Are you sure you are looking at logistic regression?
I'm having some difficulty making sense of a logistic regression.
I think I have the interpretation right based on a university tutorial on how to interpret data.
The outcome being observed is visual acuity change after surgery (day 1).
Is it correct to say that with each increase in age (10 years) there is a decrease in 0.05 units?
The categorical variables confuse me a bit more.
Do residents of the South experience fewer gains, specifically 0.02 unit less compared to those living in the North?
Should the intercept be reported? Or can it be ignored in some circumstances?
Last edited by jle; 02-12-2015 at 12:46 AM.
Are you sure you are looking at logistic regression?
Stop cowardice, ban guns!
Sorry I meant linear regression
Your interpretations seem correct for the most part. You would want to include the intercept unless you standardized the beta coefficients. South experiences on average a 0.02 decrease in visual acuity compared to the North.
Stop cowardice, ban guns!
Thanks hlsmith, could you explain to me how the intercept is taken into the interpretation?
Is is something that can be excluded from academic writing or should it always be included for completeness?
You are actually constructing a linear combination. So you are plotting a combination of the covariates in relationship to the dependent variable, so picture a line on a 2 axes graph. Now the intercept tells you where the line crosses (intercepts) the y-axis. Now if you wanted to interpret your model in a picture you would take the intercept plus all of the beta coefficients(multiplied by the value of interest) to get the slope of the line that predicts y.
This probably does a better job of describing whether you should include it depending on the purpose of the analysis.
http://blog.minitab.com/blog/adventu...nt-y-intercept
Stop cowardice, ban guns!
Tweet |