NEAREST NEIGHBOUR PSM

[natiesim]

Dear all, I want to estimate the ATT using the nearest neigbour method (1 to 1). I have used psmatch2 for that purpose but the problem is that at the end of the estimation I get 430 individuals in the treatment group and 2452 in the control group on the common support . I am confused because I thought that when the matching is one to one, I am suppose to have the 430 individuals in the treatment group and 430 in the control group on the common support. Thanks for your answers.

[Navid Asgari]

Hi,

Nearest neighbor matching may use more than one observation from the control group with a member of the treatment group, depending on what type of nearest neighbor matching you use: http://www.cemmap.ac.uk/docs/psmatch2_features.pdf

[natiesim]

thanks Navid