Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: comparing nested models

From   Austin Nichols <>
Subject   Re: st: comparing nested models
Date   Thu, 3 Feb 2011 11:59:51 -0500

Neil Shephard <> :
You would certainly not get a better fit by specifying
. reg wage age
instead of
. reg wage i.age
but then one must balance parsimony against parametric assumptions.
Note also that you can test your parametric assumptions, but that is
related to what the poster was asking:

sysuse nlsw88, clear
reg wage age i.age
test 35.age 36.age 37.age 38.age 39.age 40.age 41.age 42.age 43.age
44.age 45.age
reg wage c.age##c.age##c.age
test c.age#c.age c.age#c.age#c.age

Most would assume log wages are a quadratic or higher order polynomial
function of age without much theoretical justification (google "Mincer
wage equation" for discussion), but note that age is categorical in
every dataset already...  Though I suppose it could be measured with
millisecond precision as of Stata 10.

On Thu, Feb 3, 2011 at 11:36 AM, Neil Shephard <> wrote:
> On Thu, Feb 3, 2011 at 4:09 PM, Nikolas Asasa <> wrote:
>> Dear All,
>> I have two models aiming to estimate effect of age on wage. In the first model age is recorded in 5 categories. In the second model age is recorded in 10 categories.
>> The question is how to compare goodness of fit of two models?
>> In principle I would apply F-test, but I am not sure whether "test" procedure in Stata allows me to do it? Should I use ANOVA instead?
> If you have the actual age of individuals would you not get the best
> fit by simply using all of the information on age and wage...
> wage = a*age + b
> ...under a linear regression?
> There are many, many reasons why you shouldn't classify continuous
> variables into categorical.  Frank Harrell's page somes many of these
> up nicely
> Neil

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index