R tutorial: Poisson Regression

This video shows you how to run and report a Poisson regression in R. It includes testing model fit and producing incidence risk ratios. Also included is a way of creating a table of the output. Data and code can be found here drive.google.c...

[@BayesRules]

This is very useful. Thanks for sharing

[@ryanandsarah]

2nd question. If you have a negative coefficient of -0.5 , and then you exponentialize it, it becomes positive. Would you still interpret that exponential coefficient as a negative number when considering how it impacts the DV?
Thanks again!

[@DrPC_statistics_guides]

The exponent of a - figure will be less than one. The exponent of 0 (no change, is 1). So think about exponents as multipliers, so exponentialised negative coefficients gives a value less than one so when acting as a multipler it has a negative effect. So an exponentialised regression coefficient of .5 means a 50% reduction. An exponentialised coefficients of 1.5 means 50% increase. An exponentialised coefficient of 1 means no effect

[@ryanandsarah]

How do I interpret and differentiate the Exponential Coefficients from the original Coefficients?
It seems like they are both telling me the degree to which the DV is impacted. To what end would I use the Exponential Coefficient over the original? My only thought is that the Exp Co should be used when we are using exponential functions such as Poisson and Negative Binomials, and standard Coefficients should be used for Linear Regression models.
Thank you!!

[@DrPC_statistics_guides]

They tell you the same thing in a different way. Exp coefficients are more useful in giving the reader real world effects as they relate to going up by one count