I do Research on the effects of employee stock ownership on individuals idea suggestions to a corporate idea suggestion scheme (Count). We have 5 years of observations (N~650,000 observations)
- NEWIDEA_A is the Count of ideas issued per year per employee
- ESO_ONLY_BINARY is a binary measure if employees particpiate in the firms employee stock ownership scheme (1 yes / 0 no) ~ 90,000 observations
- ESO_AND_OTHER_BINARY is a binary measure if employees participate in the firms employee stock ownership scheme and purchase additional stocks thorugh private accounts (1 yes / 0 no) ~ only 3900 observations
- Employees who purchae no stocks at all also exist in the data base an account for 70% of observations.
- Several controls are included such as AGE, TENURE, etc.
Code:
xtnbreg NEWIDEA_A AGE TENURE GENDER FULLTIME DUMMY_FUNCTION_1 DUMMY_FUNCTION_2 DUMMY_LEVEL_1 DUMMY_LEVEL_2 DUMMY_LEVEL_3 SIZE ESO_ONLY_BINARY OTHER_ONLY_BINARY ESO_AND_OTHER_BINARY i.YEAR, fe iterate (30)
Code:
. xtnbreg NEWIDEA_A AGE TENURE GENDER FULLTIME DUMMY_FUNCTION_1 DUMMY_FUNCTION_2 DUMMY_LEVEL_1 DUMMY_LEVEL_2 DUMMY_LEVEL_3 SIZE ESO_ONLY_BINARY ESO_AND_OTHER_BI
> NARY i.YEAR, fe iterate (30)
note: 9709 groups (9709 obs) dropped because of only one obs per group
note: 101545 groups (463225 obs) dropped because of all zero outcomes
Iteration 0: log likelihood = -134796.23
Iteration 1: log likelihood = -133837.24
Iteration 2: log likelihood = -133821.82
Iteration 3: log likelihood = -133821.82
Conditional FE negative binomial regression Number of obs = 179,289
Group variable: NEWID Number of groups = 37,538
Obs per group:
min = 2
avg = 4.8
max = 5
Wald chi2(16) = 2043.47
Log likelihood = -133821.82 Prob > chi2 = 0.0000
--------------------------------------------------------------------------------------
NEWIDEA_A | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------------+----------------------------------------------------------------
AGE | -.0060906 .0018842 -3.23 0.001 -.0097836 -.0023976
TENURE | .0147933 .0019399 7.63 0.000 .0109911 .0185955
GENDER | .2054615 .0584715 3.51 0.000 .0908594 .3200635
FULLTIME | .258968 .0405608 6.38 0.000 .1794702 .3384658
DUMMY_FUNCTION_1 | -.5244884 .0612952 -8.56 0.000 -.6446248 -.404352
DUMMY_FUNCTION_2 | -.3110061 .0334079 -9.31 0.000 -.3764844 -.2455278
DUMMY_LEVEL_1 | .3542703 .0270299 13.11 0.000 .3012926 .4072479
DUMMY_LEVEL_2 | .2987528 .0319901 9.34 0.000 .2360534 .3614521
DUMMY_LEVEL_3 | -.0189588 .0446852 -0.42 0.671 -.1065402 .0686225
SIZE | .0015458 .0004773 3.24 0.001 .0006104 .0024812
ESO_ONLY_BINARY | .1609867 .0166471 9.67 0.000 .1283589 .1936144
ESO_AND_OTHER_BINARY | .2091466 .0453593 4.61 0.000 .120244 .2980492
|
YEAR |
2011 | -.0667986 .0104762 -6.38 0.000 -.0873315 -.0462657
2012 | -.1407842 .0108313 -13.00 0.000 -.1620131 -.1195552
2013 | -.2304315 .0113434 -20.31 0.000 -.2526642 -.2081989
2014 | -.4157918 .0123889 -33.56 0.000 -.4400736 -.3915101
|
_cons | -.2494561 .0673473 -3.70 0.000 -.3814544 -.1174578
--------------------------------------------------------------------------------------
I expected to find the measure ESO_AND_OTHER_BINARY to be positive and statistically significant if it has a stronger effect on NEWIDEA_A than ESO_ONLY_BINARY. Results indicate no statistically significant difference.
Code:
drop if ESO_ONLY_BINARY==0 & ESO_AND_OTHER_BINARY==0 xtnbreg NEWIDEA_A AGE TENURE GENDER FULLTIME DUMMY_FUNCTION_1 DUMMY_FUNCTION_2 DUMMY_LEVEL_1 DUMMY_LEVEL_2 DUMMY_LEVEL_3 SIZE ESO_AND_OTHER_BINARY i.YEAR, fe iterate (30)
Code:
xtnbreg NEWIDEA_A AGE TENURE GENDER FULLTIME DUMMY_FUNCTION_1 DUMMY_FUNCTION_2 DUMMY_LEVEL_1 DUMMY_LEVEL_2 DUMMY_LEVEL_3 SIZE ESO_AND_OTHER_BINARY i.YEAR if N
> UMBERESO > 0, fe iterate (30)
note: 6721 groups (6721 obs) dropped because of only one obs per group
note: 19149 groups (67063 obs) dropped because of all zero outcomes
Iteration 0: log likelihood = -17664.959
Iteration 1: log likelihood = -17490.792
Iteration 2: log likelihood = -17487.008
Iteration 3: log likelihood = -17486.989
Iteration 4: log likelihood = -17486.989
Conditional FE negative binomial regression Number of obs = 24,778
Group variable: NEWID Number of groups = 6,571
Obs per group:
min = 2
avg = 3.8
max = 5
Wald chi2(15) = 381.35
Log likelihood = -17486.989 Prob > chi2 = 0.0000
--------------------------------------------------------------------------------------
NEWIDEA_A | Coef. Std. Err. z P>|z| [95% Conf. Interval]
---------------------+----------------------------------------------------------------
AGE | .001378 .0062992 0.22 0.827 -.0109683 .0137242
TENURE | .0074251 .0060829 1.22 0.222 -.0044972 .0193474
GENDER | .3475998 .2242667 1.55 0.121 -.0919549 .7871545
FULLTIME | .2080055 .1186018 1.75 0.079 -.0244499 .4404608
DUMMY_FUNCTION_1 | -.5581062 .1335297 -4.18 0.000 -.8198196 -.2963927
DUMMY_FUNCTION_2 | -.3730021 .0797736 -4.68 0.000 -.5293555 -.2166486
DUMMY_LEVEL_1 | .3499385 .0805792 4.34 0.000 .1920062 .5078708
DUMMY_LEVEL_2 | .4196076 .0883678 4.75 0.000 .2464098 .5928053
DUMMY_LEVEL_3 | .1176554 .1113332 1.06 0.291 -.1005537 .3358645
SIZE | .0021156 .0017175 1.23 0.218 -.0012507 .0054819
ESO_AND_OTHER_BINARY | -.0114105 .0487064 -0.23 0.815 -.1068733 .0840523
|
YEAR |
2011 | -.0816513 .0248654 -3.28 0.001 -.1303866 -.0329159
2012 | -.1998002 .0263144 -7.59 0.000 -.2513755 -.1482248
2013 | -.3118865 .0299882 -10.40 0.000 -.3706624 -.2531106
2014 | -.5787656 .0368524 -15.70 0.000 -.6509949 -.5065363
|
_cons | .0013934 .2245913 0.01 0.995 -.4387975 .4415844
--------------------------------------------------------------------------------------
My questions are:
- Is the way I used to test the two variables against each other appropriate or would you use a different approach to test if ESO_ONLY_BINARY and ESO_AND_OTHER_BINARY have statistically significnatly different effect magnitues on NEWIDEA_A?
- In case my approach is fine, how could I do a power analysis to rule out the concern that I do not find a difference because of a small amount of observations for ESO_AND_OTHER_BINARY (<4000 observations)?
Best
Felix
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input float NEWID double(YEAR NEWIDEA_A AGE TENURE GENDER FULLTIME) float(DUMMY_FUNCTION_1 DUMMY_FUNCTION_2 DUMMY_LEVEL_1 DUMMY_LEVEL_2 DUMMY_LEVEL_3) double SIZE float(ESO_ONLY_BINARY ESO_AND_OTHER_BINARY) 1 1 0 30 15 0 1 0 0 1 0 0 18 1 0 1 2 0 31 16 0 1 0 0 1 0 0 19 1 0 1 3 0 32 17 0 1 0 0 1 0 0 20 1 0 1 4 0 33 18 0 1 0 0 1 0 0 19 0 0 1 5 0 34 19 0 1 0 0 1 0 0 21 0 0 2 1 0 49 31 0 1 0 1 0 1 0 5 0 0 2 2 0 50 32 0 1 0 1 0 1 0 6 0 0 2 3 0 51 33 0 1 0 1 0 1 0 5 0 0 2 4 0 52 34 0 1 0 1 0 1 0 6 0 0 2 5 0 53 35 0 1 0 1 0 1 0 6 0 0 4 1 0 47 14 0 0 1 0 0 0 1 6 0 0 5 1 0 31 14 0 1 0 0 0 0 0 49 0 0 5 2 0 32 15 0 1 0 0 0 0 0 40 0 0 5 3 0 33 16 0 1 0 0 0 0 0 43 0 0 5 4 0 34 17 0 1 0 0 0 0 0 35 0 0 5 5 0 35 18 0 1 0 0 0 0 0 33 0 0 6 1 0 55 14 0 0 0 1 0 0 1 2 0 0 6 2 0 56 15 0 0 0 1 0 0 1 1 0 0 6 3 0 57 16 0 0 0 1 0 0 1 1 0 0 6 4 0 58 17 0 0 0 1 0 0 1 1 0 0 6 5 0 59 18 0 0 0 1 0 0 1 1 0 0 7 1 3 35 13 0 1 0 0 0 0 0 30 0 0 7 2 1 36 14 0 1 0 0 0 0 0 29 0 0 7 3 2 37 15 0 1 0 0 0 0 0 29 0 0 7 4 6 38 16 0 0 0 0 0 0 0 30 0 0 7 5 2 39 17 0 0 0 0 0 0 0 30 0 0 8 1 0 39 13 1 1 0 1 0 1 0 4 0 0 8 2 0 40 14 1 1 0 1 0 1 0 2 0 0 8 3 0 41 15 1 1 0 1 0 1 0 5 0 0 8 4 0 42 16 1 1 0 1 0 1 0 5 0 0 8 5 0 43 17 1 0 0 1 0 1 0 7 0 0 9 1 0 39 12 1 0 0 1 0 1 0 7 0 0 9 2 0 40 13 1 0 0 1 0 1 0 7 0 0 9 3 0 41 14 1 0 0 1 0 1 0 9 0 0 9 4 0 42 15 1 0 0 1 0 1 0 8 0 0 9 5 0 43 16 1 0 0 1 0 1 0 9 0 0 10 1 0 37 11 0 1 0 1 0 0 1 8 0 0 10 2 0 38 12 0 1 0 1 0 0 1 6 0 0 10 3 0 39 13 0 1 0 1 0 0 1 8 0 0 10 4 0 40 14 0 1 0 1 0 0 1 9 0 0 10 5 0 41 15 0 1 0 1 0 0 1 11 0 0 11 1 0 47 12 0 1 0 1 0 1 0 5 0 0 11 2 0 48 13 0 1 0 1 0 1 0 6 0 0 11 3 0 49 14 0 1 0 1 0 1 0 6 0 0 11 4 0 50 15 0 1 0 1 0 1 0 6 0 0 11 5 0 51 16 0 1 0 1 0 1 0 6 0 0 12 2 0 37 13 1 0 0 1 0 1 0 7 0 0 12 3 0 38 14 1 1 0 1 0 1 0 6 0 0 12 4 0 39 15 1 1 0 1 0 1 0 6 1 0 12 5 0 40 16 1 1 0 1 0 1 0 6 0 0 14 1 0 34 12 1 1 1 0 0 1 0 10 1 0 14 2 0 35 13 1 1 0 0 0 0 1 1 1 0 14 3 0 36 14 1 1 0 0 0 0 1 1 1 0 15 1 0 41 12 1 1 0 0 0 0 1 6 0 0 15 2 0 42 13 1 1 0 0 0 0 1 6 0 0 15 3 0 43 14 1 1 0 0 0 0 1 6 0 0 17 2 0 34 14 0 1 0 1 0 1 0 2 0 0 17 3 0 35 15 0 1 0 1 0 1 0 3 0 0 17 4 0 36 16 0 1 0 1 0 1 0 2 0 0 17 5 0 37 17 0 1 0 1 0 1 0 6 0 0 18 1 0 34 13 1 1 0 1 0 1 0 14 0 0 18 2 0 35 14 1 1 0 1 0 1 0 13 0 0 18 3 0 36 15 1 1 0 1 0 1 0 5 0 0 18 5 0 38 17 1 0 0 1 0 1 0 6 0 0 19 1 0 35 7 0 1 0 1 0 0 1 5 1 0 19 2 1 36 8 0 1 0 1 0 0 1 6 1 0 19 3 0 37 9 0 1 0 1 0 0 1 6 1 0 19 4 0 38 10 0 1 0 1 0 0 1 5 1 0 19 5 0 39 11 0 1 0 1 0 0 1 5 1 0 20 5 0 53 20 0 1 0 1 0 0 1 1 0 0 21 1 0 38 12 0 1 0 1 0 1 0 3 0 0 21 2 0 39 13 0 1 0 1 0 1 0 2 0 0 21 3 0 40 14 0 1 0 1 0 1 0 4 0 0 21 4 0 41 15 0 1 0 1 0 1 0 2 0 0 21 5 0 42 16 0 1 0 1 0 1 0 2 0 0 22 1 0 34 10 1 1 0 0 0 1 0 17 0 0 22 2 3 35 11 1 1 0 0 0 1 0 4 0 0 22 3 0 36 12 1 1 0 0 0 1 0 11 0 0 22 4 0 37 13 1 1 0 0 0 1 0 11 0 0 22 5 0 38 14 1 1 0 0 0 1 0 11 0 0 23 1 0 55 32 0 1 0 1 0 1 0 20 0 0 23 2 0 56 33 0 1 0 1 0 1 0 20 0 0 23 3 0 57 34 0 1 0 1 0 1 0 21 0 0 23 4 0 58 35 0 1 0 1 0 1 0 24 0 0 23 5 0 59 36 0 1 0 1 0 1 0 11 0 0 24 1 0 38 15 0 1 0 1 0 0 1 1 0 0 24 2 0 39 16 0 1 0 1 0 0 1 1 0 0 24 3 0 40 17 0 1 0 1 0 0 1 1 0 0 24 4 0 41 18 0 1 0 1 0 0 1 1 0 0 24 5 0 42 19 0 1 0 1 0 0 1 1 0 0 25 2 0 41 14 0 1 0 1 0 0 1 1 0 0 25 3 0 42 15 0 1 0 1 0 0 1 2 0 0 25 4 0 43 16 0 1 0 1 0 0 1 3 0 0 25 5 0 44 17 0 1 0 1 0 0 1 3 0 0 27 3 0 44 15 1 1 0 1 0 1 0 2 0 0 27 4 0 45 16 1 1 0 1 0 1 0 2 0 0 27 5 0 46 17 1 1 0 1 0 1 0 2 0 0 28 1 0 46 12 0 1 0 0 0 0 0 42 0 0 28 2 0 47 13 0 1 0 0 0 0 0 39 0 0 28 3 0 48 14 0 1 0 0 0 0 0 42 0 0 end label values YEAR YEAR label def YEAR 1 "2010", modify label def YEAR 2 "2011", modify label def YEAR 3 "2012", modify label def YEAR 4 "2013", modify label def YEAR 5 "2014", modify label values GENDER GENDER label def GENDER 0 "Male", modify label def GENDER 1 "Female", modify
Listed 100 out of 652223 observations
Use the count() option to list more
0 Response to Comparing two groups using xtnbreg + power analysis
Post a Comment