Adjusting for covariates

Jeff Leek

@jtleek

www.jtleek.com

http://apps.who.int/gho/athena/data/GHO/WHOSIS_000008.csv?profile=text&filter=COUNTRY:;SEX:

https://github.com/jtleek/dataanalysis/blob/master/week4/006multipleVariables/

https://github.com/jtleek/dataanalysis/blob/master/week4/006multipleVariables/

Hu_i= b₀ + b₁ Y_i + e_i

https://github.com/jtleek/dataanalysis/blob/master/week4/006multipleVariables/

https://github.com/jtleek/dataanalysis/blob/master/week4/006multipleVariables/

Hu_i= b₀ + b₁ Y_i +b₂ F_i + e_i

b₀ - percent hungry at year zero for males

b₀ + b₂ - percent hungry at year zero for females

b₁ - change in percent hungry (for either males or females)

in one year

e_i - everything we didn't measure

​

Slide courtesy: Ingo Ruczinski

Notes and further reading

• Linear models is a whole class (no joke): https://www.coursera.org/course/regmods
• Basic thing to keep in mind is how many levels do you want to fit? What makes sense biologically?
• Great additional notes in Chapter 2 here: http://genomicsclass.github.io/book/

​