Why do Random Effects model use Z-statistics instead of T-statistics?

[Nana Lee]

Hi there.
Based on Chow, Hausman, and LM test, it is decided that RE is a good fit for one of my models.
I did a regression using RE command such as:

xtreg Y X1 X2, re vce(cluster ID)

Result displays in Z stats instead of T stats. Is there any particular reason for that?
Because, isn't Z test supposed to be used when population variance is known?


Best Regards.

[Maarten Buis]

Random effects models are estimated with Maximum Likelihood rather than Ordinary Least Squares. With ML we don't have the nice small sample results that allow us to use the t-distribution, but we have asymptotic results that allows us to use the standard normal distribution.

[Daffa Favian]

Hi guys.
My lecturer asks me to find the value of t stats from a RE model. Is it even possible? If it is possible, how do I find the t stats from a RE model?
I tried to compare t-stats from Common Effect and z-stats Random Effect and the result is slightly similar. Is it actually the same? Because I already read the explanation above and I'm confused why my lecturer asks me to find the t stats from a RE model.
Thank you.

[Joseph Coveney]

Is it even possible? If it is possible, how do I find the t stats from a RE model?

Yes. See below. Begin at the "Begin here" comment. t statistics are in the regression table, ANOVA parameterization F statistics are in the contrasts that follow.

.ÿ
.ÿversionÿ15.1

.ÿ
.ÿclearÿ*

.ÿ
.ÿsetÿseedÿ`=strreverse("1494467")'

.ÿquietlyÿsetÿobsÿ14

.ÿ
.ÿgenerateÿbyteÿpidÿ=ÿ_n

.ÿgenerateÿdoubleÿpid_uÿ=ÿrnormal()

.ÿ
.ÿgenerateÿbyteÿtrtÿ=ÿmod(_n,ÿ2)

.ÿ
.ÿquietlyÿexpandÿ2

.ÿbysortÿpid:ÿgenerateÿbyteÿtimÿ=ÿ_n

.ÿ
.ÿgenerateÿdoubleÿoutÿ=ÿpid_uÿ+ÿrnormal()

.ÿ
.ÿ*
.ÿ*ÿBeginÿhere
.ÿ*
.ÿanovaÿoutÿtrtÿ/ÿpid|trtÿtimÿtrt#tim

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿ=ÿÿÿÿÿÿÿÿÿ28ÿÿÿÿR-squaredÿÿÿÿÿ=ÿÿ0.6888
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿRootÿMSEÿÿÿÿÿÿ=ÿÿÿÿ1.48548ÿÿÿÿAdjÿR-squaredÿ=ÿÿ0.2999

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿSourceÿ|ÿPartialÿSSÿÿÿÿÿÿÿÿÿdfÿÿÿÿÿÿÿÿÿMSÿÿÿÿÿÿÿÿFÿÿÿÿProb>F
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿ-----------+----------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿModelÿ|ÿÿÿ58.62076ÿÿÿÿÿÿÿÿÿ15ÿÿÿ3.9080507ÿÿÿÿÿÿ1.77ÿÿ0.1619
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿtrtÿ|ÿÿ3.6011394ÿÿÿÿÿÿÿÿÿÿ1ÿÿÿ3.6011394ÿÿÿÿÿÿ0.87ÿÿ0.3689
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿpid|trtÿ|ÿÿ49.567789ÿÿÿÿÿÿÿÿÿ12ÿÿÿ4.1306491ÿÿ
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿ-----------+----------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿtimÿ|ÿÿ.72239349ÿÿÿÿÿÿÿÿÿÿ1ÿÿÿ.72239349ÿÿÿÿÿÿ0.33ÿÿ0.5778
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿtrt#timÿ|ÿÿ4.7294383ÿÿÿÿÿÿÿÿÿÿ1ÿÿÿ4.7294383ÿÿÿÿÿÿ2.14ÿÿ0.1689
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿResidualÿ|ÿÿ26.479827ÿÿÿÿÿÿÿÿÿ12ÿÿÿ2.2066522ÿÿ
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿ-----------+----------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿTotalÿ|ÿÿ85.100587ÿÿÿÿÿÿÿÿÿ27ÿÿÿ3.1518736ÿÿ

.ÿ
.ÿmixedÿoutÿi.trt##i.timÿ||ÿpid:ÿ,ÿremlÿdfmethod(satterthwaite)ÿnolrtestÿnolog

Mixed-effectsÿREMLÿregressionÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿÿÿÿÿ=ÿÿÿÿÿÿÿÿÿ28
Groupÿvariable:ÿpidÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿgroupsÿÿ=ÿÿÿÿÿÿÿÿÿ14

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿObsÿperÿgroup:
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿminÿ=ÿÿÿÿÿÿÿÿÿÿ2
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿavgÿ=ÿÿÿÿÿÿÿÿ2.0
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿmaxÿ=ÿÿÿÿÿÿÿÿÿÿ2
DFÿmethod:ÿSatterthwaiteÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿDF:ÿÿÿÿÿÿÿÿÿÿÿminÿ=ÿÿÿÿÿÿ12.00
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿavgÿ=ÿÿÿÿÿÿ16.99
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿmaxÿ=ÿÿÿÿÿÿ21.97

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿF(3,ÿÿÿÿ14.88)ÿÿÿÿ=ÿÿÿÿÿÿÿ1.11
Logÿrestricted-likelihoodÿ=ÿ-51.205812ÿÿÿÿÿÿÿÿÿÿProbÿ>ÿFÿÿÿÿÿÿÿÿÿÿ=ÿÿÿÿÿ0.3747

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿoutÿ|ÿÿÿÿÿÿCoef.ÿÿÿStd.ÿErr.ÿÿÿÿÿÿtÿÿÿÿP>|t|ÿÿÿÿÿ[95%ÿConf.ÿInterval]
-------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿ1.trtÿ|ÿÿÿ-1.53922ÿÿÿ.9514877ÿÿÿÿ-1.62ÿÿÿ0.120ÿÿÿÿ-3.512617ÿÿÿÿ.4341768
ÿÿÿÿÿÿÿ2.timÿ|ÿÿ-.5007235ÿÿÿ.7940227ÿÿÿÿ-0.63ÿÿÿ0.540ÿÿÿÿÿ-2.23075ÿÿÿÿ1.229303
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿtrt#timÿ|
ÿÿÿÿÿÿÿÿ1ÿ2ÿÿ|ÿÿÿ1.643939ÿÿÿ1.122918ÿÿÿÿÿ1.46ÿÿÿ0.169ÿÿÿÿ-.8026882ÿÿÿÿ4.090567
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿ_consÿ|ÿÿÿ.3715206ÿÿÿ.6728034ÿÿÿÿÿ0.55ÿÿÿ0.586ÿÿÿÿ-1.023882ÿÿÿÿ1.766923
------------------------------------------------------------------------------

------------------------------------------------------------------------------
ÿÿRandom-effectsÿParametersÿÿ|ÿÿÿEstimateÿÿÿStd.ÿErr.ÿÿÿÿÿ[95%ÿConf.ÿInterval]
-----------------------------+------------------------------------------------
pid:ÿIdentityÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿvar(_cons)ÿ|ÿÿÿ.9619984ÿÿÿ.9559371ÿÿÿÿÿÿ.1371943ÿÿÿÿ6.745478
-----------------------------+------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿvar(Residual)ÿ|ÿÿÿ2.206652ÿÿÿÿ.900862ÿÿÿÿÿÿ.9913621ÿÿÿÿ4.911741
------------------------------------------------------------------------------

.ÿcontrastÿtrtÿtimÿtrt#tim,ÿsmall

Contrastsÿofÿmarginalÿlinearÿpredictions

Marginsÿÿÿÿÿÿ:ÿasbalanced

-----------------------------------------------------------
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿÿÿÿÿdfÿÿÿÿÿÿÿÿddfÿÿÿÿÿÿÿÿÿÿÿFÿÿÿÿÿÿÿÿP>F
-------------+---------------------------------------------
outÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿtrtÿ|ÿÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ12.00ÿÿÿÿÿÿÿÿ0.87ÿÿÿÿÿ0.3689
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿÿÿÿÿtimÿ|ÿÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ12.00ÿÿÿÿÿÿÿÿ0.33ÿÿÿÿÿ0.5778
ÿÿÿÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿÿÿtrt#timÿ|ÿÿÿÿÿÿÿÿÿÿ1ÿÿÿÿÿÿ12.00ÿÿÿÿÿÿÿÿ2.14ÿÿÿÿÿ0.1689
-----------------------------------------------------------

.ÿ
.ÿexit

endÿofÿdo-file


.

[Daffa Favian]

Thanks Joseph! However, I still don't understand. Could you explain more in a simple way?