Skip to content
Prev 20208 / 20628 Next
Original Raw

Standard errors of variances; negative variance

John Maindonald 
Dear  All,
As with the square root of -1, negative variances are a convenient 
computational fiction.  The square root of -1 is an invitation to move
to a 2-dimensional word, where they make perfect sense.

A negative variance estimate, in those systems that allow it, is a useful 
warning that one needs to think and move beyond the one-dimensional 
representation in which the variances have been set.
John Maindonald
Statistics Research Associates, Wellington NZ.

Thread (16 messages)

Will Hopkins Standard errors of variances; negative variance Jun 19 Ben Bolker Standard errors of variances; negative variance Jun 19 Thierry Onkelinx Standard errors of variances; negative variance Jun 20 Sorkin, John Standard errors of variances; negative variance Jun 20 Ben Bolker Standard errors of variances; negative variance Jun 20 James Pustejovsky Standard errors of variances; negative variance Jun 20 Will Hopkins Standard errors of variances; negative variance Jun 20 Ben Bolker Standard errors of variances; negative variance Jun 20 stevedrd m@iii@g oii y@hoo@com Standard errors of variances; negative variance Jun 21 John Maindonald Standard errors of variances; negative variance Jun 21 Will Hopkins Standard errors of variances; negative variance Jun 21 John Maindonald Standard errors of variances; negative variance Jun 21 J.D. Haltigan Standard errors of variances; negative variance Jun 21 Ben Pelzer Standard errors of variances; negative variance Jun 22 Will Hopkins Standard errors of variances; negative variance Jun 22 John Maindonald Standard errors of variances; negative variance Jun 22