## 1. Matching programme (wh_match)

The following files contain downloadable STATA .ado files that implement a variant of the "simple matching estimator" proposed by Abadie and Imbens (2006). In contrast to the implementation by Abadie et al. (2004) this implementation is based primarily on MATA.

This routine implements goodness of fit tests for the normal and the Laplace distribution based on:

Stephens M A (1974) EDF statistics of fit and some comparisons, Journal of the American Statistical Association 69 (347) 703-737.

Puig P and Stephens M A (2000) Tests of fit for the laplace distribution with applications, Technometrics 42 (4) 417-424

Options:

d(lap/norm/all): if nothing is specified the GOF tests are run for both distributions

C(# obs): cuts 2*cut outliers symetrically from the left and right tail

Example 1: nola_gof var1, d(lap) C(5)

Example 2: nola_gof var1

This program was written by Dieter Gstach. It implements the Pedroni Group Z-rh, the Pedroni Group Z-t (PEDRONI 1999,2004), the Westerlund Group VRG and the Westerlund Group VRG Statistics (WESTERLUND 2005a).

Using the programme requires the residuals from an OLS regression, the number of regressors and the information wheter a trend should be taken into account or not.

Example 1: di_pcoint_wh e, reg(2) tr

Example 2: di_pcoint_wh e, reg(2)

Theoretically the statistics are valid only for balanced data. But here they are implemented also for unbalanced data using mean T_i values for standardizations.