Notice: On March 31, it was **announced** that Statalist is moving from an email list to a **forum**. The old list will shut down at the end of May, and its replacement, **statalist.org** is already up and running.

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

From |
"JVerkuilen (Gmail)" <jvverkuilen@gmail.com> |

To |
statalist@hsphsun2.harvard.edu |

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

Date |
Fri, 12 Oct 2012 20:36:01 -0400 |

On Fri, Oct 12, 2012 at 5:56 PM, Amal Khanolkar <Amal.Khanolkar@ki.se> wrote: > 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. > You are making a prediction that involves both location (i.e., the mean) and the dispersion (i.e., the variance). There are a number of models that can accommodate this kind of pattern. , but you might want to consider using simultaneous quantile regression (-sqreg-). I also did some looking and found -reghv-. This allows you to use heteroscedasticity covariates in a regression model with multiplicative variance. Finally, you might consider -betafit-, which similarly allows the use of heteroscedasticity covariates, though you'd need to linearly rescale the data to be in the unit interval. * * 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/

**References**:**st: Checking to see if the association between two variables is linear or otherwise***From:*Amal Khanolkar <Amal.Khanolkar@ki.se>

- Prev by Date:
**Re: RE: st: Checking to see if the association between two variables is linear or otherwise** - Next by Date:
**Re: st: RE: How do I properly tell Stata this data is Panel Data?** - Previous by thread:
**Re: RE: st: Checking to see if the association between two variables is linear or otherwise** - Next by thread:
**st: How do I properly tell Stata this data is Panel Data?** - Index(es):