Raw Message

Message-ID: <CALhdq6vbT-rDraXx8rpxBTWdQCYB3Xrj4Hz_naeemPp0DR7_vA@mail.gmail.com>
Date: 2014-02-25T14:54:11Z
From: Peter Crowther
Subject: mcr mcreg: Why are BCa and quantile se values calculated, stored, then the stored values set to NA?

In mcreg, there's the following snippet for bootstrap CI method of BCa
(and a very similar one for quantile):

else if (method.bootstrap.ci == "BCa") {
            bootB0 <- mc.calc.bca(Xboot = B0, Xjack = B0jack,
                xhat = glob.coef[1], alpha)
            bootB1 <- mc.calc.bca(Xboot = B1, Xjack = B1jack,
                xhat = glob.coef[2], alpha)
            bootB0$se <- glob.sigma[1]
            bootB1$se <- glob.sigma[2]
            bootB0$se <- NA
            bootB1$se <- NA
        }
[...]
        mc.res <- mc.make.CIframe(b0 = glob.coef[1], b1 = glob.coef[2],
            se.b0 = bootB0$se, se.b1 = bootB1$se, CI.b0 = bootB0$CI,
            CI.b1 = bootB1$CI)

I'd quite like to use in my code the SE values that were calculated in
mcreg.  Can any kind soul shed light on why they're calculated,
stored, then the stored values set to NA?

Thanks in advance for any insight!

- Peter
--
Peter Crowther, Director, Melandra Limited