http://www.johndcook.com/blog/2009/06/29/bancrofts-rule/ The blog of John D. Cook Sat, 11 Feb 2012 01:10:06 -0500 http://wordpress.org/?v=2.8.4 hourly 1 http://www.johndcook.com/blog/2009/06/29/bancrofts-rule/comment-page-1/#comment-48734 Anders Sat, 16 Oct 2010 22:42:04 +0000 http://www.johndcook.com/blog/?p=2556#comment-48734 Let me first say that it is an interesting post. I used to think that the Bancroft estimator is very arbitrary, but now I see that there is some logic to it. However, you only focus on the variance in your post. While the Bancroft estimator is still unbiased one should also consider that the resulting confidence interval might be centered further from the "truth" than the one resulting from the OLS estimator using the full sample (due to consistency of OLS). On the other hand my caveat just underscores your point about the virtue of the Bancroft estimator being limited to small samples. Let me first say that it is an interesting post. I used to think that the Bancroft estimator is very arbitrary, but now I see that there is some logic to it. However, you only focus on the variance in your post. While the Bancroft estimator is still unbiased one should also consider that the resulting confidence interval might be centered further from the “truth” than the one resulting from the OLS estimator using the full sample (due to consistency of OLS). On the other hand my caveat just underscores your point about the virtue of the Bancroft estimator being limited to small samples.

]]> http://www.johndcook.com/blog/2009/06/29/bancrofts-rule/comment-page-1/#comment-21010 La ley de Bancroft « Apuntes de Estadística Sat, 11 Jul 2009 18:19:32 +0000 http://www.johndcook.com/blog/?p=2556#comment-21010 [...] La ley de Bancroft Julio 11, 2009 Posted by psirusteam in Estadística, Modelos. trackback John D. Cook afirma que para encontrar la pendiente de una línea de regresión sobre un conjunto de datos, una muy buena aproximación es ajustar una línea que contemple el primer punto y el último punto. Esto se conoce como la ley de Bancroft (1944, 1964, 1972). Para más información ver la argumentación de John D. cook aquí. [...] [...] La ley de Bancroft Julio 11, 2009 Posted by psirusteam in Estadística, Modelos. trackback John D. Cook afirma que para encontrar la pendiente de una línea de regresión sobre un conjunto de datos, una muy buena aproximación es ajustar una línea que contemple el primer punto y el último punto. Esto se conoce como la ley de Bancroft (1944, 1964, 1972). Para más información ver la argumentación de John D. cook aquí. [...]

]]> http://www.johndcook.com/blog/2009/06/29/bancrofts-rule/comment-page-1/#comment-20280 Sharan Sharma Tue, 30 Jun 2009 13:39:43 +0000 http://www.johndcook.com/blog/?p=2556#comment-20280 Excellent post! Thank you. Excellent post! Thank you.

]]>