Dear Stata users,
I have a field experiment on the impact of nudges on the use of polluting vehicles among a population of individuals. I have a control group and a Placebo group. I want to estimate if the nudge works (I use Linear Probability Model) but also to estimate threshold above which in terms of distance the effect disappears.
So I run the following estimation:
reghdfe Polluting_vehic treated_AM##period1##c.ldist , absorb(Id_user id_site time2) cluster(Id_user)
Id_user are individuals, id_site the firm and time2 the time (day week). Period1 is the variable capturing the treatment period.
I have the following result:
HDFE Linear regression Number of obs = 30,479
Absorbing 3 HDFE groups F( 3, 122) = 3.31
Statistics robust to heteroskedasticity Prob > F = 0.0223
R-squared = 0.8110
Adj R-squared = 0.8077
Within R-sq. = 0.0015
Number of clusters (Id_user) = 123 Root MSE = 0.2130
(Std. Err. adjusted for 123 clusters in Id_user)
--------------------------------------------------------------------------------------------
| Robust
Polluting_vehic | Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------------------------+----------------------------------------------------------------
1.treated_AM | 0 (omitted)
1.period1 | 0 (omitted)
|
treated_AM#period1 |
1 1 | -.0709034 .0242027 -2.93 0.004 -.118815 -.0229917
|
ldist | 0 (omitted)
|
treated_AM#c.ldist |
1 | 0 (omitted)
|
period1#c.ldist |
1 | .0004862 .0009643 0.50 0.615 -.0014228 .0023951
|
treated_AM#period1#c.ldist |
1 1 | .0150103 .0058533 2.56 0.012 .0034231 .0265975
|
_cons | .6267646 .0030639 204.57 0.000 .6206993 .6328298
--------------------------------------------------------------------------------------------
Absorbed degrees of freedom:
-----------------------------------------------------+
Absorbed FE | Categories - Redundant = Num. Coefs |
-------------+---------------------------------------|
Id_user | 123 123 0 *|
id_site | 27 0 27 |
time2 | 378 1 377 |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation
.
How can I compute the Threshold regarding the distance variable?
Thank you for your help.
I have a field experiment on the impact of nudges on the use of polluting vehicles among a population of individuals. I have a control group and a Placebo group. I want to estimate if the nudge works (I use Linear Probability Model) but also to estimate threshold above which in terms of distance the effect disappears.
So I run the following estimation:
reghdfe Polluting_vehic treated_AM##period1##c.ldist , absorb(Id_user id_site time2) cluster(Id_user)
Id_user are individuals, id_site the firm and time2 the time (day week). Period1 is the variable capturing the treatment period.
I have the following result:
HDFE Linear regression Number of obs = 30,479
Absorbing 3 HDFE groups F( 3, 122) = 3.31
Statistics robust to heteroskedasticity Prob > F = 0.0223
R-squared = 0.8110
Adj R-squared = 0.8077
Within R-sq. = 0.0015
Number of clusters (Id_user) = 123 Root MSE = 0.2130
(Std. Err. adjusted for 123 clusters in Id_user)
--------------------------------------------------------------------------------------------
| Robust
Polluting_vehic | Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------------------------+----------------------------------------------------------------
1.treated_AM | 0 (omitted)
1.period1 | 0 (omitted)
|
treated_AM#period1 |
1 1 | -.0709034 .0242027 -2.93 0.004 -.118815 -.0229917
|
ldist | 0 (omitted)
|
treated_AM#c.ldist |
1 | 0 (omitted)
|
period1#c.ldist |
1 | .0004862 .0009643 0.50 0.615 -.0014228 .0023951
|
treated_AM#period1#c.ldist |
1 1 | .0150103 .0058533 2.56 0.012 .0034231 .0265975
|
_cons | .6267646 .0030639 204.57 0.000 .6206993 .6328298
--------------------------------------------------------------------------------------------
Absorbed degrees of freedom:
-----------------------------------------------------+
Absorbed FE | Categories - Redundant = Num. Coefs |
-------------+---------------------------------------|
Id_user | 123 123 0 *|
id_site | 27 0 27 |
time2 | 378 1 377 |
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation
.
How can I compute the Threshold regarding the distance variable?
Thank you for your help.
