From: mjwoo@niss.org
Sent: Wednesday, September 28, 2005 1:32 PM
To: Jerry Reiter
Cc: karr@niss.org; saki@stat.duke.edu; robin@stat.duke.edu;
aoganyan@niss.org; mjwoo@niss.org; veraf@niss.org
Subject: Re: Updates?

Hello Professor Jerry Reiter,

After the meeting, I transformed the data generated from skew-normal distribution in terms of rotations, and with respect to several masking methods. I applied the propensity score utility measure to them to figure out what happened to various rotated data and masked data. Also, I used other measures with the same data sets. The data has 1,000 observations, so the fitting algorithms didn't matter that much.

I turned to the data which was obtained from Anna (n=10,000). However, there is a fitting problem (convergency and memory) in "R" since sample size is too large. So, instead of R, I used SAS, and I could get more convergent results. Right now, I am trying to find out properties of each utility meausre. I hope I can be ready to show summary to all on friday.

Best regards,

Mi-Ja

> Hi All,
>
> I'm having problems with my car, but I'm going to try to get to NISS 
> on Friday.
>
> Are there any updates on the propensity score utility measure research 
> since we met two Fridays ago?
>
> Best,
> Jerry
>
>