 2021, 09:37

The second results you show look very close to what I was expecting, and I apologize for having suggested ## here, when # is easier to work with in this situation. All the 0#... empty notifications are expected: by construction, when ps = 0, event_time_var is always 0 (labeled as -5). The only concern I have is with -1.ps#8.event_time_var omitted because of collinearity-. It may not be a problem, but I can't tell just from the regression output. Please show the results of:

Code:

tab event_time_var ps if e(sample)
Comment
Cairone Federica

Join Date: Apr 2021

Posts: 31
#45

27 Apr 2021, 10:21

Here the results:

Code:

. tab event_time_var ps if e(sample) event_time | ps _var | 0 1 | Total -----------+----------------------+---------- -5 | 971,329 265,220 | 1,236,549 -4 | 0 21,081 | 21,081 -2 | 0 21,215 | 21,215 0 | 0 23,008 | 23,008 2 | 0 19,227 | 19,227 3 | 0 29,457 | 29,457 -----------+----------------------+---------- Total | 971,329 379,208 | 1,350,537

Then, using # instead of ## gives me a different result, but yes I prefer

regress register i.ps#i.event_time_var i.statefip i.year, cluster(statefip)

as you. Concerning t=-1, is the omitted election year, thus I think it is not a proper problem the fact it is absent; at the same time, as shown in #35 they compute the mean of this omitted time variable, even if it does not appear in any regression.
More, in the original paper, they used

[pweight= wtfinl]

, I was wondering if I should do the same...
Comment

Announcement

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment