Hi Alisa, you have either an approximately randomized experimental or, equivalently, an observational (nonexperimental) study.  CEM usually works well in that situation.

Gary
--
Gary King - Albert J. Weatherhead III University Professor - Director, IQSS - Harvard University
GaryKing.org - King@Harvard.edu - @KingGary - 617-500-7570 - Assistant: 617-495-9271

On Mon, Jul 25, 2016 at 3:55 PM, Padon, Alisa A <alisa.padon@asc.upenn.edu> wrote:

Hi,

I have a few questions regarding using CEM on an already randomized sample that I would appreciate any help with.

Briefly, I conducted an online experiment in which we randomized respondents to 1 of 3 conditions. After exposure, some respondents dropped out before the final outcome measures. In total, I have 417 respondents who completed the full study, but because of the drop-outs, there is unequal distribution of some key variables between conditions.

My questions are:

1) Would CEM be an appropriate solution to an unequal condition problem like this?

2) Does the use of the method invalidate the randomization?

3) Is there any literature that uses CEM in an already randomized sample?

Thanks in advance,

Alisa


-
--
cem Mailing List, served by HUIT
Send messages: cem@lists.gking.harvard.edu
[un]subscribe Options: http://lists.gking.harvard.edu/?info=cem
More information on cem: http://gking.harvard.edu/cem
Cem mailing list
Cem@lists.gking.harvard.edu

To unsubscribe from this list or get other information:

https://lists.gking.harvard.edu/mailman/listinfo/cem