Setting priors

One can set priors with the appropriate arguments to the model function

Argument	Used in	Applies to
prior_intercept	All modeling functions except stan_polr and stan_nlmer	Model intercept, after centering predictors.
prior	All modeling functions	Regression coefficients. Does not include coefficients that vary by group in a multilevel model (see prior_covariance).
prior_aux	stan_glm*, stan_glmer*, stan_gamm4, stan_nlmer	Auxiliary parameter, e.g. error SD (interpretation depends on the GLM).
prior_covariance	stan_glmer*, stan_gamm4, stan_nlmer	Covariance matrices in multilevel models with varying slopes and intercepts. See the stan_glmer vignette for details on this prior.

The stan_polr, stan_betareg, and stan_gamm4 functions also provide additional arguments specific only to those models:

Argument	Used only in	Applies to
prior_smooth	stan_gamm4	Prior for hyper-parameters in GAMs (lower values yield less flexible smooth functions).
prior_counts	stan_polr	Prior counts of an ordinal outcome (when predictors at sample means).
prior_z	stan_betareg	Coefficients in the model for phi.
prior_intercept_z	stan_betareg	Intercept in the model for phi.
prior_phi	stan_betareg	phi, if not modeled as function of predictors.

Example

attendance_bglm <- stan_glm(daysabs ~ math + gender + prog,
                            data = attendance, 
                            family = poisson, 
                            prior = student_t(df = 7))