Close Home Forum Sign up / Log in

Good way to present multiple regression?

T

In my papers I put multiple regression models in a table, but for a presentation at a conference I am wondering if there is a more visually appealing way to show things. For instance, say you were assessing whether 10 predictors were associated with a dependent variable, and you found that 2 or the 10 were. Is there a good way to show this other than a massive table? I am thinking not... But perhaps there is a good way to show how those 2 significant predictors related to the outcome - on a graph?

Thanks any stats buffs!

TQ

P

To compare your whole model with the null, area under ROC analysis is pretty cool. Otherwise tables is the usual way.

Would a radar chart work?

T

Thanks folks! I don't know what either of those things is but will have a google!

55271