Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | William Buchanan <william@williambuchanan.net> |
To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
Subject | Re: st: Checking to see if the association between two variables is linear or otherwise |
Date | Sat, 13 Oct 2012 07:32:55 -0700 |
I'm not sure which suggesting you're referencing. All I did was mention that there was an example of a similar idea in one of the Stata Press books and made no qualitative assumptions regarding the merits and/or shortcomings of the methodology. -Billy Sent from my iPhone On Oct 13, 2012, at 0:25, "Justina Fischer" <JAVFischer@gmx.de> wrote: > your suggestion does not resolve the problem that years sof education measure degree very inacurately. > > In A US context, 12 and 16 years of education most likely relate to high school degree and bachelor's degree attained four years later. > > That Mitchel find splines at these nots suggests that using educational-level-dummies is a very good idea indeed. > > Justina > > -------- Original-Nachricht -------- >> Datum: Fri, 12 Oct 2012 20:25:54 -0700 >> Von: William Buchanan <william@williambuchanan.net> >> An: "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> >> Betreff: Re: st: Checking to see if the association between two variables is linear or otherwise > >> It's one of the examples in Michael Mitchell's book on data visualization. >> He used knots at 12 and 16 years of education to illustrate how you could >> use splines and the different commands to show the different relationships >> between years of education and earnings. >> >> - Billy >> >> Sent from my iPhone >> >> On Oct 12, 2012, at 20:04, "JVerkuilen (Gmail)" <jvverkuilen@gmail.com> >> wrote: >> >>> On Fri, Oct 12, 2012 at 10:46 PM, David Kantor <kantor.d@att.net> wrote: >>> >>>> I agree, noting that highest grade attained is better than years in >>>> education, but it is still not appropriate, as it is not an interval >> scale. >>>> The "same" increments between values do not have the same significance. >>>> For example, the significance of the increment from grade 11 to 12 is >> not >>>> the same as from 10 to 11. >>> >>> Interesting point. I bet this could probably be ameliorated using >>> regression splines with knots at key grade milestones. >>> >>> Jay >>> -- >>> JVVerkuilen, PhD >>> jvverkuilen@gmail.com >>> >>> "Out beyond ideas of wrong-doing and right-doing there is a field. >>> I'll meet you there. When the soul lies down in that grass the world >>> is too full to talk about." ---Rumi >>> * >>> * For searches and help try: >>> * http://www.stata.com/help.cgi?search >>> * http://www.stata.com/support/faqs/resources/statalist-faq/ >>> * http://www.ats.ucla.edu/stat/stata/ >> >> * >> * For searches and help try: >> * http://www.stata.com/help.cgi?search >> * http://www.stata.com/support/faqs/resources/statalist-faq/ >> * http://www.ats.ucla.edu/stat/stata/ > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/faqs/resources/statalist-faq/ > * http://www.ats.ucla.edu/stat/stata/ * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/