Re: RE: st: Checking to see if the association between two variables is linear or otherwise

["Justina Fischer" <JAVFischer@gmx.de>]

Hi Amal,

from the education-growth literature I know that using educational levels is preferred over years in education....years in education can be very misleading (if you skipped classes, repeated classes, dropped out, etc). In addition, it is the degree that qualifies you for (certain) jobs.

The dummy variable approaches allows then the relation BMI-education to be different by educational level - in a way, this accounts for many different relations that are not linear.

There is a difference between 'linear relation' and using a 'linear model' - I am talking about (non)linear relations in a linear model (OLS), assuming your dependent variable is continuous.

The 'broad' histogramm for low educational levels may yield an insignificant coefficient for the low-education-dummy.

so try

xi: reg BMI i.education age gender (i.)income etc.  ...

and if you want you can post the results.

HTH

Justina

-------- Original-Nachricht --------
> Datum: Fri, 12 Oct 2012 23:29:45 +0000
> Von: Amal Khanolkar <Amal.Khanolkar@ki.se>
> 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

> Hi Justina,
> 
> Education is measured in years and the kind of degree qualification
> obtained - I have 6 categories of education: high school; no diploma, high
> graduate, some college, bachelors degree, masters degree etc
> 
> I see that if I plot histograms for BMI separately for each educational
> group - the distribution of BMI gets narrower as educational level increases
> (as expected) - which made me wonder that assuming a linear relationship
> between education and BMI (and thus using linear regression) may not be the
> best choice.
> 
> Do you mean, that I don't have to worry about the relationship being not
> linear then...?
> 
> Thanks,
> 
> /Amal,
> ________________________________________
> From: owner-statalist@hsphsun2.harvard.edu
> [owner-statalist@hsphsun2.harvard.edu] on behalf of Justina Fischer [JAVFischer@gmx.de]
> Sent: 13 October 2012 00:55
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: Checking to see if the association between two variables
> is linear or otherwise
> 
> Hi
> 
> unless education is measured in years I would create a set of dummy
> variables in educational level (i.education).
> 
> The effect of low education on BMI will then be driven by the systematic
> relation with BMI that exists for the _majority_ of the low-edu-population.
> 
> Pls do not forget to control for income - low education might correlate
> with low income, and low income earners may not afford healthy food.
> 
> HTH
> 
> Justina
> 
> 
> -------- Original-Nachricht --------
> > Datum: Fri, 12 Oct 2012 21:56:35 +0000
> > Von: Amal Khanolkar <Amal.Khanolkar@ki.se>
> > An: "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
> > Betreff: st: Checking to see if the association between two variables is
> linear or otherwise
> 
> > Hi,
> >
> >
> > I'm trying to figure out if linear regression is the appropriate choice
> > for my research question - I would like to analyze the association of
> BMI and
> > education (BMI is continuous and education categorical). Ideally I would
> > just run a linear regression with BMI as the outcome and education as
> the
> > principle explanatory variable.
> >
> > However my hypothesis is that low educated people are both likely to
> have
> > a low and a high BMI, i.e. the association between education and BMI is
> > probably more 'u shaped' than linear.
> >
> > What is the best way to check if the association between a continuous
> and
> > categorical variable is linear or otherwise...? Preferably, I would like
> to
> > be able to plot such a shape using Stata.
> >
> > Thanks,
> >
> > /Amal.
> > *
> > *   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/