Lars,
If male menstruation onset is treated like "missing data" and replacement
estimates are made, it will lead to inflated statistical power. My sense is
that in this instance where data doesn't apply, male and female cases need
to be analysed separately.
Max
-----Original Message-----
From: A UK-based worldwide e-mail broadcast system mailing list
[mailto:[log in to unmask]] On Behalf Of Lars Chi
Sent: 20 January 2011 01:46
To: [log in to unmask]
Subject: Query: Dealing with "conditional" input variables in Regressionþ
Hi,
I'd appreciate your help on the following. As a simple example, suppose I
have two independent variables x_1 = {Female, Male} and x_2 = {Age at which
menstruation began}. Obviously, X_2 will only be populated if X_1=Female.
Suppose I want to estimate the impact of x_1 and x_2 on some dependent
variable y. How should I code these variables in the regression analysis?
My methodology is the following: replace "missing" observations in x_2
(i.e., those corresponding to males) with zeros, and set x_1=1 if x_2 is
zero and x_1=0 otherwise. Then fit y = beta_1*x_1 + beta_2*x_2. Does this
sound right?
Thanks in advance for your help,
Lars.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|