Expected behavior: I am trying to run a difference in differences model using reghdfe with multiple fixed effects.
Actual behavior: The standard errors are not calculating and several variables are being omitted due to collinearity. However, when my colleague runs the exact same code, his calculates with no observable problems. Beginning of output is below.
. reghdfe std_score i.intervention##i.post asian black hispanic white other ell ie gt i.grade $sch_var [pw = ps_all], absorb(studentnumber sem_by_year schoolnum) vce(cluster schoolnum)
(dropped 574 singleton observations)
note: 1bn.intervention is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: 1bn.post is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: asian is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: black is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: hispanic is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: white is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: other is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: ell is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: iep is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
note: gt is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09)
(MWFE estimator converged in 44 iterations)
note: 1.intervention omitted because of collinearity
note: 1.post omitted because of collinearity
note: asian omitted because of collinearity
note: black omitted because of collinearity
note: hispanic omitted because of collinearity
note: white omitted because of collinearity
note: other omitted because of collinearity
note: ell omitted because of collinearity
note: iep omitted because of collinearity
note: gt omitted because of collinearity
Warning: variance matrix is nonsymmetric or highly singular
Do you know why I would be getting collinearity problems when others can seem to run this code exactly as is?
Stata version: 15.1
OS: macOS Catalina
Expected behavior: I am trying to run a difference in differences model using reghdfe with multiple fixed effects.
Actual behavior: The standard errors are not calculating and several variables are being omitted due to collinearity. However, when my colleague runs the exact same code, his calculates with no observable problems. Beginning of output is below.
. reghdfe std_score i.intervention##i.post asian black hispanic white other ell ie gt i.grade $sch_var [pw = ps_all], absorb(studentnumber sem_by_year schoolnum) vce(cluster schoolnum) (dropped 574 singleton observations) note: 1bn.intervention is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: 1bn.post is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: asian is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: black is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: hispanic is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: white is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: other is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: ell is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: iep is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) note: gt is probably collinear with the fixed effects (all partialled-out values are close to zero; tol = 1.0e-09) (MWFE estimator converged in 44 iterations) note: 1.intervention omitted because of collinearity note: 1.post omitted because of collinearity note: asian omitted because of collinearity note: black omitted because of collinearity note: hispanic omitted because of collinearity note: white omitted because of collinearity note: other omitted because of collinearity note: ell omitted because of collinearity note: iep omitted because of collinearity note: gt omitted because of collinearity Warning: variance matrix is nonsymmetric or highly singular
Do you know why I would be getting collinearity problems when others can seem to run this code exactly as is?
Thanks for your help.