On 18 Mar 2004 05:59:22 -0800, [EMAIL PROTECTED] (Phillip Good) wrote: > As I suspected you are unable to offer a single justification > for your actions just as Rumsfield was unable to offer a single > justification for his. Given your recklessness, may I recommend > "Common Errors in Statistics (and How to Avoid Them)," Wiley, 2003. > > Phillip Good [ snip, 145 lines of previous stuff - apparently starting with the topic of Maximum Likelihood solutions.]
Top-posting is not the usual custom in the sci.stat.* groups. Snipping out extraneous comments is a friendly custom -- I am seldom pleased to see a 119-line post followed by 151-line post, which is the same thing plus headers /tags plus 4 lines of comment. Maximum Likelihood, Phil, *is* a really big workhorse of statistical estimation, partly because the ML solutions share certain fine characteristics ... which I don't call to mind. Now, were you saying that there are specific drawbacks that make MLE (the solutions of today, ones which happen to be popular) poor for the "Subject:" of Missing Data? -- I have not heard of that, but I could listen to argument. But I guess I would be a tough sell, if you are saying there is something vastly wrong with MLE, in general. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html - I need a new job, after March 31. Openings? - . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
